Understanding protein structure-function: a complex network view

Dozent	Somdatta Sinha, Indian Institute of Sciene Education and Research (IISER)
Ort	ZBH, Bundestr. 43, Raum 16
Zeit	13. Jul 2023 16:00

This talk aims to elucidate functional variations in proteins and their mutants that have no significant difference in their three-dimensional structures using network theory.

Automated perception of chemical patterns to derive force fields from scratch

Dozent	Tobias Hüfner, Max-Planck Institute for Biophysics, Frankfurt am Main
Ort	ZBH, Bundesstr. 43, Raum 16
Zeit	06. Jul 2023 16:00

In this talk I will present a method for the automated data-driven perception of discrete chemical patterns that define force types (for instance, harmonic bond stretching and bending) in a molecular mechanics forcefield. The presented method discovers these patterns based on the parameter gradients of a forcefield model with respect to high-level reference data. Using this method, I will then demonstrate the automated derivation of a small molecule forcefield from scratch.

Bioinformatics in the clinical oncology lab

Dozent	Henning Stehr, Department of Pathology, Stanford University
Ort	ZBH, Bundesstr. 43, Raum 16
Zeit	22. Jun 2023 16:00

I will talk about the life as a bioinformatician in a clinical pathology lab and present some recent projects in clinical cancer genomics, hematology and machine learning.

Graph-theoretic concepts in bioinformatics: from modularization and assembly of protein structure complexes to analysis of signal transduction pathways using Petri nets

Dozent	Ina Koch, Zentrum für Bioinformatik an der Goethe-Universität Frankfurt a.M.
Ort	ZBH, Bundesstr. 43, Raum 16
Zeit	11. May 2023 16:00

The talk focuses on different applications of graph-theoretic concepts in bioinformatics. The first part introduces complex graphs that describe protein structure complexes at the quaternary structure level.

We discuss the application of complex graphs to define functional modules and predict assemblies of protein structure complexes. In both applications we use the human respiratory complex I as case study.

In the second part, we consider specific invariant-based analysis techniques in the context of Petri net models of signal transduction pathways. We illustrate methods and results for the TNFR1-induced signaling pathway.

Mathematical Modeling of Infection and Immune Response

Dozent	Lars Kaderali, Institute for Bioinformatics, University Medicine Greifswald
Ort	CSSB c/o DESY, Lecture Hall, Notkestr. 85, Building 15, 22607 Hamburg
Zeit	27. Apr 2023 16:00

The past two years of the Covid pandemic have made mathematical modeling of viral infections broadly visible to the general public. While models have been used to predict future infection numbers as much as number of patients in intensive care and on ventilators, there are also less visible models that describe the molecular processes within a host during infection, using a combination of high throughput experimental data, bioinformatics analysis, as well as detailed mathematical modeling using differential equations. These models can impact on drug development and antiviral treatment strategies.

The talk will bridge from intracellular models of infection and immune response over models at the tissue and organ level to patient and drug treatment data, and I will also briefly touch upon models used to describe infections at the population level.

Exploring Ultra-Large Chemical Spaces

Dozent	Kai Sommer, Computational Life Sciences, Crop Science, Bayer AG, Frankfurt
Ort	ZBH, Bundesstr. 43, Raum 16
Zeit	16. Feb 2023 16:00

Virtual Chemical Spaces are huge and offer great potential for finding compounds from the unexplored chemical space.

Since vendors traditionally provide compounds for pharmaceutical research we analyzed the spaces with respect to commercially available herbicidal compounds.

Ligand discovery for GPCRs: hits, misses and new opportunities

Dozent	Peter Kolb, Philipps University Marburg, Pharmaceutical Chemistry, Marburg
Ort	Bundesstr. 45, grosser Hörsaal Pharmazie
Zeit	19. Jan 2023 16:00

The ever-increasing number of G protein-coupled receptor structures provides ample opportunities for strucuture-based design. At the same time, we can explore novel modes of action and reflect critically on why projects have (not) worked.

Biohybrid materials based on redesigned protein cages

Dozent	Tobias Beck, Physical Chemistry, UHH
Ort	Bundesstr. 45, grosser Hörsaal Pharmazie
Zeit	12. Jan 2023 16:00

We have established a novel method for the self-organization of biomolecular building blocks and nanoparticles. Here, protein cages are engineered with opposite surface charge by employing computational redesign, and are used as an atomically precise ligand shell for the assembly of inorganic nanoparticles. 

From RNA-RNA interactions to CRISPR guide RNA design

Dozent	Jan Gorodkin, University of Copenhagen, Veterinary and Animal Sciences, Copenhagen, DK
Ort	Bundesstr. 45, grosser Hörsaal Pharmazie
Zeit	15. Dec 2022 16:00

Computational methods for transcriptome-wide RNA-RNA interaction predictions requires efficient mapping of query RNAs to the transcipts. The principles were subsequently employed to CRISPR/Cas9 guide RNAs (gRNAs) for on- and off-target scoring via RNA-DNA binding energies and deep learning. Employing the resulting energy model we observe that efficient gRNAs are confined to a sweet-spot binding energy interval.

Machine Learning around the Design-Make-Test Cycle in Drug Discovery

Dozent	Marwin Segler, Microsoft, Cambridge, U.K.
Ort	Bundesstr. 45, grosser Hörsaal Pharmazie
Zeit	01. Dec 2022 16:00

Many tasks in Medicinal and Synthetic Organic Chemistry are about Pattern Recognition. In this talk, we will discuss what modern Machine Learning can and cannot do (yet) to help in this regard.

Neural population coding in the cerebral cortex - a neuroinformatics approach

Dozent	Stefano Panzeri, Dept. of Excellence for Neural Information Processing, UKE, Hamburg
Ort	Bundesstr. 45, grosser Hörsaal Pharmazie
Zeit	24. Nov 2022 16:00

In this work I will present our algorithmic approach, based on information theory and machine learning, to understand how functions of the brain such as perception and decision-making arise from the interactions between neurons in the cerebral cortex.

Cancer therapeutic target discovery using patient based multi-omic data

Dozent	Jonathan Woodsmith, VP Advanced Analytics and AI, Indivumed GmbH, Hamburg
Ort	Bundesstr. 45, grosser Hörsaal Pharmazie
Zeit	17. Nov 2022 16:00

Indivumed GmbH leverages a global clinical network to collect and analyze thousands of cancer patient samples using a standardized approach to ensure biospecimen quality. This multi-omic resource can be combined with clinical data to drive precision oncology approaches. Here we will discuss approaches for integration of multi-omic data to refine cancer sub-type definition, and why this is important in the context of early phase therapeutic discovery pipelines.

Bringing AI into biomedicine at the UKE

Dozent	Stefan Bonn, Medical Systems Biology, UKE, Hamburg
Ort	Bundesstr. 45, grosser Hörsaal Pharmazie
Zeit	10. Nov 2022 16:00

In my presentation I will introduce the bAIome center for biomedical AI to the audience. The main focus of the center is to bring AI-driven applications from explaratory and test phases into clinical routine. I'll highlight a couple of prevalent examples.

Harnessing Computational Chemistry for Drug Discovery – an Industry Perspective

Dozent	Dr. Andreas Bergner, Boehringer Ingelheim, Medicinal Chemistry, Computer-Aided Drug Design, Vienna, Austria
Ort	ZBH, Raum 16
Zeit	30. Jan 2020 16:00

Computational chemistry has transformed most aspects of drug discovery. Beyond the application of the scientifically best and most appropriate methods, the constraints and virtues in an industry setting need to be embraced, to unleash the full potential of computational chemistry. The talk will outline BI's strategy to manage these challenges from a broader perspective, and via selected real case examples.

Molecular Mechanisms of Host-Pathogen interaction: Structure and Function of Bacterial Type III Secretion Systems

Dozent	Prof. Dr. Michael Kolbe, Helmholtz Zentrum für Infektionsforschung, Strukturelle Infektionsbiologie, Hamburg
Ort	ZBH, Raum 16
Zeit	12. Dec 2019 16:00

The Type III Secretion System is a conserved syringe-shaped protein translocation nanomachine with a mass of about 4 MDa essential for the survival and virulence of many Gram-negative bacterial pathogens. This system is composed of a membrane-embedded basal body, a cytoplasmic sorting platform and an extracellular needle that deliver effector proteins into host cells. We have analyzed and compared the structures of Type III Secretion Systems from Shigella flexneri and Salmonella typhimurium.

When you can't get high: integrative structural modeling from low-resolution data

Dozent	Dr. Jan Kosinski, European Molecular Biology Laboratory (EMBL) Hamburg, Centre for Structural Systems Biology (CSSB), Hamburg
Ort	ZBH, Raum 16
Zeit	05. Dec 2019 16:00

Recently, revolutions in electron microscopy (EM) brought structures of numerous complexes at near-atomic resolution. However, for many complexes the high resolution is still difficult to achieve. Integrative structural modeling allows determining the structure of such complexes by combining low-resolution data from complementary techniques, such as X-ray crystallography, EM, NMR, SAXS or cross-linking mass spectrometry. I will present our integrative modeling approach and how we used it to model some of the largest complexes in the cell.

Tools and challenges in time-resolved structural biology

Dozent	Prof. Dr. Arwen Pearson, Center for Free-Electron Laser Science (CFEL), The Hamburg Center for Ultrafast Imaging (CUI), Universität Hamburg
Ort	ZBH, Raum 16
Zeit	21. Nov 2019 16:00

Recent advances in detectors, sources and sample delivery methods have opened the door to a new era of time-resolved structural studies. However, challenges of how to apply these methods to all macromolecular systems and how to appropriately analyse and interpret the data obtained remain. I will present the current state of the art, and describe some of the current open challenges in the field.

Drug Design in the exascale supercomputers era - The EXSCALATE.EU Project

Dozent	Dr. Andrea R. Beccari, Dompé Pharma R&D - EXSCALATE.eu, Napoli, Italy
Ort	ZBH, Raum 16
Zeit	17. Oct 2019 16:00

The advent of exascale computing on heterogeneous supercomputers offers a tremendous opportunity to foster the rethinking the actual drug design process aiming to co-design and develop novel and more advanced tools for the identification of more valuable clinical candidates. Thanks to the H2020 European research program we present a promising new approach to the drug discovery process based on the implementation  of an exscale-ready drug discovery platform. The very first real world application is devoted to the identification of Zika inhibitors by using a poly-pharmacological approach.

Aspects of Maximum Likelihood Phylogeny Reconstruction

Dozent	Dr. Heiko Schmidt, Center for Integrative Bioinformatics Vienna, University of Vienna / Max F. Perutz Laboratories, Vienna, Austria
Ort	ZBH, Raum 16
Zeit	04. Jul 2019 16:00

The Maximum Likelihood (ML) approach is one of the most robust and frequently used to reconstruct phylogenetic trees and, thus, evolutionary histories of species or gene families. ML phylogeny reconstruction implies a number of steps, aspects and obstacles, of which several will be discussed during the talk.

Reproducibility in computational science and data science

Dozent	Prof. Dr. Hans Fangohr, European XFEL, Hamburg
Ort	ZBH, Raum 16
Zeit	06. Jun 2019 16:00

I will discuss the topic of reproducibility in science, and in particular in computational and data science. I will outline the challenge, touching on software quality in computational research, good practice, workflows, metrics for academic success, expectations from journals, open science, Jupyter Notebooks and containers.

Proteoforms – the new challenge in Proteomics

Dozent	Prof. Dr. Hartmut Schlüter, Universitätsklinikum Hamburg-Eppendorf, Institut für Klinische Chemie und Laboratoriumsmedizin, Diagnostik-Zentrum, Hamburg
Ort	ZBH, Raum 16
Zeit	09. May 2019 16:00

Most of the current approaches in proteomics follow the "one gene - one protein" hypothesis. For protein identification, proteins are digested into peptides and then analyzed by mass spectrometry (MS). By comparing the MS results with protein databases (using search engines like Mascot) proteins can be identified. Because of the digestion step, different species (proteoforms), which are all coded by the same gene, cannot be distinguished any more, thus losing information about the functions of the analyzed proteins. This problem can be circumvented by top-down mass spectrometry. However, this new approach is still in its infancy and new methods are required for analytical access to the approximately 1 billion proteoforms in the human organism.

Computational chemical biology: what is missing?

Dozent	Prof. Dr. Paul Czodrowski, Technische Universität Dortmund, Fakultät für Chemie und Chemische Biologie, Dortmund
Ort	ZBH, Raum 16
Zeit	25. Apr 2019 16:00

Several case studies from different aspects of computational chemical biology will be presented. These case studies not only encompass structure-based drug design but also contain almost historic and recent contributions to cheminformatics.

Microbial communities in contact with environmental pollutants

Dozent	Dr. Johannes Werner, Leibniz Institute of Baltic Sea Research, Department of Biological Oceanography, Rostock-Warnemünde
Ort	ZBH, Raum 16
Zeit	11. Apr 2019 16:00

Microorganisms from diverse taxonomic branches have been shown to survive and sometimes even thrive in habitats contaminated with xenobiotic substances, such as plastics or herbicides. Comparative genomics and meta-omic analysis provide leads for possible degradation pathways, evolutionary traits and the ecological impact of microbial communities.

Structural studies of integral membrane proteins using stealth carrier nanodiscs

Dozent	Prof. Dr. Hennig Tidow, The Hamburg Centre for Ultrafast Imaging (CUI) & Dept. of Chemistry, University of Hamburg
Ort	ZBH, Raum 16
Zeit	04. Apr 2019 16:00

I will present recent results on deuterium labelled ‘stealth carrier’ nanodiscs (sND) that are effectively invisible to low-resolution neutron diffraction and enable structural studies of integrated membrane proteins in a lipidic native-like solution environment.

Biotite: A comprehensive framework for computational molecular biology in Python

Dozent	Patrick Kunzmann, Fachbereich Biologie, Technische Universität Darmstadt
Ort	ZBH, Raum 16
Zeit	31. Jan 2019 16:00

Analyzing biological sequence and structure data can be tedious work. This talk presents Biotite, a Python package for computational molecular biology, that provides easy-to-use but still efficient methods for analyzing, modifying and visualizing such data.

Accelerated MD and QM to explore conformational preferences and molecular interactions

Dozent	Dr. Julian Fuchs, Boehringer Ingelheim
Ort	ZBH, Raum 16
Zeit	24. Jan 2019 16:00

An application of accelerated molecular dynamics simulations to probe conformational preferences of small molecules and macrocycles is demontrated along with an exploration of preferred pi-stacking geometries based on large-scale quantum-mechanical calculations.

What structural bioinformaticians should know about serial crystallography and the xfel

Dozent	Dr. Anton Barty, Centre for Free Electron Laser Science, DESY, Hamburg
Ort	ZBH, Raum 16
Zeit	17. Jan 2019 16:00

The European XFEL, recently completed in Schenefeld, is one of the brightest X-ray light sources in the world.  Serial crystallography for protein structure determination, unlike conventional protein crystallography, relies on the measurement of thousands of individual crystals one after the other at frames rates of up to a megahertz or more.  The high frame rate is ideal for studying protein dynamics and producing molecular movies of proteins in action.  The deluge of data challenges conventional data analysis methods, while at the same time presenting exciting new opportunities for new approaches to data analysis, structure determination and interpretation.

QIAGEN – Working as a bioinformatician in a biotech company

Dozent	Inga Lemme, QIAGEN GmbH, Biotechnology, Hilden
Ort	ZBH, Raum 16
Zeit	06. Dec 2018 16:00

A short overview about my way from being a biologist to becoming a bioinformatician followed by an introduction to QIAGEN and a description about my daily work in the software department as data analyst.

Phytoplankton Adaptation to Global Warming - Insights from Resurrection Experiments and Ecosystem Modelling

Dozent	Prof. Dr. Inga Hense, Universität Hamburg, Institute for Marine Ecosystem and Fishery Science (IMF), Hamburg
Ort	ZBH, Raum 16
Zeit	29. Nov 2018 16:00

Some phytoplankton species produce long-living resting stages that can survive decades in the sediment. We activated 100- and 2-year-old resting stages of a spring bloom dinoflagellate from the sediment of the Gulf of Finland (Baltic Sea) and investigated whether differences in temperature dependent traits exist between historic and recent strains. Our resurrection experiments indicate that the rate of resting stage formation is significantly lower in recent compared to historic strains. Using an ecosystem model that considers mutation and selection of phytoplankton, we demonstrate that lowering the resting stage formation rate is in fact beneficial under global warming. However, the magnitude of change in the resting stage formation rate is only reproduced by the model if additional factors are taken into account.

A Conceptual Framework to Explore Tissue Organization

Dozent	Prof. Dr. Olaf Wolkenhauer, Systems Biology Bioinformatics, University Rostock
Ort	ZBH, Raum 16
Zeit	08. Nov 2018 16:00

It has been argued that the primary goal of modeling is explanation, not prediction. This rings particularly true in the life sciences, where the complexity of the systems under consideration renders most representations more abstract than desirable, and thereby less predictive. As a consequence, the value of mathematical modeling in biology and medicine often lies in the process of constructing the model, enforcing a scientific habit of mind that guides decisions to which variables to measure and thereby supporting the formation of hypotheses.

Exploring tissue organization through modeling it is necessary to link mechanistic models of subcellular processes with principles by which cells interact within a tissue. With cells and tissues reciprocally determining the functioning of each other, tissue organization is a key instance of part-whole relationships. Using Mathematical General Systems Theory (Mesarovic and Takahara 1975), we can abstract from mechanistic state space representations of intracellular processes. Deriving a state space representation of intra-cellular processes, we can show that from a tissue’s perspective, there are multiple pairs of states and stimuli that yield the same cellular response. The induced equivalence classes give evidence for the idea that the cell has a degree of autonomy, allowing us to distinguish between constituent interdependence between levels of organization and intra-level causal relations. This shifts our attention from modelling how something works, mechanistically, to how something works, in principle, in a tissue.

In my case study, the autonomy of a cell within a tissue is a discovery made in the process of modeling, not a result arising from the analysis of a model. The development of a conceptual framework to study tissue organization serves as an example for model-based exploration in systems biology.

A Simplex As It Gets - RNA Seq analysis in the compositional data space

Dozent	Prof. Dr. Steve Hoffmann, Leibnitz-Institut für Alternsforschung, Fritz-Lipmann-Institut e.V., Bioinformatik für Alterungsprozesse: Der Algorithmus des Alterns, Jena
Ort	ZBH, Raum 16
Zeit	25. Oct 2018 16:00

His talk emphasises the fact that the (re-) analysis of data with novel methods may yield interesting new insights. The main focus of the presentation is the detection of differential alternative splicing. This process contributes to the phenotypic diversity in eukaryotes and has been associated with a number of diseases. With the advent of high throughput RNA sequencing protocols these qualitative changes of the transcriptome can be investigated with a much deeper resolution. However, the size and number of RNA sequencing data sets require much faster algorithms to detect differential alternative splicing. Here, we will present and discuss a new method, DIEGO, for the detection of such events. In essence, it achieves significantly better run times by the transformation of split-read alignment data, i.e. the most direct evidence for splicing events, to the simplex space. At the same time it shows comparable or even better benchmarks with respect to sensitivity and specificity.

How do you build and validate 1500 models and what can you learn from them? An automated and reproducible system for building predictive models for bioassay data

Dozent	Dr. Greg Landrum, KNIME AG
Ort	ZBH, Raum 16
Zeit	05. Jul 2018 16:00

Here we describe an automated workflow for building and training predictive models for bioassay data and the application of that workflow to train and validate more than 1500 predictive models for the assay data present in ChEMBL 23. The workflow is implemented with the KNIME Model Factory in the open-source KNIME Analytics Platform. Since we know that there is no single best machine-learning algorithm/chemical fingerprint combination for all datasets, our workflow tries a variety of different fingerprints and algorithms for each assay and selects the one that performs best. The breadth of methods we use, and the automation of the process sets this effort apart from other large-scale modeling exercises with ChEMBL.

We begin with an overview of the KNIME Model Factory itself, and then describe the individual steps used to build and validate the predictive models. We’ve also started analyzing the models and the misleadingly high predictive accuracies we observe. We will close with a presentation of what we’ve learned so far about combinations of fingerprints/algorithms/parameters which seem to work well across this very large collection of different datasets.

The KNIME Model Factory is open source and can be freely downloaded from our website (URL provided during the presentation). The files for the final models and datasets are large, but we will also make those available upon request. Although we have worked with public data (ChEMBL), applying the workflow described here to other data sources (for example an internal data warehouse) would only require modification of the section that selects and extracts the data from the database.

Use of virtual screening in the discovery of the first known inhibitors of p300/HAT, an epigenetic target for oncology

Dozent	John H. van Drie, Van Drie Research LLC, North Andover, MA, USA
Ort	ZBH, Raum 16
Zeit	14. Jun 2018 16:00

DNA is regulated epigenetically by reversible acetylation and deacetylation.  The latter is performed by histone de-acetylases (HDAC's), which have been a target for drug discovery for many years, leading to drugs like vorinostat for the treatment of cutaneous T cell lymphoma.  Acetylation is performed by a HAT's (histone acetyltransferases), targets which have proven intractable for drug discovery, because no inhibitors have been found to serve as starting points for medicinal chemistry optimization.   Today, I will describe the use of virtual screening in collaboration with Acylin Therapeutics to identify the first such molecules, against one enzyme of that family, p300/HAT.  Crucial to that success was the ability to predict possible large-scale conformational changes on binding its co-substrate, acetyl CoA.  These inhibitors were acquired by AbbVie as leads for their med chem optimization which culminated in A-485, a potential drug candidate with excellent pharmacokinetics and animal efficacy in androgen receptor-positive prostate cancer.  

Genome sequencing paves the way for genomic medicine: the pulmonary arterial hypertension point of view

Dozent	Dr. Stefan Gräf, University of Cambridge, Department of Medicine, Cambridge, U.K.
Ort	ZBH, Raum 16
Zeit	07. Jun 2018 16:00

The NIHR BioResource for Rare Diseases in the UK has sequenced the entire genomes of more than 13,000 individuals diagnosed with one of 15 rare diseases, one of which is pulmonary arterial hypertension (PAH) also known as high blood pressure in the lungs. I will introduce the project, talk about some of the challenges we had to overcome during such a large scale project and present most recent results that already emerged from the analysis of this rich whole genome sequencing data resource. In particular, I will focus on the identification of novel rare and common protein-coding and non-coding sequence variation underlying PAH.

Bacterial gene regulation: The interdependent network of gene regulation and metabolism is robust where it needs to be

Dozent	Prof. Dr. Marc-Thorsten Hütt, Computational Systems Biology, Life Sciences & Chemistry, Bremen
Ort	ZBH, Raum 16
Zeit	31. May 2018 16:00

Recently, a framework for interdependent networks has emerged in the context of Statistical Physics. In a first quantitative application of this framework to Systems Biology, we study the interdependent network of gene regulation and metabolism for the model organism Escherichia coli in terms of a biologically motivated percolation model. Considering its response to perturbations that are localized with respect to functional criteria, we find the interdependent system to be sensitive to gene regulatory and protein-level perturbations, yet robust against metabolic changes. We expect this approach to be applicable to a range of other interdependent networks.

Fragments are fabulous. Fragment-based docking, molecule comparisons, and compound development

Dozent	Prof. Dr. Peter Kolb, Philipps-University Marburg, Pharmaceutical Chemistry, Marburg
Ort	ZBH, Raum 16
Zeit	17. May 2018 16:00

Fragments as low-molecular-weight compounds have a high likelihood of binding to a protein target of interest. They are useful as building blocks in order to extend small-molecule binders and for suggesting novel ligands. Moreover, we use them as a means to compare ligands to each other on a more granular level.

Phylogeography and Evolution of band-winged grasshoppers: worldwide case studies

Dozent	Dr. Martin Husemann, CeNak, UHH, Zoological Museum, Hamburg
Ort	ZBH, Raum 16
Zeit	12. Apr 2018 16:00

The band-winged grasshoppers are one of the most diverse groups within Orthoptera. Due to their complex behaviour and distribution, but difficult genomic makeup, they represent an interesting taxon for evolutionary research. Here I present different evolutionary and phylogeographic studies on this diverse grasshopper taxon.

DataWarrior - An Open-Source Tool for Interactive and Automated Analysis of Drug Discovery Data

Dozent	Dr. Thomas Sander, Idorsia Pharmaceuticals Ltd, Allschwil, Switzerland
Ort	ZBH, Raum 16
Zeit	25. Jan 2018 16:00

First I will explain basic DataWarrior concepts as data visualization, data filtering and cheminformatics functionality. Then I will show two advanced use cases in the field of medicinal chemistry, which involve a scaffold analysis and an activity cliff visualization on top of a chemical space plot.

Designing Interactive Visual Analytics Tools for Exploration of Microbial Communities

Dozent	Dr. Sara Fernstad, Newcastle University, School of Computing, Newcastle upon Tyne, U.K.
Ort	ZBH, Raum 16
Zeit	18. Jan 2018 16:00

Studies of microbial ecology are increasingly common in many research areas. The rapid development in sequencing capability enables detection of large numbers of microbes in samples. Efficient analysis methods are crucial for gaining relevant insights and understanding of the ecology of microbial communities. This talk will present methods for designing visual analysis tools for interactive exploration of data from such studies. It will describe a user and task centred design approach, provide suggestions of useful visualization methods, dimension reduction approaches and dimension and sample ordering approaches to support tasks of relevance for the analysis.

Regulatory RNA - Experiment, Data Analysis & Prediction

Dozent	Prof. Björn Voß, Institute of Biochemical Engineering, Computational Biology, Stuttgart
Ort	ZBH, Raum 16
Zeit	11. Jan 2018 16:00

Prokaryotic complexity is limited by the quadratic growth of protein based regulatory networks. In order to overcome this limitation and to evolve more complex organisms, other kinds of regulation have to be employed, one of which is RNA-based regulation. In my talk, I will present the experimental and computational strategy for, as well as our first results on studying RNA-based regulatory networks.

Data mining the open medicinal chemistry literature: From selectivity profiling to time-trend analyses and beyond.

Dozent	Dr. Barbara Zdrazil, University of Vienna, Pharmaceutical Chemistry, Vienna, Austria
Ort	ZBH, Raum 16
Zeit	30. Nov 2017 16:00

The open data revolution has enabled researches to mine data from diverse data sources and across different domains. In drug-discovery, especially the availability of public small-compound bioactivity databases (e.g. ChEMBL) broadened our mechanistic understanding of certain diseases and toxic events. In addition, such data bases present a useful repertoire for studying the synthetic and testing efforts of the drug-discovery community in a temporal context.
We used these resources in both ways, first, focusing on related targets of interest for studying molecular features driving target-selectivity, secondly following a target-independent scaffold analysis of compounds being published in the medicinal chemistry literature over time.
In conclusion, data-informed decision making is a very important asset nowadays in drug-discovery which can inform in very early stages of drug discovery and has the potential to effectively reduce attrition rates.

Simulation assisted and feed-back guided automation of enzyme engineering

Dozent	Dr. Uwe Jandt, TUHH, Bioprocess and Biosystems Engineering, Hamburg
Ort	ZBH, Raum 16
Zeit	16. Nov 2017 16:00

The modification and even more de novo construction of novel enzymatic and multienzymatic bioreaction cascades is of high interest for biotechnological and medical applications. Besides high-throughput screening of numerous, often randomly generated enzyme variants, model-based rational design is increasingly applied to create enzymes with completely new properties. Ideally, both methods should be combined and complement each other in an automated manner, minimizing manual effort and possibly overcoming the problem of “combinatorical explosion” when targeting multiple modification sites in parallel. In this talk, I will present this concept and its ongoing integration based on multi-scaled simulation predictions and a robotic setup, in order to achieve the simulation-experimental feedback loop for enzyme engineering.

Adventures in ADMET Modeling

Dozent	Prof. Bob Clark, Cheminformatics at Simulations Plus, Inc., Indiana University, USA
Ort	ZBH, Raum 16
Zeit	09. Nov 2017 16:00

To work, a drug has to reach its molecular target within the body in a form that is effective, stick around until it has accomplished its task, and then be eliminated. The same is true of environmental toxins. Pharmacokinetic analysis helps anticipate potential problems in both cases, and getting that right requires robust models for absorption, distribution, excretion and toxicological (ADMET) properties. The talk will focus on the machine learning techniques involved in building such models - and knowing when you have succeeded.

Integrative approaches and workflows to identify disease specific molecular signatures

Dozent	Dr. Shailendra K. Gupta, University of Rostock, Systems Biology and Bioinformatics, Rostock
Ort	ZBH, Raum 16
Zeit	02. Nov 2017 16:00

Exploring large scale networks is an art where several analytic tools are integrated in a computational workflow to identify disease specific small regulatory modules. These small modules are subjected to a more detailed analysis using mathematical modeling for the prediction of disease signatures, i.e. set of network derived diagnostic/prognostic biomarkers and therapeutic candidates. I will present an integrative workflow to study large-scale biochemical disease networks by combining techniques from bioinformatics and systems biology in the context of tumor metastasis.

Using Machine Learning to predict and understand human disease

Dozent	Prof. Dr. Stefan Bonn, University Clinic Hamburg-Eppendorf, Center for Molecular Neurobiology, Hamburg
Ort	ZBH, Raum 16
Zeit	26. Oct 2017 16:00

With the advent of 'big' and 'quantitative' biomedical across domains it might be possible to predict and understand disease using Machine Learning approaches. I will present recent data showing that we are starting to predict basic biological processes and disease within the limits of measurement error.

Optimal decision making in a random environment

Dozent	Prof. Sören Christensen, UHH, FB Mathematik, Stochastische Prozesse und ihre Anwendungen, Hamburg
Ort	ZBH, Raum 16
Zeit	19. Oct 2017 16:00

In the case of uncertainty the danger of making wrong decisions is inevitable. In this talk I give an introduction to the mathematical field of stochastic control which deals with such problems.

New methodologies for simulating quantum dynamics in complex chemical systems

Dozent	Dr. Ralph Welsch, DESY Theory Group, Hamburg
Ort	ZBH, Raum 16
Zeit	29. Jun 2017 16:00

The fundamental and detailed understanding of molecular reactions and light-molecule interactions is a central challenge in physical chemistry. However, the theoretical and computational treatment of quantum mechanical effects of nuclei in these reactions is a challenging task. In this talk, I will describe state-of-the-art methods that exactly or approximately include nuclear quantum effects like tunneling, zero-point energy, etc. in simulations of the time-evolution of molecular systems.

(Big) Data-Driven Drug Discovery

Dozent	Dr. Gerard JP van Westen, LACDR, Medicinal Chemistry, Leiden, NL
Ort	ZBH, Raum 16
Zeit	22. Jun 2017 16:00

We aim to explore potential synergy at the border of disciplines, while basing our decisions on a background in biopharmaceutical sciences. In the talk I will outline recent and current developments in data-driven drug discovery in Leiden. I will present some previous and ongoing work. Concrete applications are mode-of-action elucidation (finding the protein(s) responsible for an observed phenotype), precision medicine (integrating natural variation in predictive affinity models), and GPU accelerated deep learning applied to cheminformatics.

Small angle X-ray scattering, structural biology(blobology) and the future

Dozent	Dr. Cy M. Jeffries, EMBL, Hamburg
Ort	ZBH, Raum 16
Zeit	15. Jun 2017 16:00

Structural biologists have a huge choice when it comes to investigating the structures of biomacromolecules. Small angle X-ray scattering (SAXS) is just one of a dazzling array of techniques on offer at large-scale facilities and infrastructures here in Hamburg. The EMBL-P12 bioSAXS beam line located at the PETRAIII synchrotron on the DESY campus provides a fully-automated approach to measuring SAXS data to obtain low-resolution structural information from proteins, protein assemblies, complexes and other biological macromolecules in solution. But why do we need low resolution ‘blobs’ and what do they tell us? And what is the future of SAXS in today’s over-crowded and ultra-competitive structural biology market: Electron microscopy, crystallography, tomography, ultra-resolution light microscopy and now XFEL? Why SAXS?

Open data and resources for drug discovery

Dozent	Dr. Andrew R. Leach, EMBL-EBI, Hinxton, U.K.
Ort	ZBH, Raum 16
Zeit	01. Jun 2017 16:00

The biological sciences have a long tradition of “open data” yet until relatively recently this has not been the case in the chemical and chemical biology sciences. The advent of open, public resources such as ChEMBL is now changing that situation. ChEMBL currently contains more than 15m activity values on more than 1.7m compounds, curated from the literature, from data depositions and from other relevant sources. In this talk I will draw upon my experience in large Pharma to illustrate how the EBI’s resources such as ChEMBL can be used to tackle a wide range of questions in drug discovery and chemical biology and to help make informed, data-driven decisions on progression. I will also discuss our plans for the evolution of these resources and invite feedback and suggestions.

Sensitive protein sequence searching and clustering for the analysis of massive metagenomic datasets

Dozent	Dr. Johannes Söding, MPI for Biophysical Chemistry, Quantitative and Computational Biology, Göttingen
Ort	ZBH, Raum 16
Zeit	11. May 2017 16:00

I will present our new method MMseqs2 (Many-against-Many sequence searching) for very fast batch protein sequence and profile searches of huge protein sequence data sets. Protein sequence searching is the main time, cost and quality bottleneck for the analysis of metagenomic datasets. While previous search methods sacrificed sensitivity for speed gains, MMseqs2 is as sensitive as BLAST, more sensitive than PSI-BLAST, and two to three orders of magnitude faster. I will explain the ideas that led to this massive improvement. MMseqs2 searching and clustering will considerably increase the fraction of annotatable metagenomic ORFs. I will also present Linclust, a linear time-complexity clustering method with which we could cluster 1.7 billion protein sequences in 1.5 days on a single server.

No stress - heat shock proteins and life cycle control in Leishmania parasites

Dozent	Dr. Joachim Clos, Bernhard Nocht Institute for Tropical Medicine, Hamburg
Ort	ZBH, Raum 16
Zeit	04. May 2017 16:00

David versus Goliath: Antibiotic Inhibition of Ribosomes and Protein Synthesis

Dozent	Prof. Dr. Daniel N. Wilson, UHH, Biochemie und Molekularbiologie, Hamburg
Ort	ZBH, Raum 16
Zeit	20. Apr 2017 16:00

Templates for semantic metadata annotations of data sets

Dozent	Dr. Manfred Kohler, Fraunhofer IME, Life Science Data Solutions, Hamburg
Ort	ZBH, Raum 16
Zeit	13. Apr 2017 16:00

In life sciences, a huge number of experiments produces a lot of small data sets. This high frequency combined with a variability of experiments, demand high workload on documentation. In reality the quality of documentation is far away from optimal.
We propose a simpler way to create useful descriptions of experimental work by using predefined templates accessible via PIDs. The templates themselves can be used to create a graph database showing the hierarchical structure of experimental conditions, useful for finding the correct experimental template as well as for correct analytical combinations of experimental work based on semantic information across different research areas. The intended method could be used for new as well as for existing experiments. The presentation should be used as a starter for a community based discussion.

Controlling Molecules for the recording of action movies

Dozent	Prof. Dr. Jochen Küpper, Center for Free-Electron Laser Science, DESY and University of Hamburg
Ort	ZBH, Raum 16
Zeit	26. Jan 2017 16:00

Leveraging linked open data – from transporter interaction profiles to toxicity prediction

Dozent	Prof. Dr. Gerhard F. Ecker, University of Vienna, Department of Pharmaceutical Chemistry, Wien, Austria
Ort	ZBH, Raum 16
Zeit	19. Jan 2017 16:00

With the public availability of large data sources such as ChEMBL and the Open PHACTS Discovery Platform, retrieval of data sets for certain protein targets of interest, measured under consistent assay conditions is no longer a time consuming process. Especially the use of workflow engines such as KNIME or Pipeline Pilot allows the submission of complex queries and enables the search for several targets simultaneously. Within this talk, case studies for the exploitation of linked open data for the development of ligand-transporter interaction models and their use for predicting complex in vivo toxicity endpoints will be presented. Furthermore, we will outline workflows for analysis of phenotypic screening data and transporter profiling.

Applications of in-memory database technology in the field of healthcare and life sciences

Dozent	Dr. Hans-Christian Ehrlich, SAP SE, Innovation Center Potsdam, Healthcare & Life Science, Potsdam
Ort	ZBH, Raum 16
Zeit	12. Jan 2017 16:00

Advances in the analytical capabilities of in-memory database systems over the last decade made them a promising technology for applications in the field of healthcare and life sciences. At the same time, the data volume acquired in healthcare and life science is increasing exponentially and already at a point where using a solid database technology is unavoidable to leverage the maximal information from the available data. 

In this talk, I like to give an overview of SAP's activities in the field of healthcare and life sciences and demonstrate how scientists can benefit from SAP technology.

COSMO-RS: The currently most accurate solvation model and its extension to drug design

Dozent	Prof. Dr. Andreas Klamt, COSMOlogic GmbH & Co. KG, Leverkusen
Ort	ZBH, Raum 16
Zeit	01. Dec 2016 16:00

The COSMO-RS method is a unique combination of quantum chemistry, continuum solvation theory and statistical thermodynamics. Currently it provides the most accurate predictions of solvation free energies in pure and mixed solvents, and at variable temperature. It is widely used and proven in chemical engineering and pharmaceutical development, but its special description of molecular interactions also allows for interesting extensions to drug design.

On Using Structural Models

Dozent	Dr. Manfred S. Weiss, Helmholtz-Zentrum für Materialien und Energie, Macromolecular Crystallography, Berlin
Ort	ZBH, Raum 16
Zeit	24. Nov 2016 16:00

When using structural models from the Protein Data Bank, care must be taken to account for the limitations that such models present. Validation tools often provide the necessary information, but the validation results themselves need to be interpreted properly. In the talk, I will present several structure validation tools and I will elaborate on several cases, in which careful validation would have shed light on serious deficiencies in structures determined by crystallography.

Computational drug design: Applying new algorithms for novel challenges

Dozent	Prof. Dr. Gerhard Wolber, FU Berlin, Institute of Pharmacy, Berlin
Ort	ZBH, Raum 16
Zeit	17. Nov 2016 16:00

Virtual screening using ligand- and structure-based methods has become an essential part of drug design. Medicinal chemists typically apply such algorithms in a pragmatic and creative way and finally use evaluation metrics such as enrichment factors or receiver operating characteristics to benchmark their methods for a specific application scenario. The nature of the algorithms is mostly not an important part of the design of a study and the potential of choosing the right method often underestimated.

In this talk several lead discovery and optimization success stories will be presented using 3D pharmacophores, a method, for which knowledge bias is part of model design. Such knowledge can be gained from experimental data or other computational tools like molecular dynamics simulations. Existing 3D pharmacophore screening methods will be critically reviewed and novel ideas for generating 3D pharmacophore density functions from MD simulations (dynophores) will be presented.

Zwitterions and polar polymers: stealth reagents for tumor targeting and antifouling surfaces

Dozent	Prof. Dr. Wolfgang Maison, Pharmazeutische und Medizinische Chemie, UHH
Ort	ZBH, Raum 16
Zeit	10. Nov 2016 16:00

Conjugation of zwitterions and polar polymers such as oligoethylene glycol (OEG) to drugs can alter their pharmacokinetics dramatically. In addition, zwitterionic and polar coatings can be used to create tailored materials for applications in implant medicine and clinical hygiene. The lecture discusses the use of zwitterions and OEG as stealth reagents for tumor imaging and the design of antifouling surfaces. New imaging reagents based on the clinically used chelator DOTA and antimicrobial coatings for clinically relevant metals and plastics will be described.
 

MD-Simulations in Pharmaceutical Drug Discovery

Dozent	Dr. Thomas Fox, Boehringer Ingelheim Pharma GmbH & Co. KG, Ingelheim am Rhein
Ort	ZBH, Raum 16
Zeit	27. Oct 2016 16:00

Atomistic molecular dynamics (MD) simulations are close to becoming routine computational tools for drug discovery. Their main advantage is in explicitly treating structural flexibility and entropic effects, thus allowing to gain insight into the thermodynamics and kinetics associated with drug?target recognition and binding. I will give a short introduction to the drug discovery process in a pharmaceutical company, highlighting some of the possible pitfalls one has to be aware. Then I will present a few examples how we employ MD simulations to support drug discovery projects, ranging from conformational analysis of macrocycles, the prediction of binding poses, the mapping of interaction sites on protein surfaces to the calculation of binding free energies.

Application of rigidity theory to the thermostabilization of proteins

Dozent	Prof. Dr. Holger Gohlke, Pharmaceutical and Medicinal Chemistry, University of Düsseldorf
Ort	ZBH, Raum 16
Zeit	07. Jul 2016 16:00

Application of rigidity theory to the thermostabilization of proteins
Prakash Chandra Rathi¹, Alexander Fulton², Karl-Erich Jaeger², and Holger Gohlke¹*
¹Institute for Pharmaceutical and Medicinal Chemistry, Department of Mathematics and Natural Sciences, Heinrich Heine University, Düsseldorf, Germany
²Institute of Molecular Enzyme Technology, Heinrich-Heine-University Düsseldorf and Research Centre Jülich, Jülich, Germany
Protein thermostability is a crucial factor for biotechnological enzyme applications. Protein engineering studies aimed at improving thermostability have successfully applied both directed evolution and rational design. However, for rational approaches, the major challenge remains the prediction of mutation sites and optimal amino acid substitutions. Recently, we showed that such mutation sites can be identified as structural weak spots by rigidity theory-based thermal unfolding simulations of proteins.^1-4 Here, we describe and validate a unique, ensemble-based,⁵ yet highly efficient strategy to predict optimal amino acid substitutions at structural weak spots for improving a protein’s thermostability. For this, we exploit the fact that in the majority of cases an increased structural rigidity of the folded state has been found as the cause for thermostability. When applied prospectively to lipase A from Bacillus subtilis, we achieved both a high success rate in predicting thermostabilized lipase variants and a remarkably large increase in those variants’ thermostability based on single amino acid mutations. The results suggest that our strategy is a valuable complement to existing methods for rational protein design aimed at improving thermostability.

1. Radestock, S. & Gohlke, H. Exploiting the link between protein rigidity and thermostability for data-driven protein engineering. Eng. Life Sci. 8, 507-522 (2008).
2. Radestock, S. & Gohlke, H. Protein rigidity and thermophilic adaptation. Proteins 79, 1089-1108 (2011).
3. Rathi, P.C., Radestock, S. & Gohlke, H. Thermostabilizing mutations preferentially occur at structural weak spots with a high mutation ratio. Journal of Biotechnology 159, 135-44 (2012).
4. Rathi, P.C., Jaeger, K.E. & Gohlke, H. Structural Rigidity and Protein Thermostability in Variants of Lipase A from Bacillus subtilis. PLoS One 10, e0130289 (2015).
5. Pfleger, C. & Gohlke, H. Efficient and robust analysis of biomacromolecular flexibility using ensembles of network topologies based on fuzzy noncovalent constraints. Structure 21, 1725-34 (2013).

Control of Th17 cells

Dozent	Prof. Dr. Samuel Huber, Dr. Nicola Gagliani, UKE, Molekulare Immunologie und Gastroenterologie, Hamburg
Ort	ZBH, Raum 16
Zeit	23. Jun 2016 16:00

Th17 cells play a key role for the defense against harmful pathogens. However, if uncontrolled, Th17 cells can also promote chronic inflammatory and autoimmune disease. Thus a proper control of Th17 cells is essential. This control is exerted on several layers. i)Th17 cells can be suppressed by regulatory T cells, such as Tr1 cells. ii) Th17 cell associated cytokines can be blocked by endogenous inhibitors. iii) Th17 cells can be transdifferentiated into regulatory T cells. The aim of our research is to understand these mechanisms in order to develop new therapeutic approaches for autoimmune and chronic inflammatory diseases.

Computational Design in Pharmaceutical Drug Discovery

Dozent	Prof. Dr. Hans-Joachim Böhm, F. Hoffmann-La Roche AG, Basel, Switzerland
Ort	ZBH, Raum 16
Zeit	16. Jun 2016 16:00

Computational Design approaches are an important and integral part of modern drug discovery. Over the last years, the impact of Computational Chemistry has significantly improved due to i) rapid access to 3D structural information on the target protein, ii) new approaches to analyze and score protein-ligand interactions and iii) new approaches to make the large amounts of available structure-property data actionable. I will present some recent work to illustrate these points.

Navigation in Chemical Space Towards Biological Activity

Dozent	Dr. Peter Ertl, Novartis Institutes for BioMedical Research, Basel, Switzerland
Ort	ZBH, Raum 16
Zeit	02. Jun 2016 16:00

One of the most common tasks that cheminformatics experts in pharmaceutical industry face almost daily, is analysis and visualization of large collections of molecules. Typical areas, where this is needed, are analysis and enhancement of company compound archive, analysis of high-throughput screening data, design of combinatorial libraries, chemogenomics analyses and many others. This presentation will provide an overview of various methods used to analyze and medicinal chemists visualize chemical space with particular focus on needs of medicinal chemists.

Structure-based design of epigenetic inhibitors – Application of binding free energy calculations

Dozent	Prof. Dr. Wolfgang Sippl, Pharmazie, Martin-Luther-Universität Halle Wittenberg
Ort	ZBH, Raum 16
Zeit	26. May 2016 16:00

In the present work the structure-based design of novel epigenetic inhibitors for several histone modifying enzymes and readers is presented. With the usage of X-ray crystallography, molecular docking and binding free energy calculations nanomolar inhibitors could be obtained. The advantages of virtual screening as well as the drawbacks of binding free energy prediction will be discussed.
References:
M. Roatsch, et al. Substituted 2-(2-aminopyrimidin-4-yl)pyridine-4-carboxylates as potent inhibitors of JumonjiC domain-containing histone demethylases.
Future Med. Chem., 2016 Mar 14. [Epub ahead of print]
M. Schiedel, et al. Structure-based Development of a Sirtuin 2 Affinity Probe. Angew. Chem. Int. Ed. Engl.  55, 2252-6, 2016.
T. Heimburg, et al. Structure-Based Design and Synthesis of Novel Inhibitors targeting HDAC8 from Schistosoma mansoni for the Treatment of Schistosomiasis
J. Med. Chem, 2016 Mar 14. [Epub ahead of print]
 

Rapid Phylogenies

Dozent	Fabian Klötzl, MPI für Evolutionsbiologie, Plön
Ort	ZBH, Raum 16
Zeit	12. May 2016 16:00

Computing phylogenies is often an integral part of comparative genomics.
This is usually done via evolutionary distances calculated from a multiple sequence alignment. However, aligning large samples of whole genomes is slow. To compute distances for thousands of genomes, we developed the anchors distance, which approximates distances without computing a full alignment. In this presentation I will talk about how suffix arrays can be leveraged to quickly look up exact matches between sequences. Unfortunately, alignment-free methods do not allow the calculation of support values via bootstrapping. To overcome this, we have recently implemented two strategies for support values which do not require an MSA. Lastly, I will give a sneak peak at the MSA tool we are trying to build with our knowledge on fast sequence matching.

Linking phenotypic differences between species to differences in their genomes

Dozent	Dr. Michael Hiller, MPI of Molecular Cell Biology and Genetics, Dresden
Ort	ZBH, Raum 16
Zeit	21. Apr 2016 16:00

Cell Based Assays In Drug Discovery

Dozent	Dr. Sheraz Gul, Fraunhofer-IME SP, Hamburg
Ort	ZBH, Raum 16
Zeit	07. Apr 2016 16:00

The changes in the approaches to assay development and high throughput screening for the major drug target classes in early stage drug discovery will be described.
This will followed by examples illustrating the advantages, disadvantages and challenges associated with screening using purified protein targets, cell-based reporter systems, primary cells, 3D-cultures and re-programmed cells.
 

opm: an R package for analysing OmniLog® phenotype microarray data and the specific task of metadata management

Dozent	Dr. Lea Vaas, Fraunhofer-Institut für Molekularbiologie und Angewandte Oekologie IME, Hamburg
Ort	ZBH, Raum 16
Zeit	07. Jan 2016 16:00

The talk will give an overview about the measurement technique, describe the resulting data type and introduce a dedicated R-package including a workflow for the data analysis.
With this basis, we will see what types of metadata could be added and which strategies are available for their management.
Literature: LAI Vaas, J Sikorski, B Hofner, A Fiebig, N Buddruhs, HP Klenk, M Göker (2013) opm: an R package for analysing OmniLog® phenotype microarray data. Bioinformatics 29 (14), 1823-1824 LAI Vaas, J Sikorski, V Michael, M Göker, HP Klenk (2012) Visualization and curve-parameter estimation strategies for efficient exploration of phenotype microarray kinetics. PLoS One 7 (4), e34846

Meth camp: the revelation of differential cytosine methylation

Dozent	Dr. Steve Hoffmann, University Leipzig, Transcriptome Bioinformatics
Ort	ZBH, Raum 16
Zeit	10. Dec 2015 16:00

The methylation of DNA cytosines (5-mC) is an epigenetic modification that has an impact on gene regulation. Bisulfite sequencing protocols allow one to measure 5-mC levels of whole genomes in a matter of days. The avalanche of sequence data and methylation rates, however, makes the comparison of signals between different samples and conditions computationally quite expensive. Here, we will take a look at methods to detect regions of differential methylation (DMRs), introduce a novel algorithm that uses a non-parametric test and develop some ideas to investigate this important epigenetic layer.

Ecological Prospects from Microbial Metagenomes: Hydrothermal Vents as Model Systems

Dozent	Prof. Mirjam Perner, Universität Hamburg, Microbiology & Biotechnology, Molecular Biology of Microbial Consortia
Ort	ZBH, Raum 16
Zeit	26. Nov 2015 16:00

Hydrothermal vents host rich metabolically and phylogenetically distinct microbial communities. The local chemolithoautotrophic microorganisms live off of the inorganic energy sources that are provided by the hydrothermal fluids and thus provide the basis of life in these hydrothermal systems. Until now, it has been enigmatic which environmental parameters influence the distribution of these microorganisms. Here we present some highlights suggesting some of the environmental parameters that are important for determining the hydrothermal vent community.

Local adaptation in European populations of Daphnia - is it phenotype, genotype, both?

Dozent	Prof. Mathilde Cordellier, Universität Hamburg, Biozentrum Grindel
Ort	ZBH, Raum 16
Zeit	19. Nov 2015 16:00

Structural insights into nutrient uptake transporters

Dozent	Dr. Christian Löw, EMBL Hamburg
Ort	ZBH, Raum 16
Zeit	29. Oct 2015 16:00

The Nature of Known Drug Space

Dozent	Dr. Jóhannes Reynisson, School of Chemical Sciences, University of Auckland, New Zealand
Ort	ZBH, Raum 16
Zeit	22. Oct 2015 16:00

Known drug space (KDS) is defined as all molecules that are used in the clinic [1]. It can be used as a metric in exploring ideas within the framework of drug discovery programmes, i.e., the presence of certain undesirable molecular moieties [2]. This begs the question what is the nature of KDS? It was analysed for the occurrence of natural products and their derivatives. It was found that 10% of the drugs on the market are unaltered natural products, 29% are their derivatives (semi-synthetics) and the rest (61%) have a synthetic origin [3]. Also, KDS occupies a larger volume in chemical space with respect to drug-like chemicals in respect to mainstream molecular descriptors (see figure) [3]. Furthermore, the size of chemical space is estimated to be 1.6×109 compounds based on the power growth ((f(x) = AxB, Carbon=x)) of known organic compounds. The statistical distribution of KDS and drugs with good oral-bioavailability (drug-like) based on their carbon atom count was used to deduce the size of these sub-regions within chemical space. KDS is predicted to be 2.0×106 molecules and drug-like chemical space is calculated to be 1.1×106 compounds [4]. Finally, it was found that serendipity has played a crucial part in populating KDS [5].
[1] A. Mirza, R. Desai, J. Reynisson, Eur. J. Med. Chem., 44 (2009) 5006-5011.
[2] P. Axerio-Cilies, I.P. Castañeda, A. Mirza, J. Reynisson, Eur. J. Med. Chem., 44 (2009) 1128-1134.
[3] R. Bade, H.-F. Chan, J. Reynisson, Eur. J. Med. Chem., 45 (2010) 5646-5652.
[4] K.L.M. Drew, H. Baiman, P. Khwaounjoo, B. Yu, J. Reynisson, J. Pharm. Pharmaco., 64 (2012) 490–495.
[5] E. Hargrave-Thomas, B. Yu, J. Reynisson, W. J. Clin. Oncol., 3 (2012) 1-6.

Protein translation and mutations: the good, the bad and the aging

Dozent	Prof. Dr. Zoya Ignatova, Institut für Biochemie, Universität Hamburg
Ort	CHEMIE, Hörsaal B
Zeit	15. Oct 2015 16:15

Exploring protein dynamics for ligand design

Dozent	Prof. Rebecca C. Wade, Heidelberg Institute for Theoretical Studies (HITS), Molecular and Cellular Modeling
Ort	ZBH, Raum 16
Zeit	08. Sep 2015 16:00

The dynamic nature of protein structures provides challenges and opportunities for ligand design. I will first illustrate this by discussing results of recent computational studies of thymidylate synthase and cytochrome P450 enzymes [1.2]. I will then describe computational tools that we have developed to compare protein binding sites [3] and to explore protein motions affecting ligand binding to identify transient pockets for ligand design [4].
References
[1] Salo-Ahen, O.M. H., Tochowicz, A., Pozzi, C., Cardinale, D., Ferrari, S., Boum,Y., Mangani, S., Stroud, R.M., Saxena, P., Myllykallio, H., Costi, M.P., Ponterini, G., and Wade, R.C. Hotspots in an Obligate Homodimeric Anticancer Target. Structural and Functional Effects of Interfacial Mutations in Human Thymidylate Synthase J. Med. Chem., (2015), 58, 3572–3581.
[2] Yu, X., Cojocaru, V., Mustafa. G. Salo-Ahen, O.M.H., Lepesheva, G.I. and Wade, R.C. Dynamics of CYP51: implications for function and inhibitor design. J. Molec. Recogn. (2015), 28, 59-73.
[3] Fuller, J.C., Martinez, M., Henrich, S., Stank, A., Richter, S. and Wade, R.C. LigDig: a web server for querying ligand-protein interactions. Bioinformatics, (2014), doi:10.1093/bioinformatics/btu784. http://mcm.h-its.org/ligdig/
[4] Kokh DB, Richter S, Henrich S, Czodrowski P, Rippmann F, and Wade RC. Trapp: a tool for analysis of transient binding pockets in proteins. J. Chem Inf Model, (2013) 53:1235-52. http://mcm.h-its.org/trapp/

Robust workflows for probing genome-wide DNA methylation in the context of social insect comparative genomics

Dozent	Prof. Dr. Volker Brendel, Department of Biology, Indiana University, Bloomington, USA
Ort	ZBH, Raum 16
Zeit	04. Sep 2015 14:00

DNA methylation has been implicated in epigenetic control of gene expression and organism-level phenotypes in honeybees and other social insects.
We have developed a complete workflow for experimental DNA methylation data analysis that ensures reproducible and comparable treatment of data, including correction of technical biases that blur DNA methylation calls.
Re-analysis of several recent studies on the Hymenoptera gives consistent results that suggest considerable sample variation in methylation levels and sites and thus caution against over-interpretation of these data with respect to general roles of DNA methylation.

From computer-aided design to clinical testing: Creating chemical probes for the investigation of GPCR pharmacology

Dozent	Prof. Robert C. Glen, University of Cambridge, Department of Chemistry
Ort	ZBH, Raum 16
Zeit	09. Jul 2015 16:00

Apelin (APJ) is a class-A G-protein-coupled receptor (GPCR). It is activated by the Apelin family of peptides, which are hormones of various lengths cleaved from the N-terminus of the 77 residue precursor hormone preproapelin. Apelin-13 (QRPRLSHKGPMPF) and (Pyr1)apelin-13 (pyroglutamate-RPRLSHKGPMPF) are two isoforms present in vivo and the latter is the most predominant isoform in human heart tissue. Recent work has highlighted several roles for apelin in development, water intake, cardiovascular tone, as an HIV co-receptor and in angiogenesis. Apelin is the most potent inotrope yet discovered. This peptide has remarkable properties, and we have developed biased agonists and antagonists, which we have used to probe the receptor pharmacology of Apelin in Cardiovascular disease, Diabetes and Cancer. We recently reported the first in human study [1] of a biased Apelin agonist and the antagonist has unmasked the role of Apelin in sugar uptake. In cancer, Apelin antagonists arrest the growth of tumours. This presentation will describe the computer-aided molecular design of biased agonists and antagonist and their pharmacology.
[1] Design, characterization, and first-in-human study of the vascular actions of a novel biased apelin receptor agonist. AL Brame, JJ Maguire, P Yang, A Dyson, R Torella, J Cheriyan, M Singer, RC Glen, IB Wilkinson, AP Davenport – Hypertension (2015) 65, 834 (DOI: 10.1161/hypertensionaha.114.05099)

GDCh-Kolloquium: Einfache und dabei genaue Quantenchemische Methoden

Dozent	Prof. Dr. Stefan Grimme, Universität Bonn
Ort	FB Chemie, Martin-Luther-King-Platz 6, Hörsaal B
Zeit	25. Jun 2015 16:15

Feature-based 3D Pharmacophore Model Extension Using Molecular Dynamics Simulation

Dozent	Prof. Thierry Langer, University of Vienna, Pharmaceutical Chemistry, Vienna, Austria
Ort	ZBH, Raum 16
Zeit	18. Jun 2015 16:00

Pharmacophore-based compound modelling, virtual screening, and bio-activity profiling has become one of the most popular in silico techniques for supporting medicinal chemists in their hit finding, hit expansion, hit to lead, and lead optimization programs. (1)
At Inte:Ligand GmbH, we developed the program LigandScout (2) as an integrated software solution containing rapid and efficient tools for automatic interpretation of ligand-protein interactions and subsequent transformation of this information into 3D chemical feature-based pharmacophore models. Additionally, pattern recognition-based algorithms were developed for ligand-based pharmacophore modelling in the absence of a target 3D structure, as well as for establishing novel accurate virtual screening procedures. Our recent interest is to incorporate the results of molecular dynamics simulation trajectories into the pharmacophore description, in order to develop pharmacophore ensembles representing the dynamic event of binding and to make this functionality available as LigandScout KNIME extensions.
(1) Langer, T., Pharmacophores in Drug Research, Mol. Inf. 2010, 29, 470-475.
(2) Wolber, G., Langer, T. ; LigandScout: 3D Pharmacophores Derived from Protein-Bound Ligands and their Use as Virtual Screening Filters, J. Chem. Inf. Model. 2005, 45, 160-169.
(3) KoNstanz Information MinEr, available from KNIME.COM AG, Zurich, Switzerland (http://knime.org)

Towards personalized cancer therapies: from high-throughput data to individualized vaccines

Dozent	Prof. Oliver Kohlbacher, Eberhard Karls Universität Tübingen, Applied Bioinformatics
Ort	ZBH, Raum 16
Zeit	11. Jun 2015 16:00

Epitope-based vaccines have gained a lot of interest in cancer immunotherapy. By raising an immune response against tumor-specific or tumor-associated antigens, a direct targeting of the tumor is possible. The personalized nature of the cancer and of a patient’s immune system implies that the vaccine has to be tailored to every patient. We show how the selection of these epitopes can be formulated as a combinatorial optimization problem and how this approach can be integrated into clinical processes.

Comprehensive druggability analysis to prioritize the untargeted kinome

Dozent	Dr. Andrea Volkamer, BioMed X Innovation Center, Heidelberg
Ort	ZBH, Raum 16
Zeit	04. Jun 2015 16:00

Protein kinases are involved in a variety of diseases including cancer, inflammation, and autoimmune disorders. Although the development of new kinase inhibitors is a major focus in pharmaceutical research, a large number of kinases remained so far unexplored in drug discovery projects. A wealth of data is available regarding kinase structures and profiling data and can be used to answer the question which kinase to address next.

A thorough analysis of the druggability ATP pockets of the entire human kinome – including crystal structures and homology models - will be presented in order to prioritize novel kinase structures for drug discovery projects. The results indicate that - from a structural perspective - around 75% of the kinome have binding site characteristics that should allow the design of drug-like compounds. Several aspects will be discussed including the top ranking kinases and the information gained from experimental hit rates.

Unraveling biomolecular structure with theoretical vibrational spectroscopy

Dozent	Prof. Dr. Christoph Jacob, TU Braunschweig, Institut für Physikalische und Theoretische Chemie
Ort	ZBH, Raum 16
Zeit	28. May 2015 16:00

The determination of biomolecular structure and dynamics is at the core of a detailed understanding of biological functionality. Methods such as X-ray diffraction and NMR spectroscopy are well established, but the time-resolution of these techniques does not allow for the study of very flexible structures or of the fast dynamical processes occurring during protein folding. This gap could be closed by vibrational spectroscopy, which potentially offers femtosecond time resolution. However, extracting structural information from the vibrational spectra of biomolecules remains a challenging task. In particular specialized chirooptical techniques such as vibrational circular dichroism (VCD) and Raman optical activity (ROA) can provide very specific spectroscopic information. Moreover, two-dimensional infrared (2D-IR) spectroscopy gives access to anharomonic couplings, which encode specific structural information. However, quantum chemical calculations are necessary for extracting structural information from these vibrational spectra.
In my talk, I will give an overview of our efforts to unravel biomolecular structure with the help of theoretical vibrational spectroscopy. This includes methods for the analysis of calculated spectra of large biomolecules in terms of local modes as well as novel approaches for the efficient calculation of anharmonic vibrational spectra.

Does Flexibility Control the Specificity of Protein-Protein and Protein-Drug Interfaces?

Dozent	Prof. Dr. Klaus Liedl, Leopold-Franzens-University Innsbruck, Centre for Chemistry and Biomedicine (CCB), Innsbruck, Austria
Ort	ZBH, Raum 16
Zeit	30. Apr 2015 16:00

Compact Indexes for Flexible Top-k Retrieval

Dozent	Dr. Simon Gog, University of Melbourne, Dept. of Computing and Information Systems, Parkville, Australia
Ort	ZBH, Raum 16
Zeit	24. Apr 2015 14:00

In this talk, I will present a self-index based retrieval system capable of rank-safe evaluation of top-k queries. The framework generalizes the GREEDY approach of Culpepper et al. (ESA 2010) to handle multi-term queries, including over phrases. We propose two techniques which significantly reduce the ranking time for a wide range of popular Information Retrieval (IR) relevance measures, such as TF×IDF and BM25. First, we reorder elements in the document array according to document weight. Second, we introduce the repetition array, which generalizes Sadakane’s (JDA 2007) document frequency structure to document subsets. Combining document and repetition array, we achieve attractive functionality-space trade-offs. We provide an extensive evaluation of our system on terabyte-sized IR collections. This is joint work with Matthias Petri (The University of Melbourne).

Towards efficient and accurate hybrid atomistic/coarse-grained simulations

Dozent	Prof. Sereina Riniker, ETH Zürich, Physical Chemistry
Ort	ZBH, Raum 16
Zeit	23. Apr 2015 16:00

Coarse-graining allows one to increase the computational speed of classical molecular dynamics simulations by one or two orders of magnitude by reducing of the number of interacting particles and a smoother potential energy surface, but at the price of a loss of information. Hybrid atomistic/coarse-grained simulations in which a part of the system, e.g. the protein, is kept at the atomistic level are therefore a promising approach to benefit from the speed-up of coarse-grained solvent while retaining atomistic detail for the protein.

Combining bioinformatics and molecular modeling to understand and design enzymes

Dozent	Prof. Jürgen Pleiss, Institute of Technical Biochemistry, University of Stuttgart
Ort	ZBH, Raum 16
Zeit	16. Apr 2015 16:00

Enzymes are promising catalysts for a wide range of biocatalytic processes. However, finding the optimal biocatalyst for a new biocatalytic application may become a tedious process. In most cases, the commercially available enzymes will be inactive or unselective. Therefore, it will be necessary to develop an optimal biocatalyst with appropriate stability, catalytic activity, wide substrate range, and the desired regio­- and stereoselectivity. Because protein engineers (in contrast to mechanical or electrical engineers) still lack a deep understanding of their system under investigation, trial-and-error strategies are widely used.
To get a deeper insight into the function of enzymes and to improve the efficiency of enzyme development processes, we follow two main strategies: mining of the rapidly increasing sequence data by systematically analyzing protein family databases, and molecular modeling to study the molecular basis of substrate recognition and solvent effects. Data mining and molecular modelling are two complementary approaches which allow us to learn about the relationships between sequence, structure, and function of enzymes, to search for promising enzyme candidates that catalyze the desired reaction, and to design enzyme variants with improved properties.

Quantum-chemical structure optimization of biomolecules: are the popular methods really the best possible choice?

Dozent	Dr. Lars Goerigk, The University of Melbourne, Australia
Ort	ZBH, Raum 16
Zeit	09. Apr 2015 16:00

With the advent of high-performance-computing architectures, the full quantumchemical treatment of biomolecular structures — even of entire proteins — has become feasible.
However, a review of the literature reveals that despite recent developments of accurate and cost-efficient methods, only a few standard approaches are usually applied in the field of computational biomolecular quantum chemistry. A common problem with those popular approaches is that they do not describe Londondispersion correctly and that they introduce artefacts due to what quantum chemists call the “basis-set-superposition error” (BSSE).
In this presentation, a short introduction into these two problems will be given before the aforementioned popular approaches will be examined for the structural optimization of biologically relevant structures. It turns out that indeed the inadequate description of London dispersion and the BSSE in these methods influence the structures of di- and tripeptides.1 More accurate alternatives are identified.
Subsequently, the linear-scaling optimization of proteins is discussed. Parallels to the results for peptide optimizations are drawn and, finally, a perspective is given on how modern quantum-chemical approaches can be combined with experiment in future applications, such as the molecular quantum refinement of biomolecular X-ray structures. 2

[1] L Goerigk, JR Reimers, J. Chem. Theory Comput. 2013, 9, 3240.
[2] L Goerigk, CA Collyer, JR Reimers, J. Phys. Chem. B 2014, 118, 14612.

1st CSSB International Symposium, 9.-11.4. in Hamburg

Dozent
Ort
Zeit	09. Feb 2015 12:59

The 1st CSSB International Symposium “Systems in Infection Biology – from Molecules to Organisms” which will take place from 9 – 11 April 2015 in the historic auditorium of Bernhard-Nocht-Institute for Tropical Medicine in Hamburg, Germany.
 
The Centre for Structural Systems Biology (CSSB) is a new research center of excellence supported by nine research partners from Northern Germany (www.cssb-hamburg.de) and is funded by the BMBF, Hamburg, Lower Saxony and Schleswig-Holstein. The research focus of the CSSB is on interdisciplinary research across structural biology, exploiting the unique X-ray infrastructure opportunities on the DESY campus in Hamburg, systems biology and infection biology.
 
The research focus of the 1st CSSB International Symposium will be on highlighting emerging opportunities in systems biology in connection with important topics in infection biology and state-of-the-art tools in structural biology. It will feature top-class international research talks in five related scientific sessions and will end with a concluding panel discussion on future perspectives of the field. The Joachim Herz Foundation (JHS), one of the largest private foundations in Germany, has kindly agreed to become the main sponsor of the symposium.
 
The following speakers have confirmed their attendance:
Patrick Aloy (Barcelona, Spain), Patrick Cramer (Göttingen, Germany), Vitor dos Santos (Wageningen, Netherlands), Andrew Emili (Toronto, Canada), James Galagan (Boston, USA), Anne-Claude Gavin (Heidelberg, Germany), Wolf-Dietrich Hardt (Zürich, Switzerland), Dirk Heinz (Braunschweig, Germany), Deborah Hung (Cambridge, USA), Edda Klipp (Berlin, Germany), Tanja Kortemme (San Francisco, USA), Andreas Kremling (Munich, Germany), Thomas Marlovits (Hamburg, Germany), John Parkinson (Toronto, Canada), Jochen Prehn (Dublin, Ireland), Maya Topf (London, UK), Jörg Vogel (Würzburg, Germany), Gerry Wright (Toronto, Canada)
 
Further information and the registration page can be found on the event website: www.cssb-symposium2015.de
 
Following the Symposium, we will hold our first CSSB Spring School “Approaches for Structural Systems Biology – From Molecules to Organisms” which will provide an introduction to advanced biophysical approaches for Structural Systems Biology. Courses in sample preparation and analysis, advanced microscopy, single particle analysis, structure determination at atomic resolution and modeling of structure and dynamic of complexes will allow young researchers (experienced students, PhD students and postdocs) to view challenging problems from interdisciplinary scientific topics like Infection Biology from various methodical angles. The material will be presented in form of lectures, demonstrations and practical work. Further information on how to apply and the programme can be found on www.cssb-hamburg.de/springschool2015

Knowledge-based Drug Discovery at Roche: Applications in structure-based design and cheminformatics

Dozent	Dr. Martin Stahl, Molecular Design and Chemical Biology, Roche Innovation Center, Basel, Switzerland
Ort	ZBH, Raum 16
Zeit	18. Dec 2014 16:00

The impact of experimental uncertainty on SAR analysis and model building in drug design

Dozent	Dr. Christian Kramer, Institute for General, Inorganic and Theoretical Chemistry, University of Innsbruck, Austria
Ort	ZBH, Raum 16
Zeit	13. Nov 2014 16:00

All experimental results contain some experimental uncertainty. This is well appreciated in almost all engineering disciplines, and good strategies for coping with the uncertainty have been developed in these disciplines. In contrast, the analysis of experimental uncertainty in physicochemical and biochemical measurements is often worked out less stringently, and the impact of the amount of experimental uncertainty on SAR analysis and drug design decisions are not rigorously taken into account.

Unraveling biomolecular dynamics - models from simulation and experiment

Dozent	Prof. Bettina G. Keller, FU Berlin, Institute for Chemistry and Biochemistry, Berlin
Ort	ZBH, Raum 16
Zeit	06. Nov 2014 16:00

Global quantification of the environmental stress response in bacteria with nucleotide resolution

Dozent	Prof. Zoya Ignatova, University of Hamburg, Biochemistry and Molecular Biology
Ort	ZBH, Raum 16
Zeit	30. Oct 2014 16:00

Environmental stress is detrimental to cell viability and requires an adequate reprogramming of cellular activities to maximize survival and adaptability. We use genome-wide approaches to analyze the response of E. coli to acute environmental stress. We compare gene expression profiles at the level of transcription, by quantifying the absolute mRNA abundance, and at the level of translation, by measuring translational efficiency using deep-sequencing approaches.

Residence time as new parameter for drug design: Insights from kinetic studies and simulations

Dozent	Prof. Dr. Christoph Sotriffer, Universität Würzburg, Pharmazie und Lebensmittelchemie, Würzburg
Ort	ZBH, Raum 16
Zeit	23. Oct 2014 16:00

The rational modulation of drug-target residence time in structure-based drug design projects is generally not straightforward because the molecular determinants governing this key parameter often remain elusive. With an enzyme as target, a detailed understanding of the mechanism of catalysis and inhibition is essential. For enoyl-ACP reductase inhibitors as model system it will be shown how simulations studies combined with accurate kinetic and structural data provide this understanding and help to rationalize the residence times on a structural basis.

Bioinformatics and High Throughput Sequencing

Dozent	Prof. Arndt von Haeseler, Center for Integrative Bioinformatics Vienna, Max F. Perutz Laboratories, Vienna, Austria
Ort	ZBH, Raum 16
Zeit	16. Oct 2014 16:00

The advent of high throughput sequencing has boosted the application of sequencing projects in modern biology. Mapping reads to a known reference genome is one of the challenges for bioinformatics. We will describe our contribution. Moreover, we will show that a good understanding of the mapping procedure is essential to come to valid biological conclusion. We show that highly polymorphic regions in a genome hinder an accurate alignment of reads, thus biasing subsequent analyses. We will discuss an illustrative biological example from Drosophila melanogaster.

Sea ice processes and observation: a view across the disciplines

Dozent	Dr. Lars Kaleschke, Center for Earth System Research and Sustainability, University of Hamburg, Partner KlimaCampus Hamburg
Ort	ZBH, Raum 16
Zeit	10. Jul 2014 16:00

The decrease of the Arctic sea ice cover is the most visible sign of climate change. Sea ice research is driven by many disciplines ranging from microbiology to climate science and applications like shipping on the Northern Sea Route. In my presentation, I will provide a general overview with a special focus on satellite remote sensing techniques.

Structural Biology - at the interface of chemistry, physics and medicine

Dozent	Dr. Markus Perbandt, UHH, Center for Ultrafast Imaging, c/o DESY, Hamburg
Ort	ZBH, Raum 16
Zeit	03. Jul 2014 16:00

The fascination and challenge of macromolecular crystallography lies in the wide range of disciplines combined in its practice. Perhaps the biggest change is that it is now conducted largely by biologists and biochemical scientists interested in specific structural problems. The talk will address this with a couple of current examples.

Binding free energy calculations - soon a standard method in drug design?

Dozent	Dr. Clara Christ, Bayer Pharma, Berlin
Ort	ZBH, Raum 16
Zeit	26. Jun 2014 16:00

Equations for the rigorous calculation of protein-ligand free energy of binding were formulated decades ago. Large scale application in industry has, however, been hampered by the tedious setup of such calculations as well as long calculation times. We present the results of an extensive evaluation of a recently developed software solution that combines a state of the art condensed phase force-field for arbitrary organic molecules, enhanced sampling techniques, statistically sound free energy estimators and GPU-based molecular simulations.

Molecular Dynamics Simulations of Biomembranes

Dozent	Dr.-Ing. Sven Jakobtorweihen, Hamburg University of Technology, Institute of Thermal Separation Processes, Hamburg
Ort	ZBH, Raum 16
Zeit	19. Jun 2014 16:00

We used molecular dynamics (MD) simulations to investigate the partition behavior of different solutes in different lipid bilayers. The umbrella sampling approach was applied to calculate free energy profiles. Subsequently, partition coefficients were calculated from these profiles. MD results were compared to the relatively new method COSMOmic and to experimental data.

Spaced Words and kmacs: new approaches to alignment-free sequence comparison

Dozent	Prof. Dr. Burkhard Morgenstern, Georg-August-Universität Göttingen, Institute for Microbiology and Genetics, Göttingen
Ort	ZBH, Raum 16
Zeit	12. Jun 2014 16:00

Alignment-free approaches are increasingly used for DNA and protein sequence comparison because they are much faster than traditional alignment-based methods. Most alignment-free algorithms compare the word-composition of sequences. Other approaches are based on the average length of longest common subwords. We propose two new methods that are based on inexact word matches. The first one is based on the composition of the input sequences from "spaced words", i.e. words with wildcard characters at certain, pre-defined positions, similar to the "spaced seeds" used in PatternHunter. Our second approach uses longest common substrings with up to k mismatches. To evaluate our approaches, we applied them to calculate phylogenetic trees which we compared to trusted reference trees. In most cases, our approaches performed better than existing alignment-free methods that are based on exact word matches.

Computational Drug Discovery in Oncology

Dozent	Dr. Friedrich Rippmann, Director Global Computational Chemistry, Merck, Darmstadt
Ort	ZBH, Raum 16
Zeit	15. May 2014 16:00

Taking the "hallmarks of cancer" as a guide, a number of highly diverse approaches to cancer drug discovery will be explored. The talk will touch on large-scale target discovery, integrins, kinases, small molecules, antibodies, and cancer vaccines. Common theme, however will be the structural understanding of disease, and the application of computational tools in support of the ultimate goal, the discovery of drugs for yet uncurable diseases.

Chemical Space Analysis of Ion Channel Ligands Identified by High-Throughput Screening

Dozent	Dr. Alexander Böcker-Felbek, Evotec AG, Hamburg
Ort	ZBH, Raum 16
Zeit	08. May 2014 16:00

At Evotec a carefully-selected maximum diversity library of more than 250,000 lead-like compounds is readily available for high throughput screening and has been applied in screening projects of several ion channel targets. We have systematically investigated the chemical space of the screening collection using principal component analysis (PCA), clustering and self-organizing maps in combination with different 2D and 3D descriptors or chemical fingerprints. The identified ion channel ligands have been compared to compounds detected as hits by cell-based HTS assays for G-protein coupled receptor ligands. Our analyses show that ion channel HTS hits are highly diverse and broadly cover 2D, 3D and fingerprint spaces. Various key physicochemical and 3D properties were further studied individually. As an example, a paucity of low molecular weight ion channel HTS hits with low lipophilicity was observed. This observation is not obvious for GPCR HTS hits. This analysis shows that simple 2D and 3D descriptors like logP/logD or molecular weight may be applied as triaging filters for HTS hit sets.

Integration of Lipidomics into Biomedical Research

Dozent	Dr. Dominik Schwudke, Forschungszentrum Borstel
Ort	ZBH, Raum 16
Zeit	17. Apr 2014 16:00

New Publication: "FISH Oracle 2: A web server for integrative visualization of genomic data in cancer research"

Dozent
Ort
Zeit	14. Apr 2014 08:14

The article describes FISH Oracle 2, a web server for the
interactive visualization of different kinds of downstream processed
genomics data typically available in cancer research. A powerful search
interface and a fast visualization engine provide a highly interactive
visualization for such data. These features support life scientists
in the process of identifying potentially tumor relevant genomic regions.

http://www.jclinbioinformatics.com/content/4/1/5/

Origins of Life Systems Chemistry

Dozent	Prof. John Sutherland, MRC Laboratory of Molecular Biology, Cambridge, U.K.
Ort	ZBH, Raum 16
Zeit	10. Apr 2014 16:00

Studying antibiotic resistant bacteria – applications for bioinformatics?

Dozent	Prof. Dr. Peter Heisig, Universität Hamburg, Inst. of Biochemistry and Molecular Biology
Ort	ZBH, Raum 16
Zeit	30. Jan 2014 16:00

The world-wide increasing prevalence of antibiotic resistant bacteria is a major cause of concern for global health. According to a recent study of the Centers for Disease Control (CDC) an estimated 2 million people are infected by antibiotic-resistant bacteria each year and at least 23.000 die from these infections causing an economical cost of antibiotic resistance of annually up to $ 20 billion in the US. Studying the molecular basis for antibiotic resistance focusses on identification of the underlying genetic variations, elucidating the cellular consequences in order to identify and evaluate potential new targets for novel drugs.
Fluorquinolones (fqs) were used as bactericidal model antibiotics targeting bacterial type II topoisomerases. Genetic characterization of consecutive mutants selected in vitro by increasing drug concentrations revealed a combination of mutations associated with high-level resistance to fluoroquinolones in vitro. While these combinations of mutations were associated with a reduced fitness in the absence of antibiotics, clinical isolates exhibiting identical resistance mutations are lacking this cost of resistance. From the study of fitness-compensated in vitro mutants, compensatory mutations affecting genes involved in DNA replication fidelity, global transcription regulation, and toxic waste disposal were identified. A transfer of these results into the clinical situation requires further characterization by high-throughput genetic analyses.

Quo vadis structure-based drug discovery in infection biology: our experiences with TB

Dozent	Dr. Matthias Wilmanns, EMBL Hamburg, Centre for Cellular and Molecular Imaging, Hamburg
Ort	ZBH, Raum 16
Zeit	23. Jan 2014 16:00

PASS and GUSAR – software to link chemical structures with biological activities / “Walking pathways” and personalized medicine

Dozent	Dr. Olga Kel-Margoulis und Dr. Alexander Kel-Margoulis, geneXplain GmbH, Wolfenbüttel
Ort	ZBH, Raum 16
Zeit	16. Jan 2014 16:00

PASS and GUSAR – software to link chemical structures with biological activities.
Abstract:
PASS (Prediction Activity Spectra for Substances) is a software to predict potential biological activities of small drug-like compounds. Its output is a list of the predicted bioactivities as well as adverse and toxic effects for each compound, accompanied by estimations of probability.  The prediction algorithm is based on Bayesian estimates of probabilities for a compound to belong to the class of active or inactive compounds. 2D chemical structures that are represented in PASS by the set of unique MNA descriptors (Multilevel Neighborhoods of Atoms). More than 7000 bioactivities can be predicted with an average accuracy of 96%.  Predicted activities are classified into pharmacotherapeutic effects, molecular mechanisms, metabolism-related effects as well as adverse and toxic effects.
GUSAR (General Unrestricted Structure-Activity Relationships) is designed to create reliable quantitative QSAR/QSPR models using a training set of 2D chemical structures and quantitative data on biological activities. These models can be further used for activity prediction. GUSAR is based on the most recent achievements in the field of QSAR modeling: consensus prediction, applicability domain assessment, validation of the models and clear interpretations of the results obtained.
Software Demo.
---
“Walking pathways” and personalized medicine.
Abstract:
Massive changes of expression of hundreds of genes as well as changes in genomic and epigenomic landscapes observed in human diseases often represent just an “echo” of relatively few causative molecular processes in the cells taking place during the transformation into the disease state, e.g. in cancer, during malignant transformation. Non-reversible structural changes in gene regulatory networks may cause a transformation of the cell homeostasis switching it from the normal state to the disease state. We call such structural network changes “walking pathways”. The analysis of this phenomenon helps us to understand the mechanisms of molecular switches (e.g. between programs of cell death and programs of cell survival) and to identify prospective biomarkers and drug targets of cancer. Such structural changes often happen due to an epigenetic “evolution” of genome regulatory regions during the realization of development programs. The structural plasticity of regulatory networks leads to formation of personalized variants of regulatory pathways in individual tumors, which requires reverse engineering and de novo reconstruction of pathways in each individual case, and explains the pure performance of the modeling efforts which are done on the basis of a single unified cancer pathway.
Online demo:

The geneXplain platform is an online workbench to assist in operating everyday computer applications in life sciences. It comprises a number of bioinformatics and systems biological modules. These modules can combine to comprehensive workflows using a workflow management system.

The geneXplain platform with the integrated TRANSFAC^® database facilitates analyses of gene regulatory regions, either proximal promoters or distant regulatory regions, e.g. peaks calculated from ChIP-seq data. Having identified transcription factors, TFs, involved in gene regulation, we analyze networks upstream of these TFs to suggest master regulatory molecules. According to our concept of upstream analysis, such master molecules provide a coordinated regulation of a number of TFs, and might be promising candidates for biomarkers and drug targets.
The concept of the upstream analysis will be illustrated by means of several publicly available examples of disease datasets from GeneExpressionOmnibus.

Using small world network concepts to study protein-ligand binding and aminoacid flexibility

Dozent	Dr. Neil R. Taylor, Desert Scientific Software, Sydney, Australia
Ort	ZBH, Raum 16
Zeit	09. Jan 2014 16:00

Network concepts have been used to build quantitative models for predicting ligand binding affinity and aminoacid flexibility in protein-ligand complexes.  Localised network descriptors were constructed in such a way as to maximise potential small-world network influences in molecular recognition.  Our aim is to better model the contributions of non-covalent interactions to ligand binding affinity and structure stability when the local environment is highly connected.

Specialized Dynamic Properties of Promiscuous Residues in Protein Complexes

Dozent	Prof. Franca Fraternali, King´s College London, Cell and Molecular Biophysics, London, UK
Ort	ZBH, Raum 16
Zeit	12. Dec 2013 16:00

Protein-protein interaction networks have been characterised by various means, but here, we have secondary structure, conservation, plasticity, promiscuity, dynamics from very large molecular dynamics simulations and more. Some of the relationships are significant, predictive and surprising.

3D-Modeling: The Next Step?

Dozent	Prof. Tim Clark, Friedrich-Alexander-Universität Erlangen-Nürnberg, Erlangen
Ort	ZBH, Raum 16
Zeit	05. Dec 2013 16:00

Techniques for 3D-molecular modeling have not exactly developed dynamically in the past two decades. Polarizable force fields remain quaint oddities, classical force fields and molecular orbital theory are stagnating and density functional theory is becoming a zoo of highly parameterized functionals, some of which may work. I will describe the development and the rationale behind the hpCADD (high-performance Computer-Aided Drug Design) force field and the techniques associated with it. The principal motivation for the project is that CADD typically uses techniques that would run adequately on a smart phone and does not take advantage of the capabilities of modern hardware.

'Big Data' & Scientific Data Analytics

Dozent	Dr. Manfred Kohler, European ScreeningPort GmbH, Hamburg
Ort	ZBH, Raum 16
Zeit	28. Nov 2013 16:00

Big Data is one of the buzz words in IT today. In this presentation the following topics will be addressed: What is 'Big Data'? 'Big Data' and Business Intelligence/Business Analytics. High level introduction to basic technologies used with 'Big Data'. What are the challenges when using Big Data solutions with scientific data. Implications using 'Big Data' approaches in projects.

Trends and barriers to lateral gene transfer during microbial evolution

Dozent	Prof. Dr. Tal Dagan, Christian-Albrechts-Universität zu Kiel, Genomische Mikrobiologie, Kiel
Ort	ZBH, Raum 16
Zeit	21. Nov 2013 16:00

Gene acquisition by lateral gene transfer (LGT) is an important mechanism for natural variation among prokaryotes. Laboratory experiments show that protein-coding genes can be laterally transferred extremely swiftly among microbial cells, inherited by most of their descendants, and adapt to a new regulatory regime within a short time. Recent advance in the phylogenetic analysis of microbial genomes using networks approach reveals a substantial impact of LGT during microbial genome evolution. Phylogenomic networks of LGT among prokaryotes reconstructed from completely sequenced genomes uncover barriers to LGT in multiple levels including (i) barriers to gene acquisition in nature including physical barriers for gene transfer between cells, (ii) genomic barriers for the integration of acquired DNA, and (iii) functional barriers for the acquisition of new genes.

The human myelin peripheral membrane protein P2 - all the way from subatomic resolution to multilayered membranes

Dozent	Prof. Petri Kursula, University of Hamburg / DESY, Dept. of Chemistry, Hamburg
Ort	ZBH, Raum 16
Zeit	14. Nov 2013 16:00

The protein P2 is abundantly expressed in vertebrate peripheral nerve myelin, where it may function in bilayer stacking and lipid transport. P2 binds to phospholipid membranes through its positively charged surface and a hydrophobic tip, and it accommodates fatty acids inside its barrel. The structure of human P2 at an ultrahigh resolution allows for detailed structural analyses, including the full organization of an internal hydrogen bonding network. The orientation of the bound fatty acid is linked to the protonation states of coordinating arginine residues. An anion binding site in the portal region may be relevant for membrane interactions and conformational changes. In membrane multilayers, P2 has a preferred orientation and gets stabilized, and the multilayer repeat distance indicates a single layer of P2 between membranes. Simulations show the formation of a double bilayer in the presence of P2, and in cultured cells, P2 induces stacked membrane domain formation. These accurate structure-function data on P2, a major component of peripheral nerve myelin, show how it can interact with two membranes simultaneously, while going through conformational changes and ligand transfer.

Toward a Dynamic Understanding of Reaction Mechanisms

Dozent	Dr. Julia Rehbein, Universität Hamburg, Physical Organic Chemistry
Ort	ZBH, Raum 16
Zeit	07. Nov 2013 16:00

Our intuitive understanding of reaction mechanisms is strongly influenced by our approach of condensing reality into models. For instance our description of reaction rates and product selectivities is based on statistical models like transition state theory (TST). These models are always correct for one-step reactions. Matters can change though if one wants to describe product selectivity of two-step reaction steps. Key preposition for TST to work in two-step reaction mechanisms is a statistically averaged energy distribution within the reactive intermediate. However, quite a few examples have been identified where the chemical transformation is faster than the intramolecular vibrational energy redistribution. Examples of so-called nonstatistical dynamic effects will be presented in the talk.^[1] Another effect that cannot be described by our current models are bifurcations (Scheme 1).^[2] In this case our traditional 2D-representation of reaction coordinate diagrams has led to a neglect of potential energy surfaces that are characterized by a ridge connecting two transition states. This ridge substitutes the path of minimum energy and the final product distribution is determined when the trajectories are approaching the valley ridge inflection point (VRI). The underlying physical properties that are the determining factors for rate and selectivity in such reactions are still elusive and are a major part of our research. The second part of the talk will give some insights into what is special of reactions on bifurcating PES and how this is linked with certain catalyzed reactions.

[1]        Carpenter, B. K. Chem. Rev. 2013, doi cr300511u

[2]        Rehbein, J.; Carpenter, B. K. Phys. Chem. Chem. Phys. 2011, 13, 20906–20922.

Algorithms and Tools for Bioinformatics on GPUs

Dozent	Prof. Dr. Bertil Schmidt, Johannes Gutenberg-Universität Mainz, Institut für Informatik, Parallele und Verteilte Architekturen, Mainz
Ort	ZBH, Raum 16
Zeit	17. Oct 2013 16:00

The enormous growth of biological sequence data has caused bioinformatics to be rapidly moving towards a data-intensive, computational science. As a result, the computational power needed by bioinformatics applications is growing rapidly as well. The recent emergence of parallel accelerator technologies such as GPUs has made it possible to significantly reduce the execution times of many bioinformatics applications. In this talk I will present the design and implementation of scalable GPU algorithms based on the CUDA programming model in order to accelerate important bioinformatics applications. In particular, I will focus on algorithms and tools for next-generation sequencing (NGS) using error correction and short read mapping as examples. I will also give an overview of several CUDA-enabled tools developed by my research group.

Selective Inhibition of the new Malaria Target PfGSK-3

Dozent	Dr. Thomas Lemcke, Institut für Pharmazie, Universität Hamburg
Ort	ZBH, Raum 16
Zeit	11. Jul 2013 16:00

Structural modelling approaches in the bioinformatics of HIV

Dozent	Dr. Olga Kalinina, Max-PLanck-Institute for Informatics, Computational Biology and Applied Algorithmics, Saarbrücken
Ort	ZBH, Raum 16
Zeit	27. Jun 2013 16:00

Human immunodeficiency virus causes AIDS, and hence presents a major health hazard worldwide. Genotyping of HIV positive patients has become a common practice in many countries. Statistical analysis of the collected data allows for successful prediction of the most effective treatment to avoid the acquired drug resistance mutations and take into account virus tropism to the co-receptor, but the structural and mechanistic basis remains unknown. Thus HIV presents a challenging target for combining sequence-driven and structure-driven bioinformatics approaches to gain new functional insights. Several case studies involving HIV capsid and envelope proteins will be discussed.

RNA Bioinformatics for Genome Analysis

Dozent	Prof. Jan Gorodkin, RTH Center for Non-coding RNA in Technology and Health, University of Copenhagen, Denmark
Ort	ZBH, Raum 16
Zeit	20. Jun 2013 16:00

While most of the human genome is transcribed, the protein coding genes take up only ~1.2% of the available sequence. Hence, the human (and the mammalian) genome has plenty of room for ncRNAs and, consistent with this, emerging in silicoand high-throughput sequence based screens suggest possibly tens of thousands of candidates. Here, I will present our efforts at in silico detection, which includes methods for structural alignments of RNA sequences as well as constructing pipelines for genomic screens. Subsequently, methods aiming to add putative information about functionality such as RNA-RNA interactions, RNA 3D information and analysis of profiles of high-throughput sequencing data are described.

Hybrid Approaches for Non-Local Optimization of Molecular Systems

Dozent	Dr. Mark P. Waller, Organisch-Chemisches Institut, Westfälische Wilhelms-Universität Münster
Ort	ZBH, Raum 16
Zeit	13. Jun 2013 16:00

Computational chemical biology is concerned with developing accurate models of complex molecular systems. Efficient and accurate conformation searching/optimization of such systems are therefore highly desirable. Typically, gradient-based local-optimization algorithms are applied[1], however these methods cannot surmount energy barriers on the potential energy surface. Therefore, if the starting point is not carefully selected, there is a high chance that the optimized conformation may not resemble the “global” best structure[2]. Here we present an efficient hybrid metaheuristic optimization (HMO) approach for non-local optimization of complex molecular systems.[3]

The ProteinModelPortal – A Comprehensive Web Resource for Protein Structure and Model Information

Dozent	Dr. Juergen Haas, Swiss Institute of Bioinformatics, Biozentrum, University of Basel, Basel, Switzerland
Ort	ZBH, Raum 16
Zeit	06. Jun 2013 16:00

The Protein Model Portal (PMP) is an open platform fostering effective use of molecular models in biomedical research by providing convenient and comprehensive access to structural information for a specific protein. For the first time, both experimental structures and theoretical models for a protein can be searched simultaneously and analyzed for structural variation. The current release which is updated at least once a month, allows searching 18.9 million model structures for 4.2 million distinct UniProt entries.Ultimately, the accuracy of a structural model determines its utility for specific applications and model quality estimation tools assist in evaluating the accuracy of generated models. We present new developments in the Protein Model Portal supporting model validation and quality estimation, which consist of (1) continuously extended service interfaces to several established modeling and model quality estimation tools (2) a novel analysis tool for protein structure variation for both models and experimental structures and (3) the CAMEO (Continuous Automated Model Evaluation, www.cameo3d.org) system for the continuous evaluation of servers predicting 3D protein structures, ligand binding site residues and the recent extension to model quality assessment programs (MQAPs). By providing a comprehensive view on structural information, the Protein Model Portal not only offers a unique environment to apply consistent assessment and validation criteria to the complete set of structural models available for a specific protein, but also allows continuous assessment of the modeling and quality estimation services registered with CAMEO.

Beyond equilibrium: From biological membranes to carbon nanotubes

Dozent	Prof. Dr. Florian Müller-Plathe, Eduard-Zintl-Institut für Anorganische und Physikalische Chemie, TU Darmstadt
Ort	ZBH, Raum 16
Zeit	30. May 2013 16:00

Molecular dynamics methods can be used to simulate dynamical processes in soft materials and biological matter. To this end, the usual boundary conditions of molecular dynamics are modified to include thermodynamic driving forces, for example concentration or temperature gradients, which cause the system to be out of equilibrium. The induced response, e.g. diffusion of molecules or flow of heat, can then be studied. This seminar represents examples of diffusion, thermal conduction and viscous flow. The systems studied range from simple fluids and ionic liquids to biological membranes and nanosystems.

Analysis and prediction of xenobiotic metabolism

Dozent	Dr. Johannes Kirchmair, Unilever Centre, University of Cambridge, UK
Ort	ZBH, Raum 16
Zeit	23. May 2013 16:00

Metabolism occurs in any living cell and alters the physicochemical and biological properties of small molecules. The significance of such metabolic processes to the performance and safety of chemical products (including drugs, agrochemicals, cosmetics, nutraceuticals, etc.) has long been underestimated. Today, an array of powerful experimental techniques is available for the determination of the metabolic fate of small molecules, but these are highly resource-demanding and time-consuming. The increasing availability of experimental data enables the development of predictive methods for metabolism, and a large variety of approaches has been reported in this area recently. These theoretical methods do not aim to generally substitute experimental techniques but to help experimentalists focus their investigations on aspects that are most relevant to solving a specific problem. This lecture starts with an analysis of >100,000 experimentally observed biotransformations, aiming to elaborate quantitative measures for the physicochemical property shifts introduced to small molecules by metabolism. In the second part we provide a brief overview of the field of predictive metabolism, in which we elaborate on the scope and limitations of state-of-the-art methods and analyse their potential in the research of bioactive molecules. Thirdly, we present novel ligand-based approaches allowing the holistic prediction of xenobiotic metabolism without being limited to a specific family of metabolic enzymes. The performance of these models is compared among established methods, and their advantages and disadvantages are highlighted.

Vertical Integration of Genome Annotation (or How to Sail off the Genomic Cliff)

Dozent	Prof. Volker Brendel, Indiana University, Biology and School of Informatics and Computing, Bloomington, IN, USA
Ort	ZBH, Raum 16
Zeit	16. May 2013 16:00

Genome annotation has struggled to keep pace with the rate of genomic sequence generation and assembly.  But without accurate annotation, genome sequences are of little value.  I will review current approaches to genome annotation as well as novel approaches pursued in my group to address the problem of fast, scalable genome annotation.

Addressing the Innovation Gap in Pharmaceutical Drug Discovery

Dozent	Dr. Björn Windshügel, European ScreeningPort GmbH, Hamburg
Ort	ZBH, Raum 16
Zeit	02. May 2013 16:00

Despite rapidly growing investments in Research and Development, the Pharmaceutical Industry is facing an innovation gap as the number of approved drugs has not increased analogous to the growth in R&D expenditure.One of the consequences is that major pharmaceutical companies are leaving less financially attractive drug discovery areas such as antibiotics research. Due to emerging resistances the antimicrobial threat has become a major global public health problem. Thus novel approaches for the development of effective antibiotics are needed of which a very recently started campaign is presented.Besides traditional drug targets (GPCRs, Kinases, etc.) new frontiers with less fierce competition are explored in order to develop innovative therapeutics. The novel area of research on protein-protein interactions (PPI) and their modulation by small organic molecules offers many attractive opportunities for drug discovery. Some of the recent success stories on the discovery of PPI modulators will be presented in this talk.

Natural history of vertebrate globins

Dozent	Prof. Dr. Wojciech Makalowski, Institute of Bioinformatics, University of Muenster
Ort	ZBH, Raum 16
Zeit	25. Apr 2013 16:00

Globins are small heme proteins that bind various external ligands, such as oxygen or nitric oxide, and they are found in all kingdoms of life. Vertebrate genomes harbor seven different globin types, including prominent hemoglobin (Hb) and myoglobin (Mb). Tetrameric Hb is present in erythrocytes and transports oxygen via blood to the inner tissues. The monomeric, muscle-specific myoglobin, on the other hand, stores oxygen, enhances the oxygen diffusion to the muscle cells and detoxifies nitric oxide. Five recent additions to vertebrate globin repertoire are cytoglobin (Cygb), neuroglobin (Ngb), globin X (GbX), globin E (GbE) and globin Y (GbY). It has been suggested that the GbE and Mb genes, as well as the GbY and proto-Hb genes, are the paralogous products of tandem gene duplications that occurred early in the evolution of vertebrates. In contrast, Ngb and GbX derive from independent duplication events that occurred before the split of Deuterostomia and Protostomia. To date little is known regarding the functions of the novel globin proteins. I will focus on our recent evolutionary studies of Ngb and GbX in which we demonstrated ancestry of these two globins reaching dating back to origins of metazoans.

Towards Novel Approaches for RNA Imaging in Living Cells

Dozent	JProf. Dr. Andrea Rentmeister, Institut für Biochemie und Molekularbiologie, Universität Hamburg
Ort	ZBH, Raum 16
Zeit	11. Apr 2013 16:00

Specific localization of mRNAs to certain subcellular regions exists in various cell types and represents a mechanism for regulating gene expression with excellent spatial-temporal control. In neurons, local protein synthesis in dendrites is required for many forms of synaptic plasticity. Studying mRNA localization, transport, translation, and degradation in subcellular regions will undoubtedly enrich our understanding of how cells and organisms function. To study the dynamics of mRNA distribution and transport, there is an increasing demand for universal probes suitable for visualization of mRNAs in living cells. We investigate RNA methyltransferases as tools for modification of 5’ cap structures. We have tested their substrate specificity on S-adenosyl-L-methionine (SAM) analogs. A colorimetric high throughput assay allows us to test RNA methyltransferase activity in high throughput in E. coli lysate. We found enzyme variants with improved properties on SAM analogs and can now produce modified capped RNAs amenable to click-chemistry. In an alternative approach we have used RNA-binding proteins and split-GFP. This system allows us to detect specific RNA-molecules in vitro and shall be extended to living cells.

Backbone order-parameter restraining in molecular dynamics simulations

Dozent	Dr. Niels Hansen, ETH Hönggerberg, Zürich, Schweiz
Ort	ZBH, Raum 16
Zeit	25. Mar 2013 16:00

Molecular dynamics simulations are used in structure refinement of biomolecules to search conformational space for structures that satisfy best a large number of spatial constraints, which originate from NMR spectroscopic or X-ray diffraction measurements: Nuclear Overhauser enhancement (NOE) atom-atom distance bounds, J-coupling constants, X-ray diffraction intensities, etc. However, limited sampling and force field inaccuracies may hamper the comparability between calculated and experimental data. To drive simulations to sample areas of conformational space that are in agreement with experimental data, a restraining function that penalizes deviation from experimental data is added to the physical interaction function. In this work we investigate the feasibility of restraining backbone N-H order parameters.

How to solve diverse, large-scale sequence comparison problems

Dozent	Dr. Martin Frith, National Institute of Advanced Industrial Science and Technology, Tokyo, Japan
Ort	ZBH, Raum 16
Zeit	21. Mar 2013 16:00

Sequence comparison has been central to computational biology for many decades, and still is today.  Modern tasks include: comparing whole genomes; aligning bisulfite-converted DNA reads to a genome; aligning long, high-error sequences from single molecule sequencers; aligning ancient or degraded DNA; comparing metagenomic DNA to a protein database.Our main argument will be that classic alignment techniques, with substitution score matrices and statistical models, can and should be used for modern data.  We will describe how to incorporate sequence quality data, and how to make the seed-and-extend method scale to large datasets by using "adaptive seeds".  Finally, we build on the classic statistical model to accurately align paired sequences, and sequences that come from disjoint genomic loci due to rearrangements.

The world of protein energy profiles in theory and practice

Dozent	Prof. Dr. Dirk Labudde & Florian Heinke, Bioinformatics, University of Applied Sciences Mittweida
Ort	ZBH, Raum 16
Zeit	28. Feb 2013 16:00

Recent studies have shown that transforming protein structures into sequences of energy values can aid in understanding the dynamics of protein structures. This includes (de-)stabilising effects caused by mutations and protein-environment interactions. However, with the plethora of available energy models, that differ greatly in detail, complexity, and computational cost, calculating and comparing so-called energy profiles is not trivial. In essence, the different properties of energy models can change the obtained energy profile data and make interpretations difficult.In this talk, the concepts of a straight-forward coarse-grained energy model are introduced, which allows the fast computation and comparison of energy profiles. On different applications and biological questions we show the efficiency and potential of this energy model for (multiple-) alignment tasks, measuring energy profile similarity, and large-scale database searching. 

Inferring genotype-phenotype relationships in the evolution of human influenza A viruses

Dozent	Prof. Dr. Alice McHardy, Heinrich Heine Universität Düsseldorf, Algorithmische Bioinformatik, Düsseldorf
Ort	ZBH, Raum 16
Zeit	31. Jan 2013 16:00

A joint analysis of genetic, phenotypic and epidemiological information for rapidly evolving RNA viruses can give novel insight into the linked process of their epidemiology and evolution. For instance, due to the rapid evolution of the surface proteins of human influenza A viruses, regular updates of the vaccine are necessary to ensure continued protection. I will discuss several computational techniques that we have recently developed for the study of viral ‘phylodynamics’ and show their application to seasonal influenza A viruses.  All methods can be applied for the study of other rapidly evolving genes that are important for adaptation.

Fast and accurate calculation of protein-protein interaction: contribution of surface and interface residues

Dozent	Dr. K. Anton Feenstra, IBIVU/Bioinformatics, Free University Amsterdam, Amsterdam, The Netherlands
Ort	ZBH, Raum 16
Zeit	10. Jan 2013 16:00

In many cellular processes physical interactions between proteins play crucial roles; to be functional, many proteins must organise into  complexes, and on the other hand, biologically non-functional complexes might give rise to a number of pathologies. Existing methods for predicting protein-protein interactions (PPIs) use statistical learning techniques or other bioinformatic approaches and appear to have limited accuracy. We therefore aim to establish a method based on physical principles using simplified molecular models for computational efficiency, while still maintaining the accuracy of calculating the interactions. We have established that we can accurately calculate potentials of mean force using the coarse-grained MARTINI forcefield where approximately four atoms are represented by a single sphere (or ‘super atom’), yielding a speedup several hundred-fold. As a second step, we can now perform extensive simulations to assess the contributions of different interface regions to the interaction. We performed random mutations on an equal number of residues in either the surface (non-interface), the interface core (buried in complex) or the interface rim (less exposed in complex) of two different protein complexes; a TCR-pMHC and an MP1-p15 scaffolding complex. Our results show that mutations on the interface core lead to less attractive interactions than mutations on the interface rim, while non-interface surface mutations hardly effect the PPI. Furthermore, when we mutate all residues in the interface core into the most likely substitutions (excluding self) according to BLOSUM, we see no measurable effect on the interaction potential, but if we mutate to the least likely we lose the attractive potential almost completely. Given the structure of the protein complex, we are now able to calculate an interaction potential between any two proteins in a matter of hours. I will discuss how we are planning to further extend this approach.

Just How Do Enzymes Immobilize?

Dozent	Dr. Lutz Hilterhaus, Institute of Technical Biocatalysis, Hamburg University of Technology, Hamburg
Ort	ZBH, Raum 16
Zeit	15. Dec 2012 16:00

The question “How to immobilize enzymes?” has been answered by a broad range of publications on immobilization and stabilization techniques in the last decades. However, what happens during this process is not known in detail. Thus, a deeper understanding of the immobilization procedure at a molecular level will help to optimize existing immobilization protocols. Additionally, the understanding of protein unfolding will allow a knowledge-based immobilization, enhancing activity yield as well as the stabilizing effect. Therefore, the interaction of molecular modeling, molecular biology and biochemistry can lead to a more rational approach for the stabilization of enzymes by immobilization.This lecture will discuss the requirements of enzymatic processes and recent developments in structure-guided approaches from the perspective of a bioprocess engineer. Further emphasis lies on the site specific enzyme immobilization and its potential beneficial effects on activity and stability of enzymes.

Mixture models for large-scale metagenome analysis

Dozent	Dr. Peter Meinicke, University of Goettingen, Institute of Microbiology and Genetics, Dept. of Bioinformatics, Goettingen
Ort	ZBH, Raum 16
Zeit	06. Dec 2012 16:00

Metagenomics opens a new dimension in the exploration of the microbial world. For the first time, it is becoming possible to take an unbiased look at the structure of complex microbial communities from all kinds of ecosystems. The analysis of the metagenome composition on the basis of large collections of anonymous sequence fragments is currently challenging bioinformatics. Mixture models provide a natural representation of metagenomes, well-suitable for estimating taxon abundances with a high computational efficiency. However, the possible accuracy and robustness of the approach largely depends on the sequence features used for statistical modeling. While our first mixture approach was based on oligonucleotide frequencies we recently investigated the possibilities of a feature space based on protein domain frequencies. Our results indicate that the protein-based representation yields more reliable estimates than the oligonucleotide-based model and at the same time enables a broader taxonomic range that includes all domains of life and viruses.

Gene Prediction with a Conditional Random Field Maximizing Expected Accuracy

Dozent	Prof. Dr. Mario Stanke, Universität Greifswald, Institut für Mathematik und Informatik, Greifswald
Ort	ZBH, Raum 16
Zeit	22. Nov 2012 16:00

Many bioinformatics tasks such as sequence alignments, prediction of RNA secondary structure and gene prediction are solved by defining a probability or score for every element of the solution space and by then devising an algorithm that searches - exactly or approximately - for an element with highest probability or score. In the case of gene finding, we construct a probability distribution over all possible gene structures through a conditional random field and find a most likely gene structure with the Viterbi algorithm. However, we recently found that on the same model an algorithm that maximizes another criterion - the expected accuracy - gives better performance than the Viterbi algorithm.

Pharmacophore Modeling in Early Drug Discovery

Dozent	Dr. Karl-Heinz Baringhaus, Sanofi-Aventis Deutschland GmbH, Frankfurt am Main
Ort	ZBH, Raum 16
Zeit	15. Nov 2012 16:00

Pharmacophores play an important role in many fields of modern Drug Design. In lead generation, pharmacophore-based virtual screening and rescaffolding turned out to be powerful, while in lead optimization, pharmacophores are often successfully coupled with sophisticated 3D-QSAR methods (e.g. CoMFA). This presentation captures several aspects of successful pharmacophore modeling. This includes training and test set selection, conformational analysis, pharmacophore building and validation as well as several examples from an industrial perspective. Pharmacophore-based in silico selectivity assessment of compounds against anti-targets (e.g. hERG, CYPs) as well as the application of pharmacophores in polypharmacology and in silico profiling of compounds will be briefly addressed.

Clinically relevant xenograft models of cancer metastasis

Dozent	Prof. Dr. med. Udo Schumacher, Universitätsklinikum Hamburg-Eppendorf, Institut für Anatomie und Experimentelle Morphologie
Ort	ZBH, Raum 16
Zeit	08. Nov 2012 16:00

Skalierbare Analyse und 3D-Präsentation komplexer Daten

Dozent	Prof. Dr.-Ing. Stephan Olbrich, Regionales Rechenzentrum der Universität Hamburg
Ort	ZBH, Raum 16
Zeit	25. Oct 2012 16:00

A general theory of protein evolution

Dozent	Prof. Dr. Erich Bornberg-Bauer, Universität Münster, Institut für Evolution und Biodiversität, Münster
Ort	ZBH, Raum 16
Zeit	15. Oct 2012 16:00

A general theory of protein evolution (Aka: Escape from Adaptive Conflict follows from weak functional trade-offs and mutational robustness.) Tobias Sikose, Hue-Sun Chan and Erich Bornberg-Bauer A fundamental question in molecular evolution is how proteins can adapt to new functions while being conserved for an existing function at the same time. Several theoretical models have been put forward to explain this apparent paradox. The most popular models include Neofunctionalization (NEOF), Subfunctionalization (SUBF) by degenerative mutations (DDC), and dosage models. All of these models focus on adaptation after gene duplication. A newly proposed model named “Escape from Adaptive Conflict” (EAC) includes adaptive processes before and after gene duplication that lead to multifunctional proteins, and divergence (SUBF). Support for the importance of multifunctionality for the evolution of new protein functions comes from two experimental observations. First, many enzymes have highly evolvable promiscuous side activities. Second, different structural states of the same protein can be associated with different functions. How these observations may be related to the EAC model, under which conditions EAC is possible, and how the different models relate to each other is still unclear. Here, we present a theoretical framework that uses biophysical principles to infer the roles of functional promiscuity, gene dosage, gene duplication, point mutations, and selection pressures in the evolution of proteins. We find that selection pressures can determine whether NEOF or SUBF is the more likely evolutionary process. Multifunctional proteins, arising during EAC evolution, allow rapid adaptation independent of gene duplication. This becomes a crucial advantage when gene duplications are rare. Finally, we propose that an increase in mutational robustness, not necessarily functional optimization, can be the sole driving force behind SUBF. We illustrate the role of bridge sequences (intersecting two neutral nets and folding into both structures with almost equal probability) for speeding up transitions in both. model proteins and stability calculations using ROSETTA and FoldX. ---------------------------------------------------------- Sikosek, E Bornberg-Bauer and HS Chan; Evolutionary dynamics on protein bi-stability landscapes have the potential to resolve adaptive conflicts; PLoS Computational Biology; in press T Sikosek, HS Chan and Erich Bornberg-Bauer; Escape from Adaptive Conflict follows from weak functional trade-offs and mutational robustness; Proc Natl Acad Sci, USA; in press. Chen T, Vernazobres D, Yomo T, Bornberg-Bauer E, Chan HS. Evolvability and single-genotype fluctuation in phenotypic properties: a simple heteropolymer model. Biophys J. 2010 98(11):2487-96. Wroe R, Chan HS, Bornberg-Bauer E. A structural model of latent evolutionary potentials underlying neutral networks in proteins. HFSP J. 2007 1(1):79-87. Epub 2007 May 21. Wroe R, Bornberg-Bauer E, Chan HS. Comparing folding codes in simple heteropolymer models of protein evolutionary landscape: robustness of the superfunnel paradigm. Biophys J. 2005 88(1):118-31. Epub 2004 Oct 22. Cui Y, Wong WH, Bornberg-Bauer E, Chan HS. Recombinatoric exploration of novel folded structures: a heteropolymer-based model of protein evolutionary landscapes. Proc Natl Acad Sci U S A. 2002 99(2):809-14. Bornberg-Bauer E, Chan HS. Modeling evolutionary landscapes: mutational stability, topology, and superfunnels in sequence space. Proc Natl Acad Sci U S A. 1999 96(19):10689-94.

Körber-Lecture: Die Entschlüsselung des Proteoms durch Massenspektrometrie

Dozent	Körber-Preisträger 2012: Prof. Dr. Matthias Mann, Max-Plack-Institut für Biochemie, Martinsried
Ort	Fachbereich Chemie, Martin-Luther-King-PLatz 6, Hörsaal B
Zeit	06. Sep 2012 16:00

Körber-Lecture mit dem Körber-Preisträger 2012 Prof. Dr. Matthias Mann

Der Physiker und Bioinformatiker Matthias Mann hat ein hochpräzises, massenspektrometrisches Verfahren entwickelt, mit dem sämtliche Proteine einer Zelle auf einen Schlag analysiert werden können. Bisher konnten sich Forscher traditionell nur mit der Wirkung eines oder einiger weniger Proteine befassen.

2010 gelang es Mann und seinem Team sogar erstmals, das komplette Proteom – die Gesamtheit aller Eiweiße eines Lebewesens – der Hefezelle (mehr als 4.000 Proteine) zu entschlüsseln. Inzwischen arbeiten Forscher in aller Welt fieberhaft daran, das komplette menschliche Proteom zu analysieren. Dieses »Human Proteome Project« verspricht einen ähnlich spannenden Wettlauf wie um die Jahrtausendwende das »Human Genome Project«. Während die Gene passiv in den Zellkernen schlummern und zeitlebens meist unverändert bleiben, ändern sich die von der Zellmaschinerie aus der DNA erzeugten Proteine ständig. Das Proteom liefert daher immer nur eine Art »Momentaufnahme«. Doch gerade dies eröffnet Forschern völlig neue Perspektiven: So verrät etwa der Vergleich des Proteoms gesunder und kranker Zellen, welche Proteine die Entstehung von Krankheiten wie Krebs oder Diabetes begünstigen oder hemmen. Für seine wegweisenden Arbeiten zur Entschlüsselung des Proteoms erhält Prof. Dr. Matthias Mann vom Max-Plack-Institut für Biochemie in Martinsried den mit 750.000 Euro dotierten Körber-Preis für die Europäische Wissenschaft 2012. 

Seine Forschungsergebnisse präsentiert der Preisträger 

*am Donnerstag, den 6. September 2012, um 16 Uhr,* 

*im Fachbereich Chemie der Universität Hamburg,* 

*Martin-Luther-King-Platz 6, Hörsaal B*** 

in einer akademischen Vorlesung mit dem Titel »Decoding the Proteome by Mass Spectrometry– Die Entschlüsselung des Proteoms durch Massenspektrometrie« dem interessierten Fachpublikum. Zu dieser Veranstaltung laden wir Sie im Namen des Präsidiums der Universität Hamburg, des Dekanats der Fakultät für Mathematik, Informatik und Naturwissenschaften der Universität Hamburg und des Vorstands der 
Körber-Stiftung ganz herzlich ein. Der Vortrag findet in englischer Sprache 
statt und dauert voraussichtlich 45 Minuten. Im Anschluss an die Veranstaltung laden wir zum Sektempfang.  Gerne können Sie diese Einladung auch an Kolleginnen und Kollegen, den wissenschaftlichen Nachwuchs und Studierende weiterreichen.

Gerne können Sie diese Einladung auch an Kolleginnen und Kollegen, den wissenschaftlichen Nachwuchs und Studierende weiterreichen. Über eine Anmeldung per E-Mail unter wissenschaft(at)koerber-stiftung.de würden wir uns sehr freuen. 

Der *Körber-Preis *zeichnet jährlich herausragende und in Europa tätige Wissenschaftler für deren zukunftsträchtige Forschungsarbeiten aus. Prämiert werden exzellente und innovative Forschungsansätze mit hohem Anwendungspotenzial auf dem Weg zur Weltgeltung. Ausgewählt wird der jeweilige Preisträger von einem international besetzten Kuratorium unter dem Vorsitz von Prof. Dr. Peter Gruss, Präsident der Max-Planck-Gesellschaft. Der Preis wird am 7. September 2012 in Anwesenheit des Ersten Bürgermeisters der Freien und Hansestadt Hamburg im Hamburger Rathaus verliehen.

Functional modules in biochemical systems

Dozent	Prof. Dr. Ina Koch, Goethe-University, Molecular Bioinformatics, Frankfurt am Main
Ort	ZBH, Raum 16
Zeit	15. Jul 2012 16:00

The automatic computation of biological, functional modules becomes more and more important in the analysis of big and complex biochemical networks. In the last years, several unique definitions of functional modules have been developed. We define functional modules on the basis of Petri nets. The straightforward definition of Petri nets and their intuitive visualization is perfectly suitable for applications in interdisciplinary fields, such as computational systems biology. Additionally, several simulation techniques provide the possibility to study system’s dynamics, even if kinetic parameters are incomplete. Moreover, Petri net formalism provides many useful analysis techniques that can be easily adapted to biochemical systems. For an exhaustive overview see [1]. After a short introduction to Petri nets, using biological examples, the talk will mainly address recent developments for module detection. We start with differentiation between steady-state based modeling and non steady-state based methods. The main part will consider modularization and network reduction techniques, such as invariant properties, MCT-sets, T-clusters, Mauritius Maps, which have been developed in biological context to handle the huge complexity of biochemical systems. We also discuss the applicability of these concepts to signal transduction systems, gene regulatory networks, and combinations of different networks. [1] Koch, Ina; Reisig, Wolfgang; Schreiber, Falk (Eds.) Modeling in Systems Biology: The Petri Net Approach, Series: Computational Biology, Vol. 16, 1st Edition, 2011, 364 p., ISBN: 978-1-84996-473-9

Design of PresCont, a classifier for protein-protein interfaces and biochemical validation of interface predictions

Dozent	PD Dr. Rainer Merkl, Institut für Biophysik und Physikalische Biochemie, Universität Regensburg
Ort	ZBH, Raum 16
Zeit	12. Jul 2012 16:00

We have implemented the in silico tool PresCont, which assigns to each surface residue of a protein a probability for belonging to a protein-protein interface (PPI). PresCont is based on a two-class support vector machine and requires as input a multiple sequence alignment of homologous sequences and a pdb-file representing the 3D structure of the protein under study. This input data is the basis to deduce not more than four parameters utilized for classification. By analyzing large datasets, we have shown that the classification performance of PresCont compares favorably with other state-of-the-art classifiers.

PcrB (heptaprenlyglyceryl phosphate synthase) is a bacterial enzyme, which forms a dimer and whose PPI cannot be deduced unambiguously from the crystal structure. Using PresCont, the two most likely PPIs were identified and these predictions were assessed i) by amino acid substitutions that led to monomerization and ii) the introduction of an unnatural amino acid that allowed crosslinking of two protomers. The search for sites tolerating these mutations and the choice of residues aimed at monomerization were supported by computations based on the ROSETTA software suite. Thus, by combining in silico and experimental methods, we identified the dimerization interface, which is not the one specified in PDB.

Computational high-throughput screening of advanced battery electrolyte solvents and additives

Dozent	Prof. Dr. Martin Korth, Institut für Theoretische Chemie, Universität Ulm, Ulm
Ort	ZBH, Raum 16
Zeit	28. Jun 2012 16:00

Increasing global energy demand, rising carbon dioxide emission, finite fossil fuel supplies and the expectation of soaring fuel prices have brought about the urgent need for renewable energy sources. Not only the harvesting but also the storage of energy from these sources presents us with daunting scientific and technological problems. One important aspect is the development of advanced batteries as essential prerequisite for future e-mobility concepts. The last years have seen great interest in this field also from theoretically working scientists, but especially the computer-assisted development of advanced electrolyte systems is still in it's infancy. An overview of existing work in this direction is given and supplemented with a new evaluation study

The cow rumen metagenome - discovery of biomass-degrading genes and genomes

Dozent	Dr. Alexander Sczyrba, Computational Metagenomics, Bielefeld University Bioinformatics Service, Center for Biotechnology, Bielefeld
Ort	ZBH, Raum 16
Zeit	14. Jun 2012 16:00

The paucity of enzymes able to degrade lignocellulosic biomass efficiently represents a major bottleneck in the industrial production of biofuels. To identify and characterize biomass degrading genes and genomes of the cow rumen, we sequenced and analyzed more than 250 Gbp of metagenomic DNA isolated directly from microbes adherent to rumen incubated switchgrass. Despite the complexity of the microbial community, ultra-deep sequencing technology enabled the de novo assembly of numerous draft genomes from uncultured biomass degrading microbes. The authenticity of predicted genes and individual assembled genomes was validated by complementary methods, including single cell sequencing. Enzymatic activity assays verified the cellulolytic activity of candidate cellulases, emphasizing the value of deep metagenomic sequencing for targeted enzyme discovery from uncultured microbes.

Magnesium as Implant Material

Dozent	Prof. Dr. Regine Willumeit, Helmholtz-Zentrum Geesthacht, Zentrum für Material- und Küstenforschung, Geesthacht
Ort	ZBH, Raum 16
Zeit	24. May 2012 16:00

Magnesium as degradable metal has a large potential for biomedical applications. However, the mechanism of degradation is not yet clarified in spite of available in vivo data and clinical trials. Moreover, the site of possible implant application is of huge relevance for the expected degradation time due to factors like tissue fluid flow rate or tissue hydrogen dissolution coefficient. Additionally, the sensitivity of tissue cells to the magnesium specific degradation environment may considerably vary. In this presentation an overview about the challenges of magnesium implant design will be given.

Accuracy matters: 3D alignment in pharmacophore perception and virtual screening

Dozent	Prof. Dr. Gerhard Wolber, Freie Universität Berlin, Pharmaceutical Chemistry, Berlin
Ort	ZBH, Raum 16
Zeit	10. May 2012 16:00

Virtual screening using three-dimensional arrangements of chemical features (3D pharmacophores) has become a highly relevant method for virtual hit identification and lead optimization. Although frequently used, considerable differences exist in the interpretation of chemical features and the implementation of their corresponding 3D overlay algorithms. Based on a pattern-matching 3D alignment algorithm, we have developed a work-flow for high-performance virtual database screening that delivers a higher degree of geometric accuracy than previously published or commercially available approaches. The impact of increased geometric accuracy on virtual screening with respect to selectivity and sensitivity

Nucleotide prodrugs as chemical trojan horses and a potentially new target for anti-HIV chemotherapy

Dozent	Prof. Dr. Chris Meier, Organic Chemistry, Dept. of Chemistry, Faculty of Science, University of Hamburg
Ort	ZBH, Raum 16
Zeit	03. May 2012 16:00

Lipophilic prodrugs of the highly hydrophobic phosphorylated nucleosides may be a promising alternative to improve the biological activity of nucleoside analogs in antiviral chemotherapy. Two of the most successful pronucleotide-systems are the cycloSal-pronucleotides and the nucleoside arylphosphoramidates. As the first approach is based on a selective chemical hydrolysis, the second technology requires an enzyme-mediated activation. Due to their synthesis pathways, derivatives belonging to these two classes of pronucleotides were always obtained as 1:1 mixtures of diastereomers with respect to the configuration at the phosphorus center. The mixtures of diastereomers could be separated only in limited cases. However, the individual diastereomers have different antiviral activity (see below), toxicity and hydrolysis stabilities. For this reason we have developed chemical pathways to synthesize isomerically pure pronucleotides using a convergent synthesis combined with the chiral auxiliary approach.Moreover, we have introduced the first successful delivery system that releases nucleoside diphosphates intracellularly (DiPPro-approach). The basic idea, chemistry and prove-of-concept will be summarized.Finally, in a collaborative approach involving different research groups we focused on the identification of a new cellular target that may be used for the development of antiviral compounds. In our case, the cellular factor eIF-5A involved in the export of viral RNA from the nucleus into the cytoplasm was addressed as a potential target. As the starting point, two known inhibitors of this factor were modified by computational chemistry and “wet” chemistry. In antiviral assays several of our compounds were found to be highly active against HIV-1 replication. Thus, the concept may be suitable to combat upcoming drug resistance.

Computational Chemistry for Crop Protection

Dozent	Prof. Dr. Klaus-Jürgen Schleifer, Vice President, Computational Chemistry & Biology, BASF SE, Ludwigshafen
Ort	ZBH, Raum 16
Zeit	26. Apr 2012 16:00

BASF’s Crop Protection division applies different strategies to discover new active ingredients. Via the classical, so-called screening, chemicals are directly tested on harmful organisms. In parallel, the pharma-like mechanism based approach focuses on the isolated target proteins in vitro. The presentation will give a brief overview of these complementing techniques and emphasis on computational methods supporting the research process. Based on different input information, examples will be presented demonstrating integration and value of theoretical approaches for lead identification and optimization.

Making light work of malarial structural biology

Dozent	Dr. Matthew Groves, EMBL Hamburg
Ort	ZBH, Raum 16
Zeit	12. Apr 2012 16:00

>Malaria remains a pressing concern for global human health and the lack of an effective vaccine, as well as the spread of multi-drug resistant strains, 
necessitates the development of novel therapies (http://www.who.int/topics/malaria/en/). This presentation will detail developments in sample characterisation and biophysics that have been used to aid the structural study of metabolic pathways in P. falciparum (the causative agent of the most lethal form of malaria). The plasmodial aspartate aminotransferase will be used as a case example.

Bioconjugates to Specifically Render Inhibitors Water-Soluble

Dozent	Dr. Anna K. H. Hirsch, Stratingh Institute for Chemistry, University of Groningen, The Netherlands
Ort	ZBH, Raum 16
Zeit	02. Feb 2012 16:00

Water solubility is a recurring problem in medicinal chemistry. In drug discovery, an increase in affinity is often accompanied by a concomitant increase in hydrophobicity. Thus, the resulting inhibitors may suffer from low water solubility, precluding in vitro studies. A novel solubilisation strategy is presented: a polymer–peptide conjugate is designed, in which the polymer moiety affords water solubility, while the peptide sequence interacts specifically with the insoluble inhibitor,[1] leading to a soluble inhibitor–bioconjugate complex.[2] This concept was successfully applied to render small-molecule inhibitors of the kinase IspE,[3] an attractive antimalarial target, water-soluble. REFERENCES [1]. Hartmann, L., Börner, H. G., Adv. Mater., 21, 3425–3431, 2009. [2] . Hirsch, A. K. H., Diederich, F., Börner, H. G., Antonietti, M., Soft Matter, 6, 88–91, 2010. [3]. Hirsch, A. K. H., Alphey, M. S., Lauw, S., Seet, M., Barandun, L., Eisenreich, W., Rohdich, F., Hunter, W. N., Bacher, A., Diederich, F., Org. Biomol. Chem., 6, 2719–2730, 2008.

Resistance distances are meaningless in large graphs

Dozent	Dr. Ulrike von Luxburg, MPI for Intelligent Systems, Machine Learning Theory, Tübingen & University of Hamburg, Computer Science
Ort	ZBH, Raum 16
Zeit	26. Jan 2012 16:00

We study distance functions between vertices of a graph.  The resistance distance (also known as commute distance) between two vertices is defined as the expected time a random walk needs to travel from one vertex to the other and back. We study the behavior of the resistance distance as the size of the graph increases. We prove that, on many kinds of graphs, it converges to an expression that does not take into account the structure of the graph and is completely meaningless as a distance function on the graph. Consequently, the use of the raw commute distance is strongly discouraged for large graphs and in high dimensions. An alternative can be provided by the family of p-resistances (p >= 1).  We prove nice structural results about the family of p-resistances and show for which choices of p it leads to reasonable distance functions.

Computational Chemistry in Pharmaceutical Research – Achievements, Pitfalls and Challenges

Dozent	Dr. Herbert Köppen, Boehringer Ingelheim Pharma GmbH & Co. KG, Lead Identification and Optimization Support, Biberach
Ort	ZBH, Raum 16
Zeit	12. Jan 2012 16:00

Computational chemistry is an integral part of pharmaceutical research in all larger companies. Based on the author’s experience at Boehringer Ingelheim the presentation will illustrate the role and the major tasks of computational chemistry in pharmaceutical research. Progress in experimental methods, in particular in high-throughput technologies and in protein crystallography, has had a major impact on the focus and the integration of computational chemistry in the research process. It will be shown that an industrial setting with many parallel research projects and considerable synthesis and biological testing capacities creates special conditions for computational chemistry with impact on the required tools and methods. Amongst other topics the synergy of virtual screening and combinatorial chemistry and the application of the recently published GPCR structures in project work will be discussed.

It's not the size that matters but what you do with it. DNA sequence principles in the era of genome-wide datasets

Dozent	Dr. Marc Rehmsmeier, University of Bergen, Computational Biology, Norway
Ort	ZBH, Raum 16
Zeit	10. Jan 2012 16:00

Genome-wide data is being generated at an alarming rate. But what does it all mean? In this talk I will ponder on how DNA sequence principles can make sense of this evergrowing mass.

Compressed Suffix Trees in Practice

Dozent	Dr. Simon Gog
Ort	ZBH, Raum 16
Zeit	15. Dec 2011 16:00

Compressed Suffix Trees (CSTs) are space-efficient full-text index data structures which provide more functionality than their uncompressed counterparts while at the same time using only a fraction of the space. In addition, most of the operations on a CST can be performed in the same time complexity as on the uncompressed data structure.

An interesting question is, whether CSTs can also keep up with uncompressed suffix trees in practice. In this talk I will answer this question and review different existing implementations. Last, I will present my C++ library sdsl (succinct data structures library), which offers configurable implementations of different CST proposals.

Bioinformatics Software Tools for DNA Sequence Analysis and Genome Annotation

Dozent	Dr. Alexander Goesmann, University of Bielefeld, Bioinformatics Resource Facility, Bielefeld
Ort	ZBH, Raum 16
Zeit	08. Dec 2011 16:00

Today, the analysis of DNA sequences is pushed forward by high-throughput DNA sequencing instruments. The handling and evaluation of the huge amounts of sequence data currently pose a major challenge for bioinformaticians as well as biologists. Within this talk, I will present the following software tools developed at the Bielefeld Center for Biotechnology to support the automated data analysis in various fields of genome research:

i) GenDB - A genome annotation system that is currently being extended and adapted for eukaryotic genomes

ii) CARMEN - A tool for metabolic pathway reconstruction

iii) EDGAR - A platform for comparative genomics

iv) Conveyor - A workflow engine for bioinformatics

v) SARUMAN - Our latest developments to utilize GPU programming for ultrafast short read mapping

vi) VAMP - A short read browser for resequencing and RNAseq data analysis

Identification of causative genes for monogenic disorders: From established technologies to next-generation sequencing

Dozent	Prof. Dr. Kerstin Kutsche, Institute of Human Genetics, University Medical Center Hamburg-Eppendorf, Hamburg
Ort	ZBH, Raum 16
Zeit	01. Dec 2011 16:00

Proteins and PetaFlops

Dozent	Dr. Olav Zimmermann, Simulation Laboratory Biology Juelich Supercomputing Centre, Juelich
Ort	ZBH, Raum 16
Zeit	24. Nov 2011 16:00

The Simulation Laboratory Biology in Juelich supports the bioscience research community in porting life science applications to the Juelich supercomputers and to future supercomputing architectures. In addition to its support activities the Simulation Laboratory has its own research agenda with a focus on structural bioinformatics. We are especially interested in protein folding and peptide aggregation. Our simulation approach is based on Replica Exchange Monte Carlo and we are actively (co)developing two open source Monte Carlo simulation codes: SMMP and ProFaSi.In a second line of research we have developed several tools for local structure prediction, based on Support Vector Machines, a machine learning approach.The talk will give an overview about the achievements as well as the current limitations of our methods, outline how physical simulations and machine learning based approaches can be combined for structure prediction,and discuss the role of supercomputing in structural bioinformatics.

Marine Bioinformatics

Dozent	Prof. Dr. Frank Oliver Glöckner, Max Planck Institute for Marine Microbiology, Jacobs University Bremen gGmbH
Ort	ZBH, Raum 16
Zeit	17. Nov 2011 16:00

Recent developments in ‘Omics’ technologies as well as improvements in sampling and laboratory equipment have opened a new dimension in research biodiversity and ecosystem functioning. Leading institutes of marine sciences started to adopt these new technologies. Reflecting the size and complexity of the marine ecosystem, an unprecedented amount of environmental data, biodiversity data, as well as functional and organism-specific data are already produced or envisaged. This ever growing deluge of electronically available data needs to be processed, integrated and visualized for in depth analysis by domain experts with respect to a better understanding of evolution, niche adaptations and ecosystems functioning. The talk will provide an overview about the bioinformatic challenges in phylogenetic and functional diversity analysis for (marine) environmental research. It will further introduce ELIXIR – a potential European Infrastructure for Biological Information.

Exploiting Many-Core Architectures in Molecular Modeling

Dozent	Dr. Oliver Korb,The Cambridge Crystallographic Data Centre, Cambridge, UK
Ort	ZBH, Raum 16
Zeit	10. Nov 2011 16:00

In recent years the introduction of many-core architectures enabled the scientific community to tackle computationally intensive problems on desktop machines which were previously only tractable using large clusters. While many-core processors and graphics processing units offer enormous computational power, the redesign of existing sequential algorithms as well as the design of novel algorithms exploiting the massively parallel nature of these devices is a challenging task. In this talk an overview of existing many-core applications in the field of molecular modeling will be presented and common pitfalls when designing and comparing those approaches will be highlighted. The acceleration of molecular docking applications using graphics processing units will serve as an example. 

(1)         Stone, J. E.; Hardy, D. J.; Ufimtsev, I. S.; Schulten, K. GPU-accelerated molecular modeling coming of age. Journal of Molecular Graphics and Modelling2010, 29, 116-125.

(2)          Korb, O.; Stuetzle, T.; Exner, T. E. Accelerating Molecular Docking Calculations Using Graphics Processing Units. Journal of Chemical Information and Modeling 2011, 51, 865-876.

Design and synthesis of novel pyrazinones as selective PDGFR inhibitors

Dozent	Prof. Dr. Christian Peifer, University of Kiel, Pharmaceutical Chemistry, Kiel
Ort	ZBH, Raum 16
Zeit	27. Oct 2011 16:00

In this study we report on the discovery of 3,5-aryl-substituted pyrazin-2(1H)-ones as a lead for the development of potent and selective ATP competitive PDGFRb inhibitors. In order to enhance the potency and selectivity of pyrazin-2(1H)-ones for PDGFRb, molecular modelling studies using a structure based drug design approach were performed, yielding an optimization strategy for the inhibitors. A novel flexible synthetic route via microwave mediated ring closure towards asymmetric 3,5-substituted pyrazin-2(1H)-ones was developed. SAR for a series of compounds and highly potent PDGFR inhibitors are presented.

Introducing Theoretical Chemistry at the ZBH

Dozent	JProf. Dr. Tobias Schwabe, University of Hamburg, ZBH, Theoretical and Computational Chemistry
Ort	ZBH, Raum 16
Zeit	20. Oct 2011 16:00

With this talk, I introduce the new junior research group for Theoretical and Computational Chemistry. A brief overview of the general topic, quantum chemistry, is given. Further, I discuss my current research interests. One main project is the improvement of double-hybrid density functionals. These methods have been proven to be very accurate and efficient approaches in quantum chemistry. Some of the recent developments are shown. The other main project is the description of solvation effects within quantum chemistry. Among the most promising approaches to this problem are hybrid methods between quantum mechanics and classical mechanics (QM/MM). Recent advances with the polarizable embedding method are demonstrated.

The evolution of chemotactic behaviour in digital organisms

Dozent	Prof. Richard Goldstein, MRC, National Institute for Medical Research, Mathematical Biology, London, UK
Ort	Raum 16, ZBH
Zeit	23. Aug 2011 15:00

Incorporation of protein flexibility and dynamics into protein-ligand docking

Dozent	Prof. Markus A. Lill, Purdue University, Dept. of Medicinal Chemistry and Molecular Pharmacology, West Lafayette, IN, USA
Ort	ZBH, Raum 16
Zeit	12. Jul 2011 16:00

Molecular recognition between receptors and ligands through non-covalent association plays a fundamental role in virtually all biochemical processes in living organisms. Several computational concepts have been devised to study protein-ligand binding.  These techniques are routinely used in academia and industry for identifying and optimizing potential drug candidates. While those methods have been widely used to attain a qualitative understanding of ligand binding to proteins, a current challenge is to quantify their interaction in a reasonable amount of time. One major issue is that the protein in reality can adapt its shape and properties upon each individual ligand binding to it (induced protein fit). In this context, I will present our development of new concepts incorporating protein flexibility and protein dynamics to identify binding modes for ligands of biomedical interest (Software: Limoc and CorLps) and to quantify their interaction with the target protein (Software: Raptor). I also will highlight our efforts towards the development of a user-friendly open-source computer-aided drug discovery platform based on PyMOL. Several third-party and in-house molecular modeling programs are combined in a single platform. This concept originated from the development of a new course in Computer-Aided Drug Design allowing the students to focus on science rather than the usage of software.

Biomolecules in a structured solvent - a nonlocal electrostatics treatment

Dozent	Prof. Dr. Andreas Hildebrandt, Johannes-Gutenberg-Universität Mainz
Ort	ZBH, Raum 16
Zeit	23. Jun 2011 16:00

The successful development of new drugs is one of science's most difficult tasks today. Even under optimal circumstances, it takes years to finish, has a very low probability of success, and is immensely expensive. In this talk, I want to highlight two key aspects of how computer science can help to make drug design cheaper, faster, and more reliable: the estimation of interaction energies for automated approaches on the one hand and the manual inspection and manipulation of virtual representations of molecular systems by domain experts on the other. In the first part, I will focus on the accurate computation of electrostatic effects which play an important role in the energetics of biomolecules. Many of those effects are dominated by the shielding effect of the water that is always present in biochemical reactions. Therefore, a highly accurate computation of electrostatic potentials of biomolecules in water is an important precursor for many applications in bioinformatics, like the mentioned computer aided development of inhibitors for disease related enzymes. In the literature, nonlocal extensions of classical macroscopic electrostatics have been proposed to capture the effects of the water on the electric potential. We propose a reformulation of the resulting equations, which we can be addressed numerically. The method has been shown to yield very accurate results on small systems like mono- or polyatomic ions and initial results on selected proteins are highly promising. In the second part, I will introduce our recent work in molecular visualization which allows interactive real-time ray tracing of biomolecular systems. The use of such ray tracing techniques can help to greatly improve visual quality and - more importantly - depth and geometry perception by the viewer.

Docking to identify GPCR ligands and predict enzyme function

Dozent	Dr. Peter Kolb, Philipps-Universität Marburg, Institut für Pharmazeutische Chemie, Marburg
Ort	ZBH, Raum 16
Zeit	19. May 2011 16:00

Computer-aided drug design for G-protein coupled receptors (GPCRs) has so far mostly relied on ligand-based methods and homology models due to the lack of x-ray structures. The recently solved structures of the ?­₁-, ?₂-adrenergic and the A_2A-adenosine receptors now allow the use of structure-based methods. The impact of these structures is two-fold: one, they can directly be utilized in structure-based screenings and two, they are likely more relevant templates to derive homology models. I will present examples for both avenues and highlight some lessons that we believe to be of a general nature with respect to ligand design for GPCRs. Docking has also successfully been applied to the prediction of enzyme function. Candidate substrate molecules are docked to an enzyme active site, and from the analysis of the chemotypes of the top-ranking molecules, a possible reaction is proposed. This will be exemplified with several case studies of enzymes in the amidohydrolase superfamily.

Actin-based motility of the malaria parasite - from structures to understanding function

Dozent	Prof. Inari Kursula, Helmholtz Centre for Infection Research and University of Hamburg / University of Oulu, Department of Biochemistry, Oulu, Finland
Ort	ZBH, Raum 16
Zeit	05. May 2011 16:00

A spatial and temporal map of C. elegans gene expression

Dozent	Dr. Stefan Henz, MPI for Developmental Biology, Dept. of Molecular Biology, Tübingen
Ort	ZBH, Raum 16
Zeit	28. Apr 2011 16:00

C.elegans is a completely sequenced model organism. In our analysis we linked gene expression of this organism to developmental stages and cell types by making use of tiling arrays. Machine-learning-analysis detected transcripts corresponding to established gene models, but novel transcriptionally active regions (TARs) in noncoding domains were also detected. Surprisingly, our results show that about 75% of the expressed transcripts and genes are differentially expressed. We used self-organizing maps to define groups of co-regulated transcripts and applied regulatory element analysis to identify known regulatory elements as well as novel ones.

Defining the fold space of membrane proteins

Dozent	Prof. Dr. Dmitrij Frishman, Technische Universität München, Wissenschaftszentrum Weihenstephan, Freising
Ort	ZBH, Raum 16
Zeit	15. Apr 2011 16:00

Recent progress in structure determination techniques has led to a significant growth in the number of known membrane protein structures, and the first structural genomics projects focusing on membrane proteins have been initiated, warranting an investigation of appropriate bioinformatics strategies for optimal structural target selection for these molecules. What determines a membrane protein fold? How many membrane structures  need to be solved to provide sufficient structural coverage of the membrane protein  sequence space? In my talk I will describe the CAMPS database (Computational Analysis of the Membrane Protein Space) that automatically classifies a-helical membrane proteins into fold classes. I will also present our latest results on predicting interacting helices in membrane proteins and deriving helix connectivity diagrams that represent a convenient level of abstraction for comparing predicted membrane protein structures. Finally, I will review the difficulties in classifying experimentally determined three-dimensional structures of membrane proteins, focusing on difference between the SCOP and CATH databases.

Taming the beasts: A poor mans approach to unlock cereal genomes

Dozent	Dr. Klaus Mayer, Helmholtz Center Munich, German Research Center for Environmental Health, Neuherberg
Ort	ZBH, Raum 16
Zeit	03. Feb 2011 16:00

Barley is an important food and feed crop. It is closely related to wheat and its 5,2 gigabase genome is archaetypical for the Triticeae. Genome complexity and a repeat content of >80% are major obstacles to unlocking the genome by systematic sequencing. We use a powerful approach of chromosome sorting, low pass shotgun sequencing, gene-array hybridisation and bioinformatic integration of model grass genome synteny information to obtain detailed insight into the organization of the barley genome. The gene content of barley is estimated to comprise ~32,000 genes and roughly 22000 genes were positioned along the barley genome. Detailed comparisons to rice, sorghum and Brachypodium reveal vast differences in local conservation with regions that have probably undergone rearrangements and deletions in barley, rice, sorghum or Brachypodium. The resolution also allows one to undertake genome scale comparisons to wheat and to analyse for genes that are under selection in the respective species. We present a highly ordered and information-enriched scaffold of the barley genome that reaches - at a fraction of the cost - a level of information density and resolution for a representative diploid Triticeae genome.

Building an in silicoADMET platform for hit-to-lead and lead optimization

Dozent	Prof. Dr. Alexander Hillisch, Bayer Schering Pharma AG, Wuppertal
Ort	ZBH, Raum 16
Zeit	20. Jan 2011 16:00

During the past years an in silico ADMET IT platform has been built up at Bayer Schering Pharma (BSP). The platform consists of various physicochemical and ADMET prediction tools and is seamlessly integrated into Pix, our pharmacophore informatics software which is accessible to all scientists working in the early phases of drug discovery at BSP. An overview of the individual components will be given. Certain combinations of tools and scenarios to predict drug-like properties and to guide hit-to-lead and lead optimization will be discussed. Application examples will be shown and an analysis of the tool’s impact will be presented.

Drug Design - Solved and Unsolved Problems

Dozent	Prof. Dr. Hugo Kubinyi, BASF SE and University of Heidelberg, Germany
Ort	ZBH, Raum 16
Zeit	04. Nov 2010 16:00

Chemoinformatics and molecular modeling have reached a high level: QSAR, pharmacophore searches, docking and scoring, and ADMET predictions aim to contribute to the discovery of new drugs. However, there is still a gap between the achieved results and the medicinal chemists' needs in lead discovery and optimization. In Hermann Hesse‘s virtual land Castalia, intellectual efforts have no purpose other than the preservation and advancement of intellectual foundations of culture and humanity ... [they] engage in an intellectual exercise, which aims at connecting scientific and cultural values within a formal framework of mathematics and music (The Glass Bead Game“). Does this also describe the current situation in QSAR, molecular modeling and virtual screening? QSAR has the problem of insufficient external predictivity, most often resulting from inappropriate model selection and validation; modeling often does not adequately consider molecular properties and their influence on pharmacophore definitions; virtual screening and docking suffer from the lack of reliable scoring functions. A promising strategy in medicinal chemistry is the systematic variation of molecular scaffolds. There are three essential requirements for successful scaffold hopping: same or very similar geometry of the exit vectors of the original and the new scaffold; similar shape of the new scaffold; and compatible pharmacophore properties - otherwise no binding or different binding modes may result. Many blockbuster drugs resulted from simple or sophisticated scaffold hops, among them various receptor ligands, e.g. ß-blockers and CNS-active agents, but also drugs like atorvastatin. Even nature used scaffold hops, e.g. in the design of different ß-lactame antibiotics. Experience shows that scaffold hops may result in enhanced biological activity, modulation of selectivity and/or physico¬chemical properties and intellectual property of „me too“ analogs. ADMET predictions often fail because of the complexity of absorption and metabolism, especially the formation of toxic metabolites. Future computational chemistry research should focus on solutions to these problems, instead of keeping to the beaten tracks.

Interplay of Binding, Conformational Change and Reactivity

Dozent	Dr. Petra Imhof, Universität HeidelbergInterdisziplinäres Zentrum für wissenschaftliches Rechnen
Ort	ZBH, Raum 16
Zeit	28. Oct 2010 16:00

Enzymes play a fundamental role in diverse biochemical processes such as metabolism, motion and gene regulation. They are extraordinary catalysts in that they are both highly efficient yet specific. Their specificity is often explained by either a lock-and-key scenario or an induced-fit mechanism. I will present results from molecular simulations which provide a dynamic view on substrate binding and regulation. Our data suggests that a delicate interplay of enzyme-substrate interactions and conformational dynamics leads to enzymatic specificity, rather than either of the two extremes described above.

Algorithms for Efficient Exact Motif Discovery

Dozent	Prof. Dr. Sven Rahmann, Computer Science Department, TU DortmundBioinformatics for High-Throughput Technologies
Ort	ZBH, Raum 16
Zeit	17. Sep 2010 12:15

The motif discovery problem consists of finding over-represented patterns in a collection of biosequences. It is one of the classical sequence analysis problems, but still has not been satisfactorily solved in an exact and efficient manner. This is partly due to the large number of possibilities of defining the motif search space and the notion of over-representation. Even for well-defined formalizations, the problem is frequently solved in an ad hoc manner with heuristics that do not guarantee to find the best motif. We show how to solve the motif discovery problem (almost) exactly on a practically relevant space of IUPAC generalized string patterns, using the p-value with respect to an i.i.d. model or a Markov model as the measure of over-representation. In particular, (i) we use a highly accurate compound Poisson approximation for the null distribution of the number of motif occurrences. We show how to compute the exact clump size distribution using a recently introduced device called probabilistic arithmetic automaton (PAA). (ii) We describe an efficient algorithm to discover the optimal pattern with respect to the p-value. The method exploits monotonicity properties of the compound Poisson approximation and new bounds on the expected motif clump size and is two orders of magnitude faster than our previous algorithm presented at ISMB'09.

Multiscale Modelling - From the active centre to the chemical reactor / Molecular Modelling of Carrier-Mediated Transport

Dozent	Prof. Dr. F. J. Keil / Dr.-Ing. S. Jakobtorweihen, TU Hamburg-Harburg, Institut für Chemische Reaktionstechnik
Ort	ZBH, Raum 16
Zeit	15. Jul 2010 16:00

Abstracts: Multiscale Modeling - From the active centre to the chemical reactor Prof. Dr. Frerich J. Keil A hierarchical multiscale approach is presented to calculate effective rates of reaction for zeolite catalysed reaction systems. The first step in this approach involves the determination of intrinsic rate coefficients for all elementary reactions by means of quantum chemical calculations combined with transition state theory. The second step is the calculation of adsorption isotherms and diffusion coefficients of all species by means of Monte Carlo and molecular dynamics simulations. The third step comprises a continuum description of a zeolites crystal based on the reaction-diffusion equation. For coupling the intrinsic rate coefficients obtained in the first step to the continuum variables, a model is proposed that calculates the local concentration of reactants at the catalytically active centres from the total concentrations of adsorbed species. The adsorption isotherms and diffusivities are coupled to the continuum variables by analytical theories such as the ideal adsorbed solution theory and the Maxwell-Stefan formulation of multicomponent diffusion. The fourth step involves fixed-bed reactor simulations based on the continuum description of single zeolites crystals. The approach is illustrated by means of the alkylation of benzene with ethene over zeolites H-ZSM-5. Molecular Modelling of Carrier-Mediated Transport Dr.-Ing. Sven Jakobtorweihen Transport through the mitochondrial membrane is of great relevance for metabolism. Experimental data concerning these transport processes are scarce or even unavailable. Transport through the mitochondrial membrane is facilitated by transport proteins. We are aiming for a molecular understanding of the steps involved in the substrate transport facilitated by these transporter proteins. In particular, we are studying the mechanism of the oxoglutarate-malate-shuttle. First steps are the comparative modelling of the oxoglutarate transporter and the binding of the substrate. With molecular dynamics simulations it is possible to simulate the dynamics of a molecular system. However, this approach is limited to a few hundred nanoseconds, whilst the substrate transport into the mitochondrion takes place on larger time scales. To overcome this constraint special simulation techniques have to be applied (e.g., biased simulation techniques).

Biological imaging with X-ray free electron lasers

Dozent	Dr. Anton Barty, Centre for Free Electron Laser Science (CFEL), Hamburg
Ort	ZBH, Raum 16
Zeit	08. Jul 2010 16:00

Identifying unknown metabolites and other hard problems in bioinformatics

Dozent	Prof. Dr. Sebastian Böcker, Lehrstuhl für Bioinformatik, Universität Jena
Ort	ZBH, Raum 16
Zeit	10. Jun 2010 16:00

The structural elucidation of organic compounds in complex biofluids and tissues remains a significant analytical challenge. For mass spectrometry, the manual interpretation of collision-induced dissociation (CID) mass spectra is cumbersome and requires expert-knowledge, as the fragmentation mechanisms of ions formed from small molecules is not completely understood. The automated identification of compounds is generally limited to searching databases in spectral libraries. In my talk, I will present a method for interpreting the CID spectra by computing fragmentation trees that establish not only the molecular formula of the compound and all fragment ions, but also the dependencies between fragment ions. I will also report some preliminary results regarding the automated comparison of such trees for, say, clustering unknown compounds. It turns out that computationally hard problems lie at the core of this application. This is the case for many problems in bioinformatics, and prominent examples include the reconstruction of phylogenetic trees, comparative genomics, or network comparisons. Parameterized algorithmics tries to circumvent the complexity of these problems by exploiting problem-specific parameters. In the second part of my talk, I will present three examples of NP hard problems for which we have successfully developed algorithms that guarantee to find the exact solution: Cluster Editing for clustering objects, Flip Supertrees for computing phylogenetic supertrees, and Tanglegrams for visually comparing phylogenetic trees.

Molecules on Surfaces: A Challenging Hybrid System

Dozent	Prof. Dr. Roland Wiesendanger, University of Hamburg, Institute of Applied Physics and Interdisciplinary Nanoscience CenterERC Advanced Research Group FURORE
Ort	ZBH, Raum 16
Zeit	20. May 2010 16:00

Molecular Dynamics Simulations on DNA-binding Proteins

Dozent	Dr. Harald Lanig, Friedrich-Alexander-Universität Erlangen/Nürnberg, Computer-Chemie-Centrum, Erlangen
Ort	ZBH, Raum 16
Zeit	06. May 2010 16:00

Restriction-modification (R-M) systems play a crucial role in modulating the horizontal transfer of genes in bacterial populations - a major factor in the transmission of antibiotic resistance between bacterial species. The emergence of multi-drug resistant strains through horizontal transfer of antibiotic resistance genes represents an increasing threat to world health. An understanding of the molecular mechanisms of gene regulation in R-M systems, and their impact on the flow of genetic information in bacterial populations, is thus of considerable interest. The tetracycline repressor TetR is in many ways representative for signal-transduction proteins. It can be used as a switchable gene regulator for both eukaryotic and prokaryotic organisms, which encouraged many structural, mechanistic, and theoretical investigations. Molecular dynamics simulations on several variants (normal and reverse phenotype) of TetR, both in the absence and presence of an inducer, show significant differences in the structures and dynamics of the induced and non-induced forms of the protein. In this talk, I will discuss how we can use MD simulations to gain atomistic insight into the mechanisms of action of these important classes of regulatory protein complexes.

Alignment-free inter-and intraspecific sequence comparison

Dozent	Prof. Dr. Bernhard Haubold, Max-Planck-Institute for Evolutionary Biology, PloenDepartment of Evolutionary Genetics
Ort	ZBH, Raum 16
Zeit	29. Apr 2010 16:00

Alignment-free sequence comparison is an efficient alternative to classical alignment-based sequence com¬parison. However, until recently alignment-free distance metrics were not directly convertible into evolutionary distances. To fill this gap, we have derived a moment-based alignment-free estimator of the substitution rate. Its implementation based on an enhanced suffix array can be used to rapidly compute the pairwise distances between genome samples ranging from 825 HIV strains to 12 species of Drosophila. The corresponding software is freely available from guanine.evolbio.mpg.de/kr2/Our approach is particularly suited for the comparison of closely related sequences. Hence we are now extending it to intra-specific sequence comparison and have derived an alignment-free maximum-likelihood estimator of the number of pairwise mismatches. I will discuss its application in a sliding window analysis of genome-wide local genetic diversity in Drosophila melanogaster.

Towards a structural characterization of ethylene reception in plants

Dozent	Dr. Jochen Müller-Dieckmann, EMBL Hamburg
Ort	ZBH, Raum 16
Zeit	22. Apr 2010 16:00

Despite its simple character, ethylene regulates numerous developmental effects in plants like growth, ripening, senescence, seed germination and responses to stress and pathogens. In Arabidopsis thaliana five partly redundant, membrane bound receptors (ETR1, ETR2, ERS1, ERS2, EIN4) sense ethylene and initiate the signal transduction pathway. The immediate downstream target of ethylene receptors in A. thaliana is CTR1, a putative Raf-like MAPK kinase kinase. Just like the ethylene receptors, CTR1 is a negative regulator of ethylene signalling. In vivo mutants of CTR1 with abolished protein kinase activity show a constitutive triple response (CTR), a hallmark of ethylene signalling. Thus, ethylene acts as an inverse agonist which inhibits activity of its receptors releasing the negative regulation of ethylene signalling by CTR1.

The protein species concept

Dozent	Prof. Dr. Hartmut Schlüter, Universitätsklinikum Hamburg-EppendorfInstitut für Klinische Chemie
Ort	ZBH, Raum 16
Zeit	15. Apr 2010 16:00

Protein species is the term for protein variants, which are coded by one single gene, but vary in their exact chemical composition. This term covers splicing variants, truncated proteins and post translational modified proteins. The importance of the knowledge of the exact chemical composition of a protein species is given by the relationship between its composition and its function. It has been known for several decades that post translational modifications such as phosphorylation critically determine the activity of enzymes. E.g. proteolytic truncations can activate proteases, peptide hormones or receptors. However, despite this knowledge, the relationship between the exact chemical composition of a protein and its function is not yet fully taken into account in many investigations of proteins. In many proteomics studies, protein identification is based on sequence coverage significantly less than 100 %. Furthermore, post translational modifications are more or less ignored. A second drawback concerning the comprehensive description of protein species derives from the absence of a nomenclature, which describes the exact chemical composition. Therefore, we have had to deal with a huge ambiguity concerning the identity of a protein and its function. Since functions have often been assigned to proteins as coded for by a single gene, rather than specific protein species, multiple functions have been assigned to single proteins. As a result, problems occur when searching of protein knowledge databases for defined functional properties of proteins. These problems arise from synonyms, inaccuracies according the assignment of function and exact chemical composition of proteins and missing links to the experimental data, as will be shown by examples. To overcome this problem a terminology for protein species has been developed.

Cheminformatics and Metabolism at the European Bioinformatics Institute

Dozent	Dr. Christoph Steinbeck, European Bioinformatics Institute (EBI), Hinxton, Cambridge, UKChemoinformatics and Metabolism
Ort	ZBH, Raum 16
Zeit	08. Apr 2010 16:00

The Quest for novel Protein Kinase Inhibitors - Rational Design of new Hinge-binders

Dozent	Dr. Hans BriemBayer Schering Pharma AG, GDD-LGO-MC VII, Computational Chemistry, Berlin
Ort	ZBH, Raum 16
Zeit	07. Jan 2010 16:00

Protein Kinases constitute a very important drug target class, in particular in cancer research. As a consequence, over the last few years nearly every major pharmaceutical company has been involved in the search for new protein kinase inhibitors. Thus, patent space in this field has become more and more crowded. To tackle this problem, we have developed a structure-based design workflow which has proven its potential to deliver novel, patentable lead structures in running kinase projects at Bayer Schering Pharma.

Regulatory Genomics

Dozent	Prof. Dr. Martin Vingron, Max-Planck-Institut für molekulare Genetik, Berlin
Ort	ZBH, Raum 16
Zeit	17. Dec 2009 16:00

In this talk we review a biophysical method for prediction of transcription factor affinity to binding sites on the DNA. The affinity prediction is calibrated to reproduce ChIP-chip values where these are available, while also allowing for prediction solely based on a weight matrix description of a binding site. We also computed statistics for the significance of the affinity values, which allows comparing predicted binding behavior of different factors. Further, we computed tissue specific transcription factors by analyzing promoters from sets of tissue specific genes. Results confirm established knowledge and provide several new predictions.

Towards a transcriptional taxonomy of in vitro stem cell preparations

Dozent	Dr. med. Franz-Josef Müller, Zentrum für Integrative Psychiatrie, Kiel
Ort	Raum 16
Zeit	10. Dec 2009 16:00

Human self-renewing multipotent and pluripotent stem cells that can be grown in culture hold great promise for understanding and treating human disease. Functional assays, such as the teratoma assay, characterize stem cells by only a few stem cell markers and fail to recognize important differences between seemingly “identical” stem cell preparations. In order to better understand the molecular control of stem cell phenotypes, we performed comprehensive studies examining genome wide transcriptional profiles on a large and diverse set of samples of human neural, embryonic, induced pluripotent stem cells and pluripotent embryonal carcinoma cell lines. This database - termed Stem Cell Matrix - allowed us to discover with a purely data driven approach clusters of diverse stem cell types based on global gene expression that distinguish human embryonic stem cell preparations from more restricted stem cell and progenitor populations. We found for example, that pluripotent stem cells can be defined by means of an unique transcriptional phenotype in global gene expression data. This phenotype can also be discovered with other genome-wide profiling technologies, such as whole genome methylation and miRNA -profiling. This algorithm could find its first application in identifying successfully reprogrammed induced pluripotent stem cell preparations without cumbersome and strenuous animal experimentation like it is required for the teratoma assay. We used further bioinformatic analysis to uncover a protein-protein network (PluriNet) consisting of interacting proteins that is shared by the pluripotent cells. Analysis of published data showed that this protein interaction network appears to be a common characteristic of many types of pluripotent cells, including murine embryonic stem cells and induced pluripotent cells and human oocytes, is preserved across species boundaries and recapitulates many recent discoveries in ESC and iPSC research. Our proof of principle study demonstrates that large-scale high throughput approaches such as whole genome transcriptional profiling are a valuable method to characterize and monitor stem cell preparations. This approach is analogous to taxonomical approaches that have been historically used to classify zoological specimens.

Adapting Algorithms from Bioinformatics to Chem(o)informatics

Dozent	PD Dr. Michael Hutter, Zentrum für Bioinformatik, Universität des Saarlandes, Saarbrücken
Ort	ZBH, Raum 16
Zeit	26. Nov 2009 16:00

Using statistically derived information from a large number of DNA or protein sequences to make predictions about new or unknown sequences is a common concept in Bioinformatics. The underlying method that leads to the log odds scores present in BLOSUM and PAM matrices is the evaluation of the frequency of pair-wise exchanges. Using the corresponding matrix entries it is possible to predict the most likely exchange. This likeliness concept has been transferred to other prediction problems in the field of Chem(o)informatics. Using only atom type combinations in chemical compounds it was possible to predict their suitability as pharmaceutical drugs. Together with other molecular descriptors this allows a fast classification of compounds into drugs and nondrugs. Evaluation of reoccurring chemical modifications in marketed drugs led to a so-called bioisosteric exchange matrix. It turned out that hydrogen-bonding properties are less conserved than usually assumed. Using the SMILES notation, this method furthermore allows the alignment and the computation of the homology between molecules. In virtual screening, this produces somewhat different hits than conventional similiarity searches.

Deciphering the Swine-flu pandemics of 1918 and 2009

Dozent	Prof. Richard A. Goldstein, National Institute for Medical Research, Mathematical Biology, London, UK
Ort	ZBH, Raum 16
Zeit	19. Nov 2009 16:00

The devastating ‘Spanish flu’ of 1918 killed an estimated 50 million people worldwide, ranking it as the deadliest pandemic in recorded human history. It is generally believed that the virus transferred from birds directly to humans shortly before the start of the pandemic, subsequently jumping from humans to swine. By modelling how the viral sequences changes in human, avian, and swine hosts, we find it likely that the Spanish flu of 1918, like the current 2009 pandemic, was a 'swine-origin' influenza virus. Now that we are faced with a new pandemic, can we understand how influenza is able to change hosts? Again by modelling the evolutionary process, we can identify locations that seem to be under different selective constraints in humans and avian hosts, identifying changes that may have facilitated the establishment of the 2009 swine-origin flu in humans.

Forces and Conformatoinal Dynamics in Biomolecular Nanomachines

Dozent	Prof. Helmut Grubmüller, Max-Planck-Institute for Biophysical Chemistry, GöttingenTheoretical and Computational Biophysics
Ort	ZBH, Raum 16
Zeit	12. Nov 2009 16:00

Proteins are biological nanomachines. Virtually every function in the cellis carried out by proteins - ranging from protein synthesis, ATP synthesis, molecular binding and recognition, selective transport, sensor functions, mechanical stability, and many more. The combined interdisciplinary efforts of the past years have revealed how many of these functions are effected on the molecular level. Computer simulations of the atomistic dynamics play a pivotal role in this enterprise, as they offer both unparalleled temporal and spatial resolution. With state of the art examples, this talk will explain the basics of this high performance computing method, the type of questions that can (and cannot) be addressed, and (current) limitations. The examples include aquaporin selectivity, mechanics of F-ATP synthase, flexible recognition by nuclear pore transporters, and the mechanical properties of viral capsids.

Recent advances in biological small-angle X-ray scattering from solutions

Dozent	Dr. Dmitri I. Svergun, European Molecular Biology Laboratory, Hamburg Outstation c/o DESY, Hamburg
Ort	ZBH, Raum 16
Zeit	05. Nov 2009 16:00

Small-angle X-ray scattering (SAXS) is experiencing a renaissance in the studies of macromolecular solutions allowing one to study the structure of native particles and to rapidly analyze structural changes in response to variations in external conditions. Novel data analysis methods [1] have significantly enhanced the resolution and reliability of structural models provided by the technique. Emerging automation of the experiment, data processing and interpretation make solution SAXS a streamline tool for large scale structural studies in molecular biology. The method provides low resolution macromolecular shapes ab initio and is readily combined with other structural and biochemical techniques in multidisciplinary studies. In particular, rapid validation of predicted or experimentally obtained high resolution models in solution, identification of biologically active oligomers and addition of missing fragments to high resolution models are possible. For macromolecular complexes, quaternary structure is analyzed by rigid body movements/rotations of individual subunits. Recent developments made it possible also to quantitatively characterize flexible macromolecular systems, including intrinsically unfolded proteins. The novel methods will be illustrated by advanced SAXS applications to solutions of biological macromolecules. [1]. Petoukhov, M.V.Svergun, D. I. (2007) Curr Opin Struct Biol. 17, 562-571.

Protein-Protein Interaction: Interface, SurfaceSpecificity (PPI:ISS)

Dozent	Dr. K. Anton Feenstra, IBIVU/Bioinformatics, Free University Amsterdam
Ort	ZBH, Raum 16
Zeit	22. Oct 2009 16:00

Many protein families involved in protein-protein interactions (PPI) contain sub-families that interact with different protein binding partners, or have no interaction partner. Specificity in these interactions is often critical to the function of the proteins, so this specificity may be used to pinpoint sites that are critical for the interaction. We have applied the Sequence Harmony method to detect specificity sites from multiple sequence alignments (MSAs) containing an interacting and a non-interacting sub-family. For each set of orthologous interacting proteins, a set of non-interacting paralog proteins was included. Interaction and non-interaction was based on published so-called 'socio-affinity scores' for S. cerevisiae [Genome Biol (2007) 8:R197]. Exploiting the differences in sequence composition between the binding and non-binding groups by Sequence Harmony, we demonstrate that predicted specificity residues turn out to reside on the protein surface; between 70% and 95% of selected sites are surface residues, depending on the protein family examined. We also show that we can select interface residues using the 78 available S. cerevisiae protein complex crystal structures as reference, with coverage (true-positive rate) for interface residue detection from 27% and error (false-positive rate) from 14%, up to 99% and 45% respectively, depending on the Sequence Harmony cutoff selected.

LocARNA for Comparative Analysis of Structural RNA

Dozent	Dr. Sebastian Will, Institute for Computer Science, Albert-Ludwigs University FreiburgChair for Bioinformatics
Ort	ZBH, Raum 16
Zeit	16. Jul 2009 16:00

Recent discoveries show that non-coding RNA plays a central role in organisms, in many cases ascribed to the structure of the RNA. For many years, simultaneous alignment and folding has been considered the gold-standard for the comparison of structural RNAs. However, only recent achievements have permitted the practical application of this technique. One of the most successful algorithms was developed in the LocARNA system for multiple alignment of RNA. I will overview the core functionality of LocARNA, explain extensions, and discuss selected applications. These comprise performance evaluation in terms of alignment accuracy, genome-wide clustering of potential non-coding RNA, locality in RNA alignment, and the use of alignment reliability.

SAR Landscapes and Analysis Functions

Dozent	Prof. Dr. Jürgen Bajorath, Department of Life Science Informatics, University of BonnLIMES Program Unit Chemical Biology and Medicinal Chemistry
Ort	ZBH, Raum 16
Zeit	09. Jul 2009 16:00

In medicinal chemistry, the exploration of structure-activity relationships (SARs) plays a fundamental role. We develop computational methods to systematically explore SARs and compare global and local SAR features. For this purpose, the concept of activity landscapes is of fundamental importance. SAR analysis functions that relate compound similarity and potency relationships to each other are combined with molecular network representations for large-scale SAR profiling, the analysis of heterogeneous SAR microenvironments, and the identification of key compounds. Using SAR analysis functions and network representations, SAR information can also be extracted from screening data sets to prioritize hits for chemical exploration. Screening data are shown to contain SAR pathways that consist of series of similar compounds with steadily increasing potency and lead to local activity cliffs. Along such SAR pathways scaffold transitions are observed.

Computational Short Read Metagenomics

Dozent	Prof. Dr. Jens Stoye, Technische Fakultät, Genominformatik, Universität Bielefeld
Ort	ZBH, Raum 16
Zeit	02. Jul 2009 16:00

Metagenomics is a new field of research on metagenomes, where natural microbial communities are studied. The new sequencing techniques like 454 or Solexa-Illumina sequencing promise new possibilities as they are able to produce huge amounts of data in much shorter time and with less efforts and costs than the traditional Sanger sequencing. But the data produced comes in even shorter reads (35-50 base pairs with Solexa-Illumina, 100-300 basepairs with 454 sequencing). CARMA [1] is a new pipeline for the characterization of the species composition and the genetic potential of microbial samples using 454-sequenced reads. The species composition can be described by classifying the reads into the taxonomic groups of organisms they most likely stem from. By assigning the taxonomic origins to the reads, a profile is constructed which characterizes the taxonomic composition of the corresponding community. The CARMA pipeline has already been successfully applied to 454-sequenced communities [2,3] including the characterization of a plasmid sample isolated from a wastewater treatment plant [4]. Using samples from a biogas plant we examined the applicability of this approach for the ultra-short Solexa-Illumina reads by comparing the results with those obtained by the 454-sequenced sample [5,6]. Our results using 77 million 50 bp-reads revealed that this approach indeed produces consistent results. Most diffences we have found are in the taxa of higher order, e.g. in the species level, and in general for species with a very low presence. In order to apply CARMA to high-throughput sequencing data, we had to improve the accuracy and speed of our method in various ways: A preprocessing assembly phase using an adapted q-gram index [7]; adaptation of the pipeline to take the information of mated reads into account to increase read length; modification of the amino acid sequence distance function for the construction of the phylogenetic tree; and implementation of a protein-q-gram index over a multiple alignment for the read-against-Pfam protein family matching.References [1] L. Krause, N.N. Diaz, A. Goesmann, S. Kelley, T.W. Nattkemper, F. Rohwer, R.A. Edwards, J. Stoye (2008) Phylogenetic classification of short environmental DNA fragments. Nucleic Acids Res. 36(7):2230-2239. [2] E.A. Dinsdale, O. Pantos, S. Smriga, R.A. Edwards, F. Angly, L. Wegley, M. Hatay, D. Hall, E. Brown, M. Haynes, L. Krause, E. Sala, S.A. Sandin, R. Vega Thurber, B.L. Willis, F. Azam, N. Knowlton, F. Rohwer (2008) Microbial Ecology of Four Coral Atolls in the Northern Line Islands. PLoS ONE 3(2):e1584. [3] S.A. Sandin, J.E. Smith, E.E. DeMartini, E.A. Dinsdale, S.D. Donner, A.M. Friedlander, T. Konotchick, M. Malay, J.E. Maragos, D. Obura, O. Pantos, G. Paulay, M. Richie, F. Rohwer, R.E. Schroeder, S. Walsh, J.B.C. Jackson, N. Knowlton, E. Sala (2008) Baselines and Degradation of Coral Reefs in the Northern Line Islands. PLoS ONE, 3(2):e1548. [4] A. Schlüter, L. Krause, R. Szczepanowski, A. Goesmann, A. Pühler (2008) Genetic diversity and composition of a plasmid metagenome from a wastewater treatment plant. J. Biotechnol. 136(1-2): 65-76. [5] L. Krause, N.N. Diaz, R.A. Edwards, K.-H. Gartemann, H. Krömeke, H. Neuweger, A. Pühler, K.J. Runte, A. Schlüter, J. Stoye, R. Szczepanowski, A. Tauch, A. Goesmann (2008) Taxonomic composition and gene content of a methane-producing microbial community isolated from a biogas reactor. J. Biotechnol. 136(1-2), 91-101. [6] A. Schlüter, T. Bekel, N.N. Diaz, M. Dondrup, R. Eichenlaub, K.-H. Gartemann, I. Krahn, L. Krause, H. Krömeke, O. Kruse, J.H. Mussgnug, H. Neuweger, K. Niehaus, A. Pühler, K.J. Runte, R. Szczepanowski, A. Tauch, A. Tilker, P. Viehöver, A. Goesmann (2008) The metagenome of a biogas-producing microbial community of a production-scale biogas plant fermenter analysed by the 454-pyrosequencing technology. J. Biotechnol. 136(1-2):77-90. [7] K. Rasmussen, J. Stoye, E.W. Myers (2006) Efficient q-Gram Filters for Finding All epsilon-Matches over a Given Length. J. Comp. Biol. 13(2):296-308.

The mobile metagenome: a structure-to function route

Dozent	Prof. Bridget Mabbutt, Department of Chemistry and Biomolecular Sciences, Macquarie University, Sydney, Australia
Ort	ZBH, Raum 16
Zeit	18. Jun 2009 16:00

The acknowledged extent of lateral gene transfer (LGT) presents challenges to genomic studies of microorganisms: no one cell line is representative of the gene content of a species as a whole. We have been directly recovering mobile gene cassettes from environmental sources to screen the bacterial mobile gene pool, a particularly rich reservoir for adaptive functions. Due to the high sequence novelty of the recovered ORFs, it is only through 3D protein structure elucidation that their functions can be annotated. I present our successfully exploitation of structural genomics techniques on approximately 300 ORFs, and outline the functional repertoire of the many protein crystal structures solved.

Drug Discovery in the academic setting: the European ScreeningPort model ?

Dozent	Dr. Philip Gribbon, European ScreeningPort GmbH, Hamburg
Ort	ZBH, Raum 16
Zeit	11. Jun 2009 16:00

The European ScreeningPort identifies starting points for pre-clinical research programs by providing industrial scale drug discovery facilities to academic organizations. The inputs are first in class therapeutic targets, originating from academic Institutions, which have yet to be picked up by biotech/Pharma. The outputs cover the range from tool compounds, for mechanistic studies and PoC programs, through to high quality drug-like validated hits suitable for progression to lead finding and pre-clinical candidate stage. In this talk, the process and challenges of academic drug discovery will be discussed in detail using examples from in-house and pharmaceutical industry projects.

A flexible approach to induced fit docking

Dozent	Dr. Sander B. Nabuurs, Computational Drug Discovery, Center for Molecular and Biomolecular Informatics, Radboud University NijmegenMedical Centre
Ort	ZBH, Raum 16
Zeit	14. May 2009 16:00

Reliably predicting the binding modes of a set of (proposed) compounds, taking both receptor and ligand flexibility into account, and subsequently ranking these compounds by their binding affinity remains a tremendous challenge. First, I will present the methodology behind and recent developments in our induced fit docking tool Fleksy. The second part of my talk will focus on realistic large-scale induced fit cross-docking studies using sets of receptor-ligand complexes with known affinity. Our progress towards an affinity prediction model based on obtained docking results will be presented and potential pitfalls and requirements to arrive at accurate predictions will be discussed.

Comparative Analysis of the Venom Proteomes of Vipera ammodytes ammodytes and Vipera ammodytes meridionalis

Dozent	Prof. Dr. Christian Betzel, University of Hamburg, Institute of Biochemistry and MolecularbiologyLaboratory for Structural Biology of Infection and Imflammation
Ort	ZBH, Raum 16
Zeit	07. May 2009 16:00

Snake venoms are an extremely rich source of pharmacologically active proteins with a high potential for medical and pharmaceutical applications. However, at present this potential is only partially exploited, mainly due to the incomplete knowledge of the venom proteome and the pharmacological properties of its components, in particular those without enzymatic activity. Two venomous Viperid snakes are most widespread in East Europe: Vipera ammodytes ammodytes, inhabiting the west part of the region (Slovenia, Croatia, Serbia, West Bulgaria) and Vipera ammodytes meridionalis, a snake of high public health significance in Greece, East Bulgaria, Roumania, Albania and Turkey. Both snakes are today the most frequent cause of snakebites and in cases also with lethal consequence. The venoms of these snakes cause a number of distinct pathophysiological changes, such as local tissue damage, pain, paralysis, bleeding, fever, tachycardia, apoptosis and other systemic symptoms. The Vipera ammodytes bite is a serious event that requires immediate hospital care. To obtain insights about the venome composition we analysed in a comparative approach and in detail by FPLC, 2-D electrophoresis, sequence analysis, MS/MS and X-ray crystallography the venom proteomics of Vipera ammodytes ammodytes and Vipera ammodytes meridionalis. FPLC analysis showed the presence of L-amino acid oxidase, monomeric and heterodimeric phospholipases A2, C-type lectin protein, and proteinases in the venom of V. a. ammodytes. Representatives of the same protein families were found in the venom of the other subspecies, V. a. meridionalis. N-terminally identical PLA2 neurotoxins were identified in both venoms. However a distinct difference in the PLA2 compositions of the venoms was also observed: a monomeric protein with phospholipase A2 activity, identical in the first 20 amino acid residues to the catalytically inactive acidic component of the heterodimeric PLA2 present in both venoms, was found only in that of V. a. meridionalis. Probably, this protein represents an intermediate form of the two components of the hetereodimer. So far we analysed a few PLA2 structuresby X-ray crystallography to high resolution. 2-D electrophoresis and MS/MS analysis revealed further, that the two venoms shared a number of protein families: monomeric and heterodimeric Group II PLA2s, serine proteinases, Group I, II, and III metalloproteinases, L-amino acid oxidases (LAAOs), cysteine-rich secretory proteins, disintegrins, and growth factors. Totally we identified 38 venom components of the V. a. ammodytes, belonging to 9 protein families, and 67 components of the V. a. meridionalis venom belonging to 8 protein families. Furthermore the venom proteome of V. a. ammodytes shows larger diversity of proteins (139) in comparison to that of V. a. meridionalis (104 proteins. The protein compositions explains clinical effects of the V. ammodytes snakebites, such as difficulties in the breathing, paralysis, apoptosis, cloting disorders, hemorrhage, and tissue necrosis. The detailed knowledge of the venom toxin composition of the two vipers and biological properties of the protein constituents will allow the design of novel drugs, more effective treatment of the snake-bite consequences and assessment of important physiological processes. Details will be presented.

Support of Drug Discovery by Searching Fragment Spaces with Feature Trees

Dozent	Dr. Uta Lessel, Boehringer Ingelheim Pharma GmbHCo. KG, Biberach an der RissLeitstrukturfindung
Ort	ZBH, Raum 16
Zeit	30. Apr 2009 16:00

Virtual combinatorial chemistry easily produces billions of compounds, which cannot be screened in a conventional manner even with the fastest methods available. The generation of Fragment Spaces encoding huge numbers of virtual compounds by their fragments/reagents and rules how to combine them represents an efficient solution for such a scenario. Fragment Spaces can be screened with so-called Fragment Space searches e.g. based on the Feature Tree descriptor without ever fully enumerating all virtual products. In the presentation experiences with Fragment Space searches based on the Feature Tree descriptor are shared and a workflow to use this methodology in a pharmaceutical setup is described.

The role of single-stranded DNA binding proteins in the replication of viral genomes

Dozent	Dr. Paul Tucker, EMBL Hamburg c/o DESY, Hamburg
Ort	ZBH, Raum 16
Zeit	09. Apr 2009 16:00

Single-stranded DNA binding proteins (SSBs) are essential components of the DNA replication machinery. The structures of the SSBs of the adenoviridae and the herpesviridae will be discussed with special reference to the structural features responsible for cooperative binding to ssDNA.

Evolution of protein complexes

Dozent	Dr. Roland Krause, FU Berlin, MPI for Molecular Genetics, Berlin
Ort	ZBH, Raum 16
Zeit	05. Feb 2009 16:00

The repertoire of protein complexes in yeast is well known and of sufficient quality for in depth analysis. For a true comparative study, similar data for other organisms need to be obtained. While large scale data sets in worm, fly and human have been generated, their coverage is still low. In the last months, several data sets for the yeast Schizosaccharomyces pombe have become available. I will discuss this bioinformatic frontier, including mapping of orthologs, the importance of negative results for interaction studies and methods of tree reconciliation for reconstruction of ancestral protein complexes.

Chasing the AIDS Virus

Dozent	Prof. Dr. Thomas Lengauer, Max-Planck-Institut für Informatik, SaarbrückenComputational Biology and Applied Algorithmics
Ort	ZBH, Raum 16
Zeit	29. Jan 2009 16:00

AIDS is one of the major plights of humanity. The prime problem in applying antiviral drug therapy to AIDS patients is the variability of the AIDS virus HIV. Presented with a given drug therapy the virus follows evolutionary escape paths to drug resistance, which renders the drug therapy ineffective. Statistical computer models can help estimate the level of resistance of a given viral variant against a given antiviral drug. Furthermore, such models can estimate the evolutionary path to resistance followed by the virus. We describe the science behind the geno2pheno Server www.geno2pheno.org which offers such models via the internet: Given the relevant portion of an HIV genome, geno2pheno predicts the resistance of HIV to any of a number of antiviral drugs that are in clinical use. Furthermore the server ranks combination drug therapies with respect to their expected effectiveness against the given HIV variant. A new class of antiviral drugs - the so-called entry inhibitors - aims at preventing HIV from entering the host immune cell. Here, we are confronted with a different computational challenge: When entering the human host cell, HIV uses one of two coreceptor molecules on the cell surface. We present a statistical model offered via the geno2pheno server that predicts which of the two coreceptors the virus uses. Determining this “viral tropism” is essential for disease prognosis and in the context of applying the new entry inhibitors. Both model classes are trained using various linear and nonlinear statistical learning procedures. The training data are carefully assembled databases comprising relevant genotypic, phenotypic and clinical parameters. While the resistance model only incorporates sequence features, one version of the tropism model also involves information on the structure of the relevant portion of the viral gp120 protein that docks to the human cell. Both models are available via the webserver. The geno2pheno server has been developed in the context of the Arevir consortium, a German National research consortium targeted at the bioinformatical analysis of HIV resistance data, and is currently in prototypical use for research purposes. Members of the consortium and their associated practices treat about two thirds of the AIDS patients in Germany. Beerenwinkel, N., Sing, T., Lengauer, T., Rahnenführer, J., Roomp, K., Savenkov, I., Fischer, R., Hoffmann, D., Selbig, J., Korn, K., et al. 2005. Computational methods for the design of effective therapies against drug resistant HIV strains. Bioinformatics 21: 3943-3950. Lengauer, T., and Sing, T. 2006. Bioinformatics-assisted anti-HIV therapy. Nat Rev Microbiol 4: 790-797. Sander, O., T. Sing, et al. (2007). Structural Descriptors of gp120 V3 Loop for the Prediction of HIV-1 Coreceptor Usage. PLoS Comput Biol 3(3): e58. Altmann, A., Beerenwinkel, N. et al. (2007), Improved prediction of response to antiretroviral combination therapy using the genetic barrier to drug resistance. Antivir Ther, 12(2): p. 169-78. Lengauer, T., O. Sander, et al. (2007). Bioinformatics prediction of HIV coreceptor usage. Nat Biotechnol 25(12): 1407-10.

Structure-based drug design between serendipity and rationality

Dozent	Prof. Dr. Holger Steuber, Institute of Biochemistry, University of Lübeck, Lübeck
Ort	ZBH, Raum 16
Zeit	22. Jan 2009 16:00

Even though profound understanding of protein-ligand recognition provides the basis for successful structure-guided ligand design and optimization, currently only limited knowledge is available about the principles driving ligands to bind to their targets. In particular protein flexibility, the role of water for ligand affinity and binding geometry, as well as changes in protonation of protein residues or ligand functional groups are difficult to consider in the design process in absence of appropriate experimental data, and even more difficult to predict. The TIM-barrel folded enzyme Aldose reductase (ALR2) is a valuable model system to study structural and thermodynamic features of inhibitor binding and, furthermore, represents an excellent drug target. To prevent diabetic complications derived from enhanced glucose flux via the polyol pathway, the development of aldose reductase inhibitors (ARIs) has been established as a promising therapeutic concept. Furthermore its attraction as a test system consists in the extended mobility and adaptivity of its active site residues, giving rise to various distinct binding pocket conformers and pronounced induced-fit adaptations upon ligand binding. The presentation will emphasize the challenges in structure-based design strategies between serendipity and rational approaches.

Computing NMR Spectra Assignments and Protein-Protein Interactions with Paramagnetic NMR

Dozent	Prof. Thomas Huber, School of Molecular and Microbial Sciences, University of Queensland, Brisbane, Australia
Ort	ZBH, Raum 16
Zeit	18. Dec 2008 16:00

Pseudocontact shift effects induced by a paramagnetic lanthanide bound to a protein are enjoying increasing popularity in NMR spectroscopy, as they yield a complementary set of orientational and long-range structural restraints. These shifts reflect the Ä÷ anisotropy tensor, which leads to powerful restraints for determining protein and protein-complex structure. We present new algorithms specifically designed for the computation of the Ä÷-tensor from a experimental paramagnetic shifts. These can be used to 1) swiftly and fully automatically assign NMR spectra and 2) to determine the structure of protein-protein and protein-ligand interactions.

Computational approaches for targeting protein-protein interfaces with small molecules

Dozent	Prof. Dr. Holger GohlkePharmaceutical Institute, Christian-Albrechts-University, Kiel
Ort	ZBH, Raum 16
Zeit	04. Dec 2008 16:00

A promising way to interfere with biological processes is through the control of protein-protein interactions by means of small molecules that modulate the formation of protein-protein complexes. Although the feasibility of this approach has been demonstrated in principle by recent results, many of the small-molecule modulators known to date have not been found by rational design approaches. In large part this is due to the challenges that one faces in dealing with protein binding epitopes compared to, e.g., enzyme binding pockets. Recent advances in the understanding of the energetics and dynamics of protein binding interfaces and methodological developments in the field of structure-based drug design methods may open up a way to apply rational design approaches also for finding protein-protein interaction modulators.1 Here, we first show in a retrospective analysis of the well-investigated interleukin-2 system how I) potential binding sites in an interface can be identified from an unbound protein structure, II) the interface can be dissected in terms of energetic contributions of single residues, and III) one can make use of this knowledge for guiding the development of small-molecule modulators. When applied to a leukaemia-associated fusion protein in a prospective manner, the predictive character of the methodology is demonstrated. Another challenge arises from the fact that protein-protein interfaces are flexible. In the second part, we thus demonstrate a novel approach for including protein flexibility into protein-ligand docking. This approach is based on elastic potential grids, which provide an accurate and efficient representation of intermolecular interactions in fully-flexible docking. 1. Gonzalez Ruiz, D.; Gohlke, H., Targeting protein-protein interactions with small molecules: Challenges and perspectives for computational binding epitope detection and ligand finding. Curr Med Chem 2006, 13, 2607-2625.

Finding robust enzymes in soil metagenomes

Dozent	Prof. Wolfgang Streit, Biozentrum Klein Flottbek, MikrobiologieBiotechnologieUniversität Hamburg
Ort	ZBH, Raum 16
Zeit	27. Nov 2008 16:00

Interactions of non-coding RNAs in eukaryotes: viroid <-> host genome and 5,S rRNA <-> gene regulation

Dozent	Prof. Dr. Gerhard Steger, Heinrich-Heine-Universität Düsseldorf, Institut für Physikalische Biologie
Ort	ZBH, Raum 16
Zeit	13. Nov 2008 16:00

Viroids are minimal plant pathogens that do not code for any protein but possess sequence and structural signals in their RNA genome for replication, processing and transport by the host machinery. I will discuss some recent results on viroid-specific small RNAs that might interact with the host transcriptome similar to microRNAs. Viroids do interact in vivo with 5S ribosomal RNA; both RNAs possess a certain structural RNA element which is critical for binding of transcription factor TFIIIA to 5S rRNA. By serendipity we detected a 5S rRNA-like element in an alternatively spliced exon of TFIIIA. I will discuss this finding with respect to a novel gene regulation model mediated by the alternative splicing of an anciently exonized non-protein-coding element.

Detection of Extrasolar Planets

Dozent	Prof. Dr. Peter H. Hauschildt, Hamburger Sternwarte
Ort	ZBH, Raum 16
Zeit	06. Nov 2008 16:00

The detection and characterization of extrasolar planets is a rapidly growing field of astrophysics since about 1995, when the first extrasolar giant (gas) planet was found. Since then, more than 100 such objects have been discovered. Basically all of them are gas planets like Jupiter, the search for terrestrial planets is far more challenging. I will describe the method used to find extrasolar planets, their drawbacks and future prospects of finding an 'Earth 2'.

Computational Systems Biology in Drug Discovery: From Molecular Events to Clinical Outcome

Dozent	PD Dr. med. Klaus Prank, Medizinische Hochschule Hannover und Städtisches Klinikum Braunschweig gGmbHAbteilung Nuklearmedizin
Ort	ZBH, Raum 16
Zeit	30. Oct 2008 16:00

Historically, drug discovery and systems biology were closely related, observing the action of drugs in people, e.g. in traditional Chinese medicine, where herbal drugs were given to patients, the most relevant complex biological system for drug discovery. Today, drug discovery through the molecular target centered approach has been widely separated from the complex system level of human physiology. We are now starting to re-enter systems biology in modern guise, using mathematical and computational modeling approaches at the cellular, tissue and organ level. The goal of modeling is to provide a framework for hypothesis generation and prediction based on simulation of human disease biology across the multiple distance and time scales, from biochemical reactions to disease responses. Real progress has already been made in developing and testing computational systems biology approaches for cardiac risk assessment. To be useful, the heart model must reach down to the molecular level of proteins (receptors, transporters, enzymes etc.) and must also reconstruct physiology right up at the organ level. Besides the level of disease physiology, highly focused, problem-directed models, such as integrative cell models (e.g. cardiomyocyte) and models of biochemical pathways (e.g. NF-kB signaling in inflammation and EGF/ErbB signaling in cancer) contribute to meaningful drug development decisions by accelerating hypothesis-driven biology and by modeling specific physiological problems in target validation.

Perceptual Representation in Synchronized Neuronal Oscillations

Dozent	Dr. Alexander Maye, Integrative Neurophysiology, Universitätsklinikum Hamburg-Eppendorf
Ort	ZBH, Raum 16
Zeit	23. Oct 2008 16:00

Simulating oscillatory neuronal activity in computational models allows for joint analysis of all state variables in large networks. A mean-field model and a model of phase-coupled oscillators are used to describe the activity of clusters of feature-sensitive neurons in the primary visual cortex representing local object properties like color and orientation. Using an Eigenmode analysis, the stable dynamics of both networks is decomposed into static, spatial components, the eigenmodes, and temporal componentsthe associated order parameters. The talk will explain representational capabilities of oscillator networks and contrast the properties of the two models in response to simple visual stimuli.

The role of mathematical modeling in hypothesis-driven cell biology - a survey of models in bioinformatics and systems biology

Dozent	Prof. Dr. Olaf Wolkenhauer, SystemsBioinformatics, University of Rostock, Rostock
Ort	ZBH, Raum 16
Zeit	10. Jul 2008 16:00

A new paradigm should challenge the way research has been conducted. In this presentation, I will argue the case for systems biology, defined as a merger of systems theory with cell biology. Systems biology signals a shift of focus away from molecular characterization towards an understanding of functional activity. Relatively simple biochemical reaction networks demonstrate that an understanding of cell function demands a hypothesis-driven approach: data do not speak for themselves. I will discuss the role of mathematical modelling in the life sciences. The aim is to convince the experimentalists that there is nothing more practical than a good theory (model)! For more information see: www.sbi.uni-rostock.de

From Bioinformatics and Metabolomics to Systems Biology

Dozent	Prof. Dr. Dietmar Schomburg, Technical University Braunschweig, BioinformaticsBiochemistry
Ort	ZBH, Raum 16
Zeit	03. Jul 2008 16:00

Systems Biology = Networks + Structures?

Dozent	Dr. Michael Lappe, Max-Planck-Institute for Molecular Genetics, Berlin
Ort	ZBH, Raum 16
Zeit	26. Jun 2008 16:00

While Systems Biology is still struggling for a commonly accepted definition, the main paradigm shift seems to be a holistic network view of molecular biology. The general theme is to analyse biomolecules not in isolation but in their functional context in order to put everything back together again. What contributions can be expected from the study of available data of protein structures? The analysis and prediction of protein structures can be represented in terms of networks of non-covalent interactions among residues. Based on these residue interaction graphs (RIGs) we develop methods to reconstruct 3D structures accurately. The aim is to identify the key contacts determining the structure. Predicting protein structure and function is rephrased as de-novo inference of RIGs from sequence alone. This quest leads to the development of multi-body statistical potentials. While this interdisciplinary work finds application in protein- and drug-design it also has implications in studying emergent properties of molecular networks in general. This links the network view of protein structures into the broader context of systems biology.

Chemoinformatics meets statistics - Validation in ligand-based drug design

Dozent	Prof. Dr. Knut Baumann, Institute of Pharmaceutical Chemistry, University of Technology Braunschweig
Ort	ZBH, Raum 16
Zeit	19. Jun 2008 16:00

Ligand-based drug design techniques such as quantitative structure-activity relationships (QSAR) and similarity searching are well established tools for computer-aided drug design. If these techniques are used predictively, they need to be thoroughly validated to be really useful. Validation is defined as the process of checking if a certain procedure performs as expected. Interestingly, not all commonly accepted validation protocols are adequate to perform this task. For instance, the commonly accepted gold standard for validation in QSAR is test set prediction. However, test set validation can be completely useless when carried out in the wrong way. Likewise, in similarity searching validation results can be meaningless if the validation database composition renders the validation task trivial. In the talk a new QSAR method, which can model conformational flexibility, along with a critical assessment of validation procedures used in QSAR is presented. Guidelines for good validation practice will be given. In the second part, pitfalls in validating similarity searching techniques are outlined. It is shown that it is essential to design the database of inactive molecules adequately. Appropriate statistical techniques for this task will be presented.

Identification of novel therapeutic targets in haematological cancers

Dozent	Dr. Stefan Balabanov, UKE, Medizinische Klinik II, Onkologie Hämatologie
Ort	ZBH, Raum 16
Zeit	12. Jun 2008 16:00

Ligand-mediated modulation of the nuclear receptor CAR

Dozent	Dr. Björn Windshügel, Department of Pharmaceutical Chemistry, University of Kuopio, Kuopio, Finland
Ort	ZBH, Raum 16
Zeit	10. Jun 2008 14:30

Pakistani - German Scientific Collaboration: Ideas and Proposals + Research in anti-diabetics and anti-epileptics

Dozent	Prof. M. I. Choudhary, Center for Molecular Medicine and Drug Design, University of Karachi, PakistanInternational Center for Chemical Sciences
Ort	ZBH, Raum 16
Zeit	26. May 2008 16:30

This talk is within the framework of the visiting Pakistani delegation. We will discuss ideas for scientific collaboration and current research topics. The broad themes are medicinal chemistry and bioinformatics, but we will present some examples of research with applications in the areas of diabetes and epilepsy. Members of the delegation will be available for discussion on a wider range of topics.

Geometric Concepts Associated with Neighbor-Joining

Dozent	Wolfram Retter, TU Hamburg-Harburg
Ort	ZBH, Raum 16
Zeit	22. May 2008 16:00

First, an overview of the classical neighbor-joining algorithm is given. This algorithm reconstructs a weighted tree from the distances between its leaves. Then some of the concepts and principles underlying its consistency are extended in two directions: They are generalized to injective metric spaces, of which weighted real trees are the simplest case. And they are sharpened in a natural way that directly captures the case of trees with polytomy, resulting in a 'multiple neighbor-joining' algorithm.

Experimental Therapies for HIV-Infection

Dozent	Prof. Dr. Joachim Hauber, Heinrich-Pette-Institut für Experimentelle Virologie und Immunologie an der Universität HamburgAbteilung für Zellbiologie und Virologie
Ort	ZBH, Raum 16
Zeit	08. May 2008 16:00

Shape Abstraction in RNA Folding and Family Modelling

Dozent	Prof. Dr. Robert Giegerich, Bielefeld University, Bielefeld
Ort	ZBH, Raum 16
Zeit	24. Apr 2008 16:00

When humans reason about functional structures of RNA, they speak of long hairpins with miRNA percursors, of clover leaf structures with tRNA, of neighouring hairpins with attenuators, and so on. Most of the time, we do not care about individual base pairs or helix sizes, while the overall arrangment of helices and loops really matters. RNA folding programs, however, used to be ignorant of abstraction in RNA, deceiving us with a single, minimum free energy structure, or overwhelming us with a plethora of near-optimal structures, most of which are quite similar and therefore redundant. RNA shape abstraction alleviates this situation. RNA shapes are abstract structure images, retaining adjacency and nesting of structural features, but disregarding size information. Shape abstraction integrates well with dynamic programming algorithms, and hence it can be applied during structure prediction rather than afterwards. This avoids exponential explosion in the near-optimal folding space, and provides a non-heuristic and complete account of an RNA molecule's structural inclinations. Quite magically, some long-studied problems become easy. In the presentation, I will shortly review the notion of abstract shapes. I will then discuss several applications of this concept, including a highly effective filtering method when working with structural classes of RNA described by covariance models, such as provided by the Rfam data base.

Secbase and DrugscoreMaps - useful tools for structure-based drug design

Dozent	Oliver Koch, The Cambridge Crystallographic Data Centre, Cambridge, UK
Ort	ZBH, Raum 16
Zeit	17. Apr 2008 16:00

Inference of Biochemical Networks from High-Throughput Data

Dozent	Dr. Lars Kaderali, University of Heidelberg, Viroquant Research Group Modeling, Heidelberg
Ort	ZBH, Raum 16
Zeit	03. Apr 2008 16:00

Developments in experimental technologies such as DNA microarrays and rtPCR experiments render the quantitative measurement of expression levels of a large number of genes feasible, and make the acquisition of time series data possible. Such data can be used to learn detailed quantitative models describing the transcriptional dynamics of a biological system under consideration. These models can then be used to run simulations, and to analyze the dynamic behavior of the transcriptional network model under various conditions. This inference of models from data requires the specification of appropriate model classes, and appropriate methods for the estimation of model parameters. When using quantitative models based on differential equations, a typical problem is that the number of model parameters to be estimated usually far exceeds the number of measurements available, leading to underdetermined optimization problems. This effect is further amplified by noisy data and poses a serious problem for network inference. We describe an inference approach for gene regulatory networks from time series gene expression data which is based on Bayesian statistics. By combining differential equations models with dynamical bayesian networks, we combine the advantages of these two modelling approaches. The problem of overfitting is addressed by integrating biological knowledge in the inference process in terms of suitably chosen prior probability distributions over network parameters. A specifically designed prior distribution over regulation strengths is used that drives the inference to biologically plausible sparse networks. For parameter estimation, we evaluate the model posterior distribution. This approach enables the method to cope with the typical setting of only few time points to estimate a large number of model parameters and thus makes the method particularly tailored to the task of quantitative modeling from typical real-world experimental datasets.

Diffraction studies of structural states and motions along enzymatic reaction trajectories

Dozent	Dr. Hans-Dieter Bartunik, Max-Planck-Arbeitsgruppen für strukturelle Molekularbiologie HamburgMPG-ASMB c/o DESY, Arbeitsgruppe Proteindynamik
Ort	ZBH, Raum 16
Zeit	07. Feb 2008 16:00

A number of enzymes have been found to carry out productive catalytic reactions even under the molecular packing conditions in crystals. External initiation of reactions, trapping of intermediate states at cryo-temperatures and X-ray diffraction measurements have delivered sequences of three-dimensional structures at high resolution and provided a basis for investigating structure-function relationships of interactions with ligands and conformational changes along reaction trajectories. In addition to experimental results, methods are discussed that may provide further information on correlations between motions. Experimental applications of such methods including diffuse scattering and fast time-resolved diffraction, have until now been limited to model systems. The future use of PETRA III and the XFEL at DESY will open the way to new developments in the analysis of transient motions in proteins and protein complexes.

Software libraries for algorithmic bioinformatics

Dozent	Prof. Dr. Knut Reinert, Algorithmische Bioinformatik, Institut für Informatik, Freie Universität Berlin
Ort	ZBH, Raum 16
Zeit	17. Jan 2008 16:00

In this talk I will present the two of the main projects of the Algorithmic Bioinformatics Group at the FU Berlin, SeqAn and OpenMS. SeqAn is a new open source C++ library for the analysis of sequences with a focus on biological applications. It offers various efficient algorithms and data structures for holding, manipulating, searching and indexing sequence data. SeqAn applies a unique generic design that guarantees high performance, generality, extensibility, and integration with other libraries. It is easy to use and simplifies the development of new software tools with a minimal loss of performance (www.seqan.de) OpenMS is a C++ framework for data analysis applications in the field of shotgun proteomics. It offers data structures and algorithms for data preprocessing, protein identification, quantitation, alignment of samples, statistical analysis and visualization. Most functionality is already turned into tools which can be combined into powerful analysis pipelines (www.OpenMS.de).

Kinetic Modeling and Metastability of Dynamical Systems with Applications to Biomolecular Dynamics

Dozent	Dr. Frank Noe, Computational Molecular Biology, FU Berlin, BerlinDFG Research Center Matheon
Ort	ZBH, Raum 16
Zeit	13. Dec 2007 16:00

Biomolecules are high-dimensional stochastic dynamical systems with some particular interesting properties, including metastability, i.e. the ability - or curse - to remain in parts of state space for long times before leaving them. Here, I will give an overview how to model the complex stochastic dynamics with Markov- or Master-Equation models. I'll go into some methodological detail, such as spectral properties, optimal estimation of kinetic models and how to coarse-grain the kinetic description into metastable states. I'll show some examples how these methods can be applied to study conformational changes and folding of biomolecules.

Speciation in phytophagous insects - the role of host plants versus geography

Dozent	Prof. Dr. Susanne Dobler, Zoologisches Institut und Museum, Universität Hamburg
Ort	ZBH, Raum 16
Zeit	06. Dec 2007 16:00

Phytophagous insects represent one of the major radiations on earth and are highly amenable to studying speciation mechanisms. Most species are specialized for a few host plants and related taxa often differ in their host use. Speciation in this group is therefore often thought to be associated with host switches. In this work, we discuss several species of chrysoelid beetles and question how far ecological experiments, in combination with population genetic analyses, can trace the causes and courses taken of speciation mechanisms in these groups.

Schemata in Reaction-Diffusion Systems

Dozent	Prof. Dr. Andreas G. Fleischer, University of HamburgDepartment of Biology, Informatikum
Ort	ZBH, Raum 16
Zeit	15. Nov 2007 16:00

It is a principle of biological systems that the complexity of possible states has to be reduced to obtain an efficient functional structure. These states may be of a different nature. You may consider the protein space of a molecule, interactions between molecules in a cell, interactions in a neural network or interactions between biological agents. During the evolutionary process a functional structure has to develop a sufficient fitness or it will end. A genetically determined protein may appear to evolve slowly. However, the neurophysiological system of an agent has to react immediately on demands of the environment. Thus, it is a challenging question, what these different levels of building a functional adaptive structure have in common. An attempt will be put forward about combining schema theory with the approach of reaction-diffusion systems. Experimental results of movement control are evaluated on the basis of such a model. Reaction-diffusion systems provide a rich variety of activity patterns. Different types will be discussed. How do biological patterns emerge and how does the behaviour of an agent lead to a schema. The protein space can be considered to be like a picture, each possible protein and each pixel to represent a point in a high dimensional space, respectively. Such a picture may be disturbed or noisy. However, if besides others a certain picture is stored in an association matrix a recurrent attractor net may reconstruct the original picture. However, reaching a stable attractor point, that is a static picture, would be boring. But on the basis of specific association matrices a recurrent system can be constructed which provides a sequence of consecutive states. This leads to the hypothesis that complex movements are based on such consecutive states of motion primitives. From a modelling point of view, it should be possible to develop a neural network which allows for implementing a schema, i.e. a motor engram, which interacts with feedback control in an iterative manner. Neural population coding is supposed to represent a suitable modelling approach. Considering two interacting neural layers of a reaction-diffusion system they may generate an attractor state and a moving ‘bubble’, respectively. The bubble running along a motor engram activates a set of muscles and causes the arm to move towards a target. Modelling results are evaluated against the background of experimental data obtained from hand movements. They support the hypothesis of a motor schema. It determines the trajectory of intercepting a moving target. It comes up, that in conjunction with the motor schema the transition probability of the bubble elicits a new motion primitive and its direction during the next step. The Gestalt of the movement trajectory is determined by a schema which is adapted to the requirements of the task.

Efficient algorithms for the computational design of tiling arrays

Dozent	Dr. Alexander Schliep, Max-Planck-Institute for Molecular Genetics, Berlin
Ort	ZBH, Raum 16
Zeit	08. Nov 2007 16:00

Genomic tiling arrays are a type of DNA microarrays which can simultaneously investigate complete genomes with respect to a wide range of experimental questions such as mRNA expression, comparative genomic hybridization or ChIP-Chip. The design or selection of suitable oligonucleotide probes for such arrays however is computationally difficult if features such as oligonucleotide quality and repetitive regions are to be considered. We formulate the minimal cost tiling path problem (MCTPP) for the selection of oligonucleotides from a set of candidates. Our formulation incorporates experimental constraints with respect to obligatory oligonucleotides, specific regions of interest and tradeoffs between hybridization parameters, probe quality and tiling density. We show MCTPP to be equivalent to a shortest path problem which, for the unconstrained case, can be solved in linear time using Monge theory and in log-linear time in the general case due to the properties of shortest path in MCTPP instance graphs using an efficient implementation of Dijkstra's algorithm. Hence we can compute globally optimal tiling paths from millions of candidate oligonucleotides on a standard desktop computer. We will also show results from experiments on M. smegmatis.

The Role of Computational Chemistry in Lead Identification and Optimization at Bayer HealthCare

Dozent	Dr. Alexander Hillisch, Director, Medicinal Chemistry, Bayer HealthCare AG, Wuppertal
Ort	ZBH, Raum 16
Zeit	01. Nov 2007 16:00

Today, Computational Chemistry is applied routinely in all early phases of drug discovery in pharmaceutical industry research. Especially lead identification, hit-to-lead and lead optimization processes are supported. The impact of this technology on real drug discovery successes is not equally distributed over the different phases and strongly dependent on the amount and quality of experimental data that are available. In my talk I will outline how Computational Chemistry contributes to drug discovery at Bayer HealthCare. The structure, tasks and responsibilities of the Computational Chemistry team in Wuppertal will be explained. I will give examples for the application of various methods such as ligand and protein structure-based drug design approaches. Finally, I will also explain the development of in silico ADMET methodologies and the application of these programs to drug discovery projects. Aspects for the future development of Computational Chemistry in drug discovery will be discussed.

Prediction of Transcription Factor Binding Sites Using Variable Order Markov Models and Bayesian Networks

Dozent	Prof. Dr. Stefan Posch, Martin-Luther-UniversitätHalle-Wittenberg, HalleInstitut für Informatik
Ort	ZBH, Raum 16
Zeit	25. Oct 2007 16:00

Mapping Molecular Landscapes inside Cells by Cryoelectron Tomography

Dozent	Dr. Stephan Nickell, Max-Planck-Institute of Biochemistry, Martinsried
Ort	ZBH, Raum 16
Zeit	11. Oct 2007 16:00

Cryo-electron tomography is an emerging imaging technique that has unique potential for molecular cell biology. At the present resolution large supramolecular structures can be studied in unperturbed cellular environments and, in the future, it will become possible to map molecular landscapes inside cells in a more comprehensive manner. 'Visual proteomics' aims to complement and extend mass-spectrometry-based inventories, and to provide a quantitative description of the macromolecular interactions that underlie cellular functions.

Diplomvortrag: Implementation of a metabolic network of insulin resistance

Dozent	Anja Schoettler, ZBH
Ort	ZBH, Raum 16
Zeit	27. Sep 2007 16:00

Worldwide more than 170 million individuals suffer from diabetes mellitus and a further growth of almost 50% is predicted for the year 2010. 90% of these individuals have type 2 diabetes which is closely associated with insulin resistance and obesity. Since obesity and type 2 diabetes are often associated with dysregulations in hepatic lipid synthesis, a perfect understanding of the regulation of metabolic pathways regarding fatty acid synthesis, fatty acid modification, beta-oxidation and transcription factors regulating lipid homeostasis in liver would be useful to identify potential therapeutic drugs. Therefore in this study these models were created with the SBML-compliant process diagram editor CellDesigner based on information found in the literature. Furthermore data from Eli-Lilly and the University Hospital Hamburg-Eppendorf was available for various transcription factors, enzymes, hormones and fatty acids of three mouse groups fed in different ways in order to validate the models.

Prediction the Reading Direction of Structured RNAs in Multiple Sequence Alignments

Dozent	Kristin Reiche, University of Leipzig, Dept. of Computer ScienceBioinformatics
Ort	ZBH, Raum 16
Zeit	12. Jul 2007 16:00

Genome-wide screens for structured ncRNA genes in mammals, urochordates, and nematodes have predicted thousands of putative ncRNA genes and other structured RNA motifs. A prerequisite for their functional annotation is to determine the reading direction with high precision. We present here a support vector machine (SVM) that reliably classifies the reading direction of a structured RNA from a multiple sequence alignment and provides an improvement in classification accuracy over previous approaches.

Motif Detection under RNA Structural Constraints

Dozent	Prof. Dr. Rolf Backofen, Albert-Ludwigs-Universität FreiburgInstitut für Informatik / Lehrstuhl für Bioinformatik
Ort	ZBH, Raum 16
Zeit	11. Jul 2007 16:00

In the past years, the common view of RNA has evolved from a more or less boring intermediate in protein translation to a very important player in cell regulation. The changed role was primarily the result of the detection of thousands of non-coding RNAs with many different regulatory functions in post-translational regulation. The detection of new functional RNAs requires new comparative methods for motif detection since the RNA sequence is much less conserved than the RNA structure. Hence, purely sequence-based methods for finding RNA-motifs such as multiple sequence alignment will fail, and we will discuss various approach (e.g. MARNA, LocaRNA) to sequence-structure alignment developed in our group. Beside finding structural RNAs, we will also discuss the importance of the RNA secondary structure for the regulation of the binding efficiency of RNA-binding proteins like splice enhancers/silencers.

Biomolecular Computing Models for Graph Problems and Finite State Automata

Dozent	Israel M. Martinez-Perez, Technische Universität Hamburg-Harburg
Ort	ZBH, Raum 16
Zeit	05. Jul 2007 16:00

DNA computing is a discipline that aims at harnessing individual molecules at the nanoscopic level for computational purposes. Given its vast parallelism and high-density storage, DNA computing approaches were originally employed to solve many combinatorial problems. Unfortunately, all these high expectations were rapidly truncated by the time consuming and error-prone laboratory operations along with the scaling issue of NP-complete problems. Now, the trend in DNA computing has shifted to autonomous DNA computing models, focussing on biomolecular problems that can barely be tackled by standard machines. The purpose of this talk is to present our contribution to this rapidly evolving field of research. Specifically, we will describe the first molecular-scale autonomous solution of two famous hard combinatorial problems in linear time, the decision Hamiltonian path and maximum clique problems. In addition, we will introduce two models for finite state machines: the sticker automata and computational gene models. The latter model might be able to detect and correct aberrant molecular phenotype given by mutated genetic transcripts. We will provide a 'proof of principle' for our model in vitro.

Mapping structural spaces: Protein Structure Similarity Clustering (PSSC) and Structural Classifcation of Natural Products (SCONP)

Dozent	Stefan Wetzel, Max-Planck-Institute for Molecular Physiology, DortmundDepartment IV - Chemical Physiology
Ort	ZBH, Raum 16
Zeit	28. Jun 2007 16:00

Recent years have seen a large increase of interest in charting both, chemical and protein space. While chemical spaces have been mainly characterized by molecular properties, protein space was organized primarily by protein function and protein families. However, in both cases, structure might be a more intuitive criterion to use in such approaches. One approach to structure based charting of chemical space has been presented recently. The method is based on molecular scaffolds which are dissected in a stepwise manner. Ordering by substructure relationships and a defined rule set yields one-to-many parent-child-relationships which can then be displayed as tree-diagrams. These diagrams can be used for a variety of purposes, e.g. data mining, HTS evaluation and the like. Concerning the protein space, a method based termed protein structure similarity clustering (PSSC) is employed. This concept explores the similarity around the binding site to find new ligands for orphan targets as well as additional targets for known actives. In our approach, we construct ligand-sensing cores which represent spherical sections of the proteins binding sites. Subsequent comparisons of sub-folds using 3D structural alignment methods lead to a similarity-based clustering of the ligand-sensing cores. These clusters and their intercluster relationships can then be used to map the structurally known protein space as well as to find new starting points for inhibitor development. Both approaches have provided first successful examples of application in the quest for small molecule enzyme inhibitors. Although a genereal applicability is difficult to prove the fusion of charting chemical and protein space based on structure is prone to benefit chemical genomics approaches and early stage drug discovery.

One-man-band in CADD: experiences in big pharma

Dozent	Dr. Andrea Zaliani (ZBH Center for Bioinformatics)
Ort	ZBH, Raum 16
Zeit	14. Jun 2007 16:00

The seminar will provide an overview of typical challenges met by me in a Computer-Aided Drug Design Laboratory of a big Pharmaceutical Industry during the past 6 years. The resulting approach is always a mix of several known techniques applied to practical cases. Modelling solutions as well as chemoinformatic and chemico-physical calculations will be seen as a unique tool to support discovery projects. Project's success and impact thereon has been, in my industrial experience, more stress than an exhaustive scientific validation, so that the correlated risk for the scientist to become a technologist and stay at the border of important findings without time and/or resources to produce important scientific papers is very high. Several examples from ADME predictions, virtual screening, conformational analysis and protein cavity detection will be given.

Coordinate uncertainties and protein structure comparison

Dozent	Dr. Thomas R. Schneider, EMBL Hamburg
Ort	ZBH, Raum 16
Zeit	31. May 2007 16:00

In contrast to sequence data where a nucleotide entry can only be right or wrong, the precision in the location of an atom in a crystal structure can vary over several orders of magnitude: while the position of an atom in a rigid region of a protein giving diffraction data to high resolution may be known to within 0.01 Ã…, for an atom in a flexible region of a poorly diffracting protein, the coordinate error may reach more than 1.0 Ã…. Reflecting this fact, we are developing methods for structure analysis which take coordinate uncertainties explicitly into account. We have developed a novel similarity measure for assessing whether different models of a molecules are identical within error or show significant differences. Application of the same similarity measure as a score for subsets of atoms allows to find regions of a molecule that are conformationally invariant (or behave as a rigid body) with respect to an ensemble of structural models. I shall describe the background of our methods and discuss their application to an ensemble of more than 50 crystal structures of protein kinase A. For this case, we were able to automatically partition the ensemble into groups of models corresponding to different types of ligands binding to the protein. Further analysis revealed several rigid bodies in the kinase that relate to the regulation of the enzyme. We are currently extending our methods to the analysis of close homologues and some initial results will be shown.

A protein modelling pipeline and composite scoring function for model quality assessment

Dozent	Pascal Benkert, CUBIC, Institute of Biochemistry, CologneCologne University Bioinformatics Center
Ort	ZBH, Raum 16
Zeit	10. May 2007 16:00

Protein structure prediction is still a challenging discipline. We have built a comparative modelling pipeline and will discuss its performance at the most recent CASP. The pipeline consists of a PDB-BLAST-like protocol for parent detection, a profile-profile alignment step, manual model building including a semi-automatic loop modeling procedure and a composite scoring function for final model selection. The first part of the talk will cover the modelling pipeline and its performance at CASP7. In the second part of the talk the QMEAN (Qualitative Model Energy Analysis) scoring function will be presented. QMEAN consists of 5 terms. In the course of protein structure prediction usually a considerable number of alternative models are produced from which the final model has to be selected. Thus, a scoring function for the identification of the best model within an ensemble of alternative models is a key component of most protein structure prediction pipelines. QMEAN, consists of 3 statistical potential terms investigating the major geometrical aspects of protein structures: the local geometry is analyzed by a new kind torsion potential over 3 consecutive amino acids. A secondary structure-specific distance-dependent pairwise potential at the residue-level is used to assess long-range interactions. A solvation potential investigates the residues solvent accessibly. Additionally, two simple terms describing the agreement of predicted and observed secondary structure and solvent accessibility, respectively, are integrated.

Diplomvortrag: Protein-Profiling stratifizierter Epithelien mittels Bildverarbeitung fluoreszent gefärbter Gewebeschnitte

Dozent	Thora Pommerencke, Zentrum für Bioinformatik
Ort	ZBH, Raum 16
Zeit	26. Apr 2007 16:00

Computational Chemistry in the Context of industrial ultra High Throughput Screening (uHTS)

Dozent	Dr. Christian Kirchhoff, Evotec AG, Hamburg (www.evotec.com)Molecular Modelling and Discovery Informatics
Ort	ZBH, Raum 16
Zeit	19. Apr 2007 16:00

Functional motion in protein crystal structures

Dozent	Dr. Victor S. Lamzin, Centre for Computational Biology, EMBL Hamburg
Ort	ZBH, Raum 16
Zeit	12. Apr 2007 16:00

X-ray crystallographic studies have helped to elucidate reaction mechanism of many biocatalysts, yet a handful of questions still remained unanswered. Among these are the pKa of the active site residues, conformational adaptation to the conditions and the nature of the incoming ligands and the protonation state of residues relevant for the function. We have developed a method for extracting directional motion of atoms, residues or groups of residues from high resolution crystallographic structure analysis. The tool takes into account the anisotropic behaviour within the protein and the assignment of regions of concerted motion. Results on directional motion on a few selected enzymes and their relevance to biological function will be presented. Future plans on enhancing the method will be discussed.

Diplomvortrag: Image Alignment for Time-Series Analysis of Protein Crystallisation Trials

Dozent	Gundolf Schenk, Zentrum für Bioinformatik
Ort	ZBH, Raum 16
Zeit	01. Feb 2007 16:45

Diplomvortrag: Flexible combination of filters for oligodesign

Dozent	Stefan Bienert, Zentrum für Bioinformatik
Ort	ZBH, Raum 16
Zeit	01. Feb 2007 16:00

Beyond Proteins: The Genome's Non-Coding Majority

Dozent	Prof. Peter F. Stadler, Universität LeipzigBioinformatik, Institut für Informatik
Ort	ZBH, Raum 16
Zeit	25. Jan 2007 16:00

In contrast to the fairly reliable and complete annotation of the protein coding genes in the human genome, comprehensive information for non-coding RNAs is still lacking. Both quality and coverage of ncRNA annotation is not comparable to that of protein coding genes. The ENCODE Pilot Project recently revealed, however, that the (non-repetitive part) of the Human Genome is transcribed almost in its entirety, giving rise to a complex mosaic of multiple transcripts with different fates for for most loci. The majority of these products have little or no protein coding capacity. Recently, we completed a comparative screen of vertebrate genomes for structural non-coding RNAs, which evaluates conserved genomic DNA sequences for signatures of structural conservation of base pairing patterns and exceptional thermodynamic stability. We find more than 30000 structured RNA elements, almost 1000 of which are conserved across all vertebrates. In contrast, the number of detectable structured ncRNAs in invertebrate genomes appear is at least one order magnitude smaller, with an even smaller list of candidates in unicellular organisms. In this presentation I will focus on approaches towards annotating the plethora on ncRNAs and ncRNA candidates, and first studies into the evolution of these transcripts at genome-wide scales.

Diplomvortrag: Effiziente Berechnung von Subgraphisomorphismen für multiple chemische Mustergraphen

Dozent	Kai Stierand, Zentrum für Bioinformatik
Ort	ZBH, Raum 16
Zeit	22. Jan 2007 17:15

Exact determination of most parsimonious trees in sequence space via partitioning techniques

Dozent	Prof. Dr. Hans-Jürgen Bandelt, Mathematik, Universität Hamburg
Ort	ZBH, Bundesstrasse 43, Raum 16
Zeit	11. Jan 2007 16:00

Since a DNA sequence space is a product space (a power of a 5-letter alphabet), one can aim at solving the maximum parsimony (MP) problem by partitioning the sequences into subsequences for which the MP problem is solved separately. If these solutions can be aggregated to a feasible global solution, then the MP problem for the full data is solved. Mike Hendy and his colleagues have proposed such a strategy some quarter of a century ago and successfully used it to solve the MP problem for very small-scale amino acid data at the time. More recently Holland et al. (MBE 2005) have revived this approach by providing a computer-assisted search for good partitions that could lead to successful solutions. These approaches would then deliver a tree that is provably MP tree. However, they cannot guarantee to have found all MP trees. To circumvent this problem, I would exclusively work with most parsimonious realizations (MPRs) of MP trees (which are mathematically speaking exactly the Steiner trees). Then there are two trivial situations where all MPRs of MP trees can be listed exhaustively: (1) the data set is representable by a connected 1-step network (whence the MPRs of MP trees coincide with the minimum spanning trees); (2) the distance matrix for the data set is a tree metric that admits a (necessarily homoplasy-free) realization in sequence space. By suitably combining both instances, a more general theorem can be proven under which all MPRs of MP trees can be determined efficiently. Striving for partitions each admitting such an exact approach one only needs to take care that separate solutions are compatible and can thus be aggregated to a global tree. All MPRs of MP trees for the original problem can then be retrieved via the compatible combinations of the separate MPR solutions. Theoretically, there can be arbitrarily large data sets that are indecomposable, that is, cannot be partitioned into proper parts that together yield a global solution. From the practical point of view, it seems that “reasonable” data sets, i.e. data sets that do not contain too much homoplasy (so that MP solutions may still be meaningful), can even be partitioned by hand for fewer than 100 sequences. This can be demonstrated with a real data set comprising the coding-region substitutions in 53 human mitochondrial DNAs.

Combining cross-linking, mass spectrometry and molecular modelling to determine the structures of proteins

Dozent	Prof. Thomas Huber, The University of Queensland, Brisbane, AustraliaSchool of Molecular and Microbial Sciences
Ort	ZBH, Raum 16
Zeit	21. Dec 2006 16:00

Structural genomics initiatives attempt to express, isolate and determine the structures of large numbers of proteins in high-throughput factories and thus minimise the cost per newly determined protein structure. Despite these efforts, to date the determination of a protein structure remains a rather costly endeavour. Chemical cross-linking in combination with mass spectrometry allows one to obtain cheaply and swiftly limited inter-residue distance information in proteins. While such limited distance information is not sufficient by itself to fully determine the structure of a protein, when combined with computational prediction approaches it has shown great potential for low-cost, high-throughput protein structure determination. Here, I will present our approach to combine distance constraints from chemical cross-linking experiments with molecular modelling to determine the structures of protein complexes and multi-domain proteins. Further, I will discuss experimental and algorithmic implementations, and illustrate this approach on the recent structure determination of murine acyl-CoA thioesterase, an important cellular enzyme responsible for catalysing the hydrolysis of acyl-CoA.

ON THE EVOLUTION OF COMPLEX BIOLOGICAL NETWORKS

Dozent	Prof. Erich Bornberg-Bauer, The Westfalian Wilhems University of MünsterDivision of Evolutionary Bioinformatics, Institute for Evolution and Biodiversity
Ort	ZBH, Raum 16
Zeit	14. Dec 2006 16:00

Explaining the emergence of complex biological entities such as organs, networks and multi-domain protein is at the heart of biology since it requires a profound understanding of evolution. While researchers have repeatedly tried to explain the evolution of such complex units by postulating that large scale duplications and selection for certain topologies per se has led to their current complexity, we show that the classical model explaining biological innovation is sufficient to understand the emergence of complex structures: both, the interaction networks of transcription factors families and the domain arrangements of proteins may evolve via a combination of the well-known processes of gene duplication, neo/sub-functionalisation and gene loss. However, we also find that quite different mechanisms have shaped the evolution of different protein families such that caution is warranted in postulating broad-brushed theories on molecular evolution.

Gene regulation code

Dozent	Dr. Alexander Kel, BIOBASE GmbH, Wolfenbüttel
Ort	ZBH, Raum 16
Zeit	07. Dec 2006 16:00

Regulation of gene expression is accomplished through binding of transcription factors (TFs) to distinct regions of DNA (TF binding sites, TFBS), and, after anchoring at these sites, transmission of the regulatory signal to the basal transcription complex. Gene regulation code which would enable a “translation” of DNA regulatory sequences into regulatory function they possess are still unclear. Some of TFs are specific for a particular tissue, a definite stage of development, or a given extracellular signal, but most transcription factors are involved in gene regulation under a rather wide spectrum of cellular conditions. It is clear by now that combinations of transcription factors rather than single factors drive gene transcription and define its specificity. Dynamic function-specific complexes of many different transcription factors, so called enhanceosomes are formed at gene promoters and enhancers controlling gene expression in a specific manner. At the level of DNA, the blueprints for assembling such variable TF complexes on promoter regions may be observed as specific combinations of TFBS located in close proximity to each other. We call such structures “Composite Modules (CMs)”. The multiplicity of cellular conditions in which eukaryotic genes should be expressed is the cause of polyfunctionality of the structure of their transcription regulatory regions. We believe that this polyfunctionality is governed through alternative CMs. In the lecture I will present a “fuzzy puzzle” model of the gene regulation code, which based on the principle of encoding multiple regulatory messages in the same DNA sequence. The structure of regulatory sequences on one hand and the specific features of transcription factors on the other provide a possibilities to encode several regulatory programs within one regulatory region. Such structure allows receiving and integration of multiple regulatory signals through reuse of the same DNA sequence of gene promoters and enhancers. We developed a novel tool, Composite Module Analyst (CMA) (Kel et al., 2006), that applies a novel approach for defining promoter models based on composition of single transcription factor binding sites as well as their pairs located inside local regulatory domains (corresponding to enchancer/silencer subregions). We use the genetic algorithm technique and utilize a multicomponent fitness function for defining the function specific composite modules in promoters of genes co-regulated in specific conditions (tissue, organ, stage of development, cell cycle, response to particular extracellular signal). We applied the CMA tool to analyze promoters of tissue/organ specific genes (based on the organ ontology Cytomer® and define Composite Modules characteristic for promoters of genes specifically (or abundantly) expressed in those tissues, which might be considered as the components of the gene regulation code. We think that “fuzzy puzzle” principle of regulation is the result of genome evolution of multicellular organisms that shall overcome evolutionary bottlenecks caused by the requirement of multiple ontogenetic programs to be encoded in a single genome. 1. Kel,A, Konovalova,T, Waleev,T, Cheremushkin,E, Kel-Margoulis,O, Wingender,E. (2006) Composite Module Analyst: a fitness-based tool for identification of transcription factor binding site combinations. Bioinformatics. 22, 1190-1197 (2006).

Computational Structural Proteomics and Inhibitor Discovery

Dozent	Prof. Ruben Abagyan, The Scripps Research Institute, La Jolla, CA, USADepartment of Molecular Biology
Ort	ZBH, Raum 16
Zeit	22. Nov 2006 10:00

Rapid advance of structural proteomics calls for the development of new methods to for predicting structural changes, association, function, as well as improving methods for structure based molecular design. We have developed methods for predicting a functional map of a protein, accurate docking and screening of compounds, and, structure prediction by global energy optimization, e.g. characterizing mutants and SNPs, homology modeling, protein or peptide docking, and accurate loop prediction, and ligand-induced receptor changes. Two stories in which novel compounds were discovered through the ligand-guided receptor pocket modeling followed by virtual screening of large compound libraries, were presented. First, we developed models of the androgen receptor in an antagonist-bound conformation. These models were used to discover computationally the secondary activity of anti-psychotic drugs. These drugs were then chemically altered and re-purposed to loose their binding to the serotonin and dopamin receptors, and improve their anti-androgen properties. Second, in a collaboration with the David Lomas lab, Cambridge, we identified the first small molecules to inhibit pathological polymerization of an alpha1-antitrypsin mutant which is the most common genetic cause of a lethal liver disease in childhood. The structural model used to identify the first polymerization inhibitors was built on using the ICM simulation-based flexibility analysis and multiple pocket conformation generator.

Auto-Rickshaw: An automated Crystal Structure Determination Platform as efficient tool for validation of an X-ray diffraction experiment

Dozent	Dr. Santosh PanjikarEMBL Hamburg
Ort	ZBH, Bundesstr. 43, Raum 16
Zeit	16. Nov 2006 16:00

We present an automated crystal structure determination platform, Auto-Rickshaw[1]. It contains several distinct computer coded decision-makers, which invoke a variety of macromolecular crystallographic programmes/programme packages during the structure determination process. A large number of paths are encoded in the system and the optimal path is selected by the decision-makers as the structure solution evolves. The primary aim of the pipeline is to validate the crystallographic experiment at the synchrotron site while the crystal is still at or near the beamline. Therefore, the system has been optimized for speed, so that typically within a few minutes, it is apparent whether or not the already collected data will be of sufficient quality to allow successful structure determination. The platform has been installed on a 16-processor Linux cluster and is remotely accessible to the beamline users via a web-server (www.embl-hamburg.de/Auto-Rickshaw). An overview of the Auto-Rickshaw pipeline with its architecture, functionality, some examples and the way this platform is used as a feedback system for X-ray data collection or validation of X-ray experiment, will be discussed. [1] Panjikar, S., Parthasarathy, V., Lamzin, V. S., Weiss, M. S.Tucker, P. A. Acta Cryst. 2005, D61, 449.

Privileged structures for serine proteases - Docking and directed design in the protein binding pocket

Dozent	Christoph Gerlach, Lilly Foschung GmbH, Hamburg/ Institut für Pharmazeutische Chemie, Philipps-Universität Marburg
Ort	ZBH, Raum 16
Zeit	09. Nov 2006 16:00

Successful automated computational ligand design requires rigorous validation of predicted binding geometries. Assessing the predictive power of our inhibitor design tools we selected trypsin and thrombin as well-established model cases. Via the analysis of ligand binding using ITC we factorize binding affinity into thermodynamic contributions. We start with a privileged ligand scaffold well-suited to address key interactions of the recognition pattern shared by serine proteases. Binding modes of these privileged scaffolds are generated by docking. The best ranked solutions are optimized by local searches based on knowledge-based potentials derived from CSD. For individual members of the series the crystal structure and the binding-affinity towards both enzymes are determined. Furthermore different approaches for library design are tested to construct a focused library with potential leads. Through syntheses, enzymatic assay and crystal structure analysis the most promising library members are further characterized.

Industrial Bio- and Chemoinformatics

Dozent	Dr. Friedrich Rippmann, Merck KGaA, DarmstadtGP RD / Tech / Bio- and Chemoinformatics
Ort	ZBH, Raum 16
Zeit	02. Nov 2006 16:00

Diplomvortrag: Indexstrukturen für die Feature-Tree-basierte Suche nach ähnlichen Molekülen

Dozent	Robert Fischer, ZBH
Ort	ZBH, Bundesstr. 43, Raum 16
Zeit	26. Oct 2006 16:00

Pleiomorphism of Supramolecular Assemblies

Dozent	Prof. Willy Wriggers, University of TexasScience Center at Houston, USA
Ort	ZBH, Raum 16
Zeit	24. Oct 2006 16:30

Advances in modern biophysics depend on an understanding of fundamental cellular processes, most of which involve the actions and interactions of large biomolecular assemblies. Examples currently studied by us are the actomyosin complex in muscle, molecular chaperones, the ribosome complex, apoptosis regulating complexes, virus capsids, and the RNA polymerase complex. Such systems have been called “macromolecular machines” to characterize the complexity and precision of their highly coordinated moving parts. Ideally, such large structures are visualized at the atomic level, but medium resolution data in the 5-25Å resolution range has become increasingly ubiquitous. Our goal is to characterize the structure and functionally relevant motions of these relevant systems across the length scales. The conformational change and biochemical mechanism can be modeled at near-atomic detail within a molecular dynamics simulation framework with amplified collective coordinates.

Probability, networks and evolution

Dozent	Prof. Mike Steel, Biomathematics Research Center, ChristchurchNew Zealand
Ort	ZBH, Raum 16
Zeit	12. Jul 2006 16:00

Modelling with high-level Petri nets

Dozent	Dr. Daniel Moldt / Dr. Heiko Rölke, Universität HamburgTheoretische Grundlagen der Informatik, Hamburg
Ort	ZBH, Raum 16
Zeit	06. Jul 2006 16:00

Vortrag auf Englisch) Carl Adam Petri stellte 1962 mit seiner Dissertation grundlegende Konzepte zur Beschreibung verteilter und nebenläufiger Systeme bereit. Schnell etablierte sich unter dem Begriff Petrinetze eine Forschungsrichtung, die zentrale Themen der Informatik behandelt. Bis heute werden die Anwendungsfelder der universellen Modellierungstechnik ständig ausgeweitet. Im Vortrag werden die grundsätzlichen Beschränkungen einfacher Petrinetze thematisiert. Einige aktuell verwendete Erweiterungen werden ebenso diskutiert, wie die Verwendung der Ergebnisse aus dem Bereich der Petrinetze in der Informatik. Inhalt des ersten Teils des Vortrags ist daher die Vorstellung einer grundsätzlichen Modellierungs-perspektive auf Basis von Petrinetzen und die darauf aufbauende softwaretechnische Verwendung in dem Modellierungsansatz Paose (Petrinetz- und AgentenOrientierte Software-Entwicklung). Im zweiten Teil des Vortrags wird die Agentensystemarchitektur Mulan (MULtiAgentenNetze) vorgestellt. Mulan ist ein vollständig in Referenznetzen spezifiziertes Rahmenwerk für Agenten und Agentensysteme, das nach den Standards der FIPA (Foundation for Intelligent and Physical Agents) entwickelt wurde. Mulan-Agenten sind zur Laufzeit beliebig adaptierbar und können insbesondere ihre eigene Ablaufsteuerung selbständig beeinflussen (autonome und selbstmodifizierende Agenten). Da ein Mulan-Agentensystem letztlich nur aus Agenten besteht, zieht sich die Eigenschaft der Selbstmodifikation auch in das Gesamtsystem hinein; dieses hat die Möglichkeit der Selbstorganisation.

Modellierung von Metall- Wechselwirkungsgeometrien für das Protein-Ligand Docking Problem

Dozent	Birte Seebeck, ZBH
Ort	ZBH, Raum 16
Zeit	04. Jul 2006 16:30

Technische Biokatalyse - Ein Grenzgang zwischen Chemie und Biologie

Dozent	Prof. Dr. Andreas Liese, Institute of Technical Biocatalysis, Technical University of Hamburg-Harbrug
Ort	ZBH, Raum 16
Zeit	29. Jun 2006 16:00

Proteine finden heutzutage immer weitere Anwendungen als Biokatalysatoren in der chemischen Industrie zur Erzeugung von Feinchemikalien, Agrochemikalien oder Bulk-chemikalien. Hierbei ist die Natur hoch spezialisiert, da sie es schafft, die kompliziertes-ten Reaktionen mit maximaler Stereo- und Regioselektivität durchzuführen. Des Weite-ren hat die Natur ihre Katalysatoren über tausende von Jahren optimiert und setzt sie höchst effektiv in Reaktionssequenzen ein. Im Bereich der Technischen Biokatalyse werden biokatalytische Prozesse für die indus-trielle Anwendung (Weiße Biotechnologie) entwickelt. In diesem Vortrag wird gezeigt, dass sowohl für Chemo- als auch Biokatalysatoren die gleichen Gesetzmäßigkeiten gel-ten. Für die Optimierung eines Verfahrens gibt es zwei Seiten: Das Protein-Engineering und das Substrat-Engineering. Am Beispiel der enantioselektiven C-C-Bindungsknüpfung und asymmetrischen Reduktionen werden die Unterschiede und Gemeinsamkeiten der chemokatalytischen und biokatalytischen Synthese diskutiert.

Reconstruction and analysis of genome-wide metabolic and regulatory networks

Dozent	Prof. An-Ping Zeng, Technical University of Hamburg-HarburgInstitute of Bioprocess and Biosystems Engineering, Hamburg and GBF-German Research Center for Biotechnology, Braunschweig
Ort	ZBH, Raum 16
Zeit	22. Jun 2006 16:00

Recent advances in genome sequencing and functional genomics now make it possible to reconstruct large-scale biological networks at different molecular levels. Studies of these genome scale networks have revealed several intrinsic structural and functional properties (e.g. power law and bow-tie structure) of biological systems. In this presentation, firstly a brief overview will be given on the reconstruction and structural analysis of metabolic and regulatory networks from genomic and functional genomic data. With the example of E. coli I will show in more detail that an integrated analysis of metabolic, regulatory and protein-protein interaction networks is particularly important for understanding both cellular metabolism and its regulation at a systems level. Some of our network-based studies of pathogens (e.g. Pseudomonas aeruginosa), industrially relevant organisms (e.g. yeast and Bacilli) and human cells will also be briefly highlighted, especially with respect to the analysis of transcriptomic and dynamic data.

Communication via Molecular Surfaces

Dozent	Prof. Tim Clark, Computer-Chemie Centrum, Erlangen
Ort	ZBH, Raum 16
Zeit	15. Jun 2006 16:00

Fast EMD-based Similarity Search in Large Multimedia Databases

Dozent	Prof. Dr. Thomas Seidl, RWTH Aachen UniversityData Management and Data Exploration Group, Aachen
Ort	ZBH, Raum 16
Zeit	01. Jun 2006 16:00

Content-based similarity search in large multimedia databases requires both, effective similarity models as well as efficient retrieval algorithms. In this talk, we focus on the Earth Movers' Distance (EMD) which has been recently introduced in the area of computer vision where it has been reported to well reflect human perceptual similarity for numerous applications including image, music, and shape retrieval. As its computation is complex, the direct application of the EMD to large, high-dimensional databases is not feasible. To overcome this limitation and to allow users to benefit from the high quality of the model even in larger settings, we propose novel approximations for the EMD to be used in index-supported multi-step query processing architectures. A formal analysis of our new filters shows that they are correct and complete, thus producing no false drops and no wrong answers with respect to the original EMD. Experiments on large image databases with high-dimensional feature vectors demonstrate the efficiency of our methods.

Homology Modelling and its Use in Target Discovery and Lead Optimisation

Dozent	Dr. Ursula Egner, Schering AGRCE Enabling Technologies, Berlin
Ort	ZBH, Raum 16
Zeit	18. May 2006 16:00

3-dimensional imaginary transition states for the study of biochemical reactions

Dozent	Dr. Joannis Apostolakis, LMU MünchenInstitut für Informatik, München
Ort	ZBH, Raum 16
Zeit	04. May 2006 16:00

I will present the concept of 3d-ITS, which is a simple model of the transition state of a given reaction. 3d-ITS are a 3 dimensional extension to the imaginary transition states introduced by S. Fujita 20 years ago to formally describe reactions. 3d-ITS can be automatically calculated for biochemical reactions based on a simple graph theoretic approach. In this talk I will describe the calculation of 3d-ITS, present extensive validation of the resutls and demonstrate their utility in a number of applications, ranging from metabolomics to drug design. I will argue that 3d-ITS is a very useful theoretical tool for making the connection between metabolomics and structural bioinformatics. In the context of the 3d-ITS discussion I will also describe Chill2 a new cheminformatics and modelling platform developed at our group.

Berechnung größter gemeinsamer Teilgraphen zur Visualisierung von Molekülmengen mit einheitlichem Grundgerüst

Dozent	Florian Krull
Ort	ZBH, Raum 16
Zeit	27. Apr 2006 16:45

Generierung eindeutiger linearer Molekülpräsentationen unter Berücksichtigung von Stereoisomerie

Dozent	Karin Dietrich
Ort	ZBH, Raum 16
Zeit	27. Apr 2006 16:00

SCOPPI.org: The structural classification of protein-protein interactions

Dozent	Prof. Dr. Michael Schroeder, Biotec / Department of Computing, TU Dresden, Dresden
Ort	ZBH, Raum 16
Zeit	20. Apr 2006 16:00

Abstract: SCOPPI, the Structural Classification of Protein-Protein Interfaces (http://www.scoppi.org), is a comprehensive database that classifies and annotates domain interactions derived from all known protein structures. SCOPPI uses SCOP domain definitions and a distance criterion to determine the inter-domain interface formed by two domain faces. SCOPPI presents a geometrical classification of all domain interfaces using a novel method based on a multiple sequence and multiple structural alignment of all members in a SCOP family. Various interface characteristics such as number, type and position of interacting amino acids, conservation of amino acids, permanent or transient nature of the interface and interface size are also provided. We use this wealth of information for various applications: 1. Viral mimicry of native interfaces: What do non-homologous domains, which use equivalent binding sites in a shared partner have in common? With SCOPPI we reveal examples of such convergently evolved motifs in viral proteins, which mimic native interfaces such as NEF/PKC, p35/IAP, and Cpasid/Cyclophilin. 2. From SCOPPI we derive family-based residue interaction propensities, which we use to improve docking scoring functions. For enzyme/inhibitor examples we can improve ranking 30-fold and for antibody/antigen up to 13-fold. 3. We use SCOPPI to classify transient and permanent interactions. We find that size related properties can consistenly classify interfaces with over 97% accurancy. 4. We apply SCOPPI to predict protein interactions. For a set of overexpressed genes in pancreas tumor SCOPPI predicts some kinase-inhibitor interactions. SCOPPI also provides access to all associated relevant literature via GoPubMed. GoPubMed.org is a novel search engine for PubMed articles. It uses the GeneOntology to classify all articles according to the processes, functions, and cellular components they mention. The user can then browse the results using these classifications. GoPubMed also shows for all topics key authors, journals, and number of publications over time.

Structure and dynamics of poly(propylene oxide) (PPO) from molecular dynamics simulations and neutron scattering

Dozent	Prof. Peter Ahlström, School of Engineering, University College of Boras, Boras, Sweden
Ort	ZBH, Raum 16
Zeit	27. Mar 2006 16:00

Poly(propylene oxide) [PPO] is an inherently amorphous polymer used in, e.g., Litium batteries. In this talk, results regarding structure and dynamics of PPO as obtained from molecular dynamics simulations will be presented as well as compared to and used for the interpretation of neutron scattering data. From the comparison between simulations and experimental data the degree of deuteration in the sample could be deduced. Local (side chain) and segmental dynamics have been simulated and are in excellent agreement with experimental results. The results indicate that the PPO dynamics neither follows the Rouse nor the Gaussian model.

Validation of Docking Software for Drug Design against RNA Drug Targets

Dozent	Dr. Carsten Detering, BioSolveIT GmbH, St. Augustin
Ort	ZBH, Raum 16
Zeit	09. Feb 2006 16:00

On trees, forests and machines, or: do new brooms clean better?

Dozent	Prof. Dr. Andreas Ziegler, Institut für Medizinische Biometrie und Statistik, Universität zu Lübeck, Lübeck
Ort	ZBH, Raum 16
Zeit	26. Jan 2006 16:00

Standard tools for classification like the logistic regression or artificial neural nets are widely used in applications. In the last decade, several new methods have been developed from the learning machine community. These include support vector machines as well as ensemble methods, including random forests. In the talk, I will sketch the classification problem of interest and show the standard findings from logistic regression as well as the fundamental ideas of support vector machines and random forests. The talk will close with a comparison of the different machines.

From sequence to function: in silico methods for optimizing biocatalysts

Dozent	PD Dr. Jürgen Pleiss, Institute of Technical Biochemistry, University of Stuttgart, Stuttgart
Ort	ZBH, Raum 16
Zeit	08. Dec 2005 16:00

Abstract: Nature has generated an impressing diversity of biocatalysts varying in sequence and biochemical properties. This biodiversity is being exploited by screening of cultivable and uncultivable organisms for new enzymes with improved properties. Protein engineering is a powerful tool to further improve activity, substrate specificity, and regio- or stereoselectivity of recombinant enzymes. Understanding the relationship between sequence, structure, and function of enzymes would greatly help screening and engineering; conversely, large-scale experimental data provide an invaluable knowledge base for developing molecular models of enzymes. To study the relationship of sequence, structure, and function of enzymes, we have combined two approaches: the systematic analysis of a comprehensive family-specific database which integrates annotated sequence and structure information, and molecular modeling of enzyme-substrate interactions by docking and molecular dynamics simulations. The Lipase Engineering Database has been established as a tool for systematic analysis of more than 1500 a/b hydrolases and to identify structurally and functionally relevant modules 1-3. By combining docking and molecular dynamics simulations, a general molecular model of enantioselectivity of a/b hydrolases was established. The model is general and predictive: a broad variety of substrates were ranked by selectivity and improved mutants in the substrate binding site were predicted 4. Recently, molecular dynamics simulations were successfully applied to reproduce long-range effects of mutations in a metallo-b-lactamase 5. 1. M. Fischer, J. Pleiss, The Lipase Engineering Database: a navigation and analysis tool for protein families. Nucleic Acids Res. (2003), 31, 319-321. 2. S. Barth, M. Fischer, R. D. Schmid, J. Pleiss, Sequence and structure of epoxide hydrolases: a systematic analysis. Proteins (2004), 55, 846-855. 3. E. Henke, J. Pleiss, U. T. Bornscheuer, Activity of lipases and esterases towards tertiary alcohols: insights into structure-function relationships. Angew. Chem. Int. Ed. (2002), 41, 3211-3213. 4. E. Henke, U. T. Bornscheuer, R. D. Schmid, J. Pleiss, A molecular mechanism of enantiorecognition of tertiary alcohols by carboxylesterases. ChemBioChem. (2003), 4, 485-493. 5. P. Oelschlaeger, R. D. Schmid, J. Pleiss, Modeling domino effects in enzymes: molecular basis of substrate specificity of the bacterial metallo-b-lactamases IMP-1 and IMP-6. Biochemistry (2003), 42, 8945-8956.

From sequence to function: in silico methods for optimizing biocatalysts

Dozent	PD Dr. Jürgen Pleiss, Institute of Technical Biochemistry, University of Stuttgart, Stuttgart
Ort	ZBH, Raum 16
Zeit	08. Dec 2005 16:00

Abstract: Nature has generated an impressing diversity of biocatalysts varying in sequence and biochemical properties. This biodiversity is being exploited by screening of cultivable and uncultivable organisms for new enzymes with improved properties. Protein engineering is a powerful tool to further improve activity, substrate specificity, and regio- or stereoselectivity of recombinant enzymes. Understanding the relationship between sequence, structure, and function of enzymes would greatly help screening and engineering; conversely, large-scale experimental data provide an invaluable knowledge base for developing molecular models of enzymes. To study the relationship of sequence, structure, and function of enzymes, we have combined two approaches: the systematic analysis of a comprehensive family-specific database which integrates annotated sequence and structure information, and molecular modeling of enzyme-substrate interactions by docking and molecular dynamics simulations. The Lipase Engineering Database has been established as a tool for systematic analysis of more than 1500 a/b hydrolases and to identify structurally and functionally relevant modules 1-3. By combining docking and molecular dynamics simulations, a general molecular model of enantioselectivity of a/b hydrolases was established. The model is general and predictive: a broad variety of substrates were ranked by selectivity and improved mutants in the substrate binding site were predicted 4. Recently, molecular dynamics simulations were successfully applied to reproduce long-range effects of mutations in a metallo-b-lactamase 5. 1. M. Fischer, J. Pleiss, The Lipase Engineering Database: a navigation and analysis tool for protein families. Nucleic Acids Res. (2003), 31, 319-321. 2. S. Barth, M. Fischer, R. D. Schmid, J. Pleiss, Sequence and structure of epoxide hydrolases: a systematic analysis. Proteins (2004), 55, 846-855. 3. E. Henke, J. Pleiss, U. T. Bornscheuer, Activity of lipases and esterases towards tertiary alcohols: insights into structure-function relationships. Angew. Chem. Int. Ed. (2002), 41, 3211-3213. 4. E. Henke, U. T. Bornscheuer, R. D. Schmid, J. Pleiss, A molecular mechanism of enantiorecognition of tertiary alcohols by carboxylesterases. ChemBioChem. (2003), 4, 485-493. 5. P. Oelschlaeger, R. D. Schmid, J. Pleiss, Modeling domino effects in enzymes: molecular basis of substrate specificity of the bacterial metallo-b-lactamases IMP-1 and IMP-6. Biochemistry (2003), 42, 8945-8956.

Computational tools for the optimization of antiretroviral drug therapies

Dozent	Prof. Dr. Joachim Selbig, Institut für Biochemie und Biologie, Universität Potsdam, c/o MPI für Molekulare Pflanzenphysiologie, Potsdam-Golm
Ort	ZBH, Raum 16
Zeit	01. Dec 2005 16:00

Despite the approval of 18 antiretroviral drugs and the use of combination therapy, successful treatment of HIV-infections is hampered by the emergence of drug-resistant genetic variants in response to therapy. Finding a new potent drug combination after treatment failure is considered challenging, because most accumulated mutations confer resistance to multiple drugs. We present computational methods for the integration and analysis of viral sequence data, phenotypic resistance data, and clinical data from infected patients. We design and implement a relational database schema that serves as the logical and physical basis for this task. Machine learning techniques are applied to predict phenotypic drug resistance. To support therapy optimization, we develop a scoring function for drug combinations. For more information, cf. www.geno2pheno.org.

Numerical Simulations of Protein Folding in an Off-Lattice Model

Dozent	Stefan Goltz, Technische Universität Hamburg-Harburg, Hamburg
Ort	ZBH, Raum 16
Zeit	17. Nov 2005 16:00

Studies on Molecular Docking Accuracy

Dozent	Dr. Jon Erickson, Eli Lilly, Lilly Corporate Center, Indianapolis, USA
Ort	ZBH, Raum 16
Zeit	10. Nov 2005 16:00

Automated Structure Diagram Generation of Molecular Complexes

Dozent	Katrin Stierand, Zentrum für Bioinformatik Hamburg
Ort	ZBH, Raum 16
Zeit	08. Nov 2005 16:00

Fragment Based Screening on Chemical Microarrays

Dozent	Dr. Günther Metz, Santhera Pharmaceuticals, Heidelberg
Ort	ZBH, Raum 16
Zeit	03. Nov 2005 16:00

Score functions for protein sequence optimisation: Statistical constructiontesting

Dozent	Benjamin Otto, Zentrum für Bioinformatik Hamburg
Ort	ZBH, Raum 16
Zeit	27. Oct 2005 16:30

Flexible Pattern Matching on String Indices Based On The Burrows and Wheeler Transform

Dozent	Andrea Papst, Zentrum für Bioinformatik Hamburg
Ort	ZBH, Raum 16
Zeit	27. Oct 2005 16:00

In modern biology, pattern searching in texts, e.g. genomes, is an important task. To accelerate the search, several types of indices have been developed. Fast and commonly used indices are the suffix array and the suffix tree. Their drawback is the memory consuption which is many times over the original text size. Recently the FM-Index was proposed. It needs less space than the text itself while still supporting fast search for simple patterns. In this talk, FM-index based algorithms are presented to solve more sophisticated pattern matching problems like searching for regular expressions or comparing two texts. The algorithms are based on the efficient construction of the reverse prefix tree using the FM index. Theoretical analysis and experimental comparison with the suffix array show a comparable speed while memory consumption is significantly reduced.

Simulation of conformational dynamics and energy transfer in biomolecules

Dozent	Prof. Dr. Gerhard Stock, Johann-Wolfgang-Goethe-Universität Frankfurt, Theoretische Chemie komplexer Systeme, Frankfurt am Main
Ort	ZBH, Raum 16
Zeit	09. Sep 2005 16:00

Recent multidimensional vibrational experiments on peptides and proteins have provided a wealth of new information on the conformational dynamics of these systems. To study the correlations between spectra and peptide dynamics and to obtain a microscopic understanding of the phenomena, molecular dynamics (MD) simulations of various small peptides in aqueous solution have been performed. From these simulations, time-dependent correlation functions reflecting vibrational dephasing, energy transfer and energy relaxation are computed and compared to experiment. Furthermore, nonequilibrium MD simulations of laser-induced peptide dynamics are presented, which allow us to study the folding and unfolding of small peptides in atomistic detail. Analyzing the free energy landscape, various conformational states as well as pathways of helix nucleation are identified and discussed.

Complexity, Universality, Self-similarity and Growth of Genomes

Dozent	Prof. H.-C. Paul Lee, Dept. of Physics and Dept. of Life Sciences, National Central University, Chungli, Taiwan
Ort	ZBH, Raum 16
Zeit	07. Sep 2005 16:00

We have recently completed a textual analysis of all extant complete genomes. This has shown that, despite the high degree of genomic diversity, there is an undeniable universality in the statistical property of word frequencies in complete genomes. Furthermore, each individual genome is structurally complex, yet highly self-similar. These properties strongly constrained the way genomes grew and evolved. We show that a simple, biologically viable growth model with three universal parameters can account for these complex, universal, and self-similar genomic properties. The main propositions are that maximally stochastic segmental duplication was the main mechanism driving genome growth, and that the duplication process most likely began when the genomes were smaller than 300 base pairs long. The model gives a natural explanation for many biological phenomena of genomes not related to word frequency. We discuss implications of our model in the context of genome evolution, the astonishing rapid accumulation of information by the genome, and the universal ancestor.

Troponin: a calcium-regulated switch protein that controls skeletal muscle contraction

Dozent	Dr. Paul Curmi, School of Physics, University of New South Wales, Sydney, Australia
Ort	ZBH, Raum 16
Zeit	01. Sep 2005 16:00

Troponin is a Ca2+-sensitive switch that regulates the contraction of vertebrate striated muscle by participating in a series of conformational events within the actin-based thin filament. Troponin is a heterotrimeric complex consisting of a Ca2+-binding subunit (TnC), an inhibitory subunit (TnI), and a tropomyosin-binding subunit (TnT). We have modelled the structure of the troponin complex in both the “on-state” (+Ca2+) and the “off-state” (-Ca2+) by combining a full set of small-angle neutron scattering (SANS) and small-angle x-ray scattering (SAXS) data with the crystal structure of cardiac troponin. A rigid-body Monte Carlo optimization procedure was used to yield the two models. The optimized models show significant differences when compared to the cardiac troponin crystal structure in the +Ca2+ state and provide a structural model for the switch between +Ca2+ and -Ca2+ states. More recently, we have used our models to interpret single particle electron microscopy images of intact thin filaments.

Intra-Cellular Networks

Dozent	Dr. Somdatta Sinha, Centre for CellularMolecular Biology, Hyderabad, India
Ort	ZBH, Raum 16
Zeit	28. Jul 2005 16:00

Dynamics of Protein Binding, Reaction and Structural Change

Dozent	Prof. Jeremy C. Smith, Computational Molecular Biophysics, University of Heidelberg, Heidelberg
Ort	ZBH, Raum 16
Zeit	07. Jul 2005 16:00

Computer simulation results on dynamical effects influencing ligand binding, reactions and the mechanics of large-scale conformational change in proteins will be discussed. Among the questions to be examined are the thermodynamic consequences of vibrational changes on ligand binding, the protein glass transition and function, proton transfer reactions in the light-driven proton pump protein, bacteriorhodopsin, and the mechanics of the power stroke of muscle contraction.

Effizientes Screening durch Wiederverwertung redundanter Basisfragmente

Dozent	Andreas Scherzler
Ort	ZBH, Raum 16
Zeit	23. Jun 2005 16:45

Optimization Strategies for the Docking of Flexible Molecules

Dozent	Axel Griewel
Ort	ZBH, Raum 16
Zeit	23. Jun 2005 16:00

DNA Computing

Dozent	Prof. Dr. Karl-Heinz Zimmermann, Computing Systems, Technical University Hamburg-Harburg, Hamburg
Ort	ZBH, Raum 16
Zeit	16. Jun 2005 16:00

Kernel Based QSAR Modelling - A New Approach to ADME in Silico Prediction

Dozent	Holger Fröhlich, Center for Bioinformatics, University of Tübingen
Ort	ZBH, Raum 16
Zeit	02. Jun 2005 16:00

Kernel methods, like the well known Support Vector Machine (SVM), have gained a growing interest during the last years for designing QSAR models having a high predictive strength. One of the key concepts of SVMS is the usage of a so-called kernel function, which can be thought of as a special similarity measure. In this talk we consider kernels for molecular structures, which are based on a graph representation of chemical compounds. Examples thereof are marginalized graph kernels (KashimaTsuda, 2004) and optimal assignment kernels (Froehlich et al., 2005). Thereby especially optimal assignment kernels show a promising prediction performance on several ADME datasets, which is comparable to a classical descriptor based model using only a few descriptors, that are a-priori known to be relevant. At the same time results are significantly better than with and without performing an automatic descriptor subset selection. We further explore the effect of combining the optimal assignment kernel with problem dependent descriptor information. Furthermore, we demonstrate the usefulness of an extension of optimal assignment kernels to a reduced graph representation of molecules, in which certain structural features, like e.g. rings, donors or acceptors, are represented as a single node in the molecular graph and hence allow to look at different scales of resolution at the molecular structure.

Computer simulation of protein-protein association including side chain and global flexibility

Dozent	Prof. Dr. Martin Zacharias, International University Bremen, School of Engineering and Science, Bremen
Ort	ZBH, Raum 16
Zeit	28. Apr 2005 16:00

Most current docking approaches to predict the binding geometry of protein-protein complexes use rigid protein partner structures. However, protein complex formation can involve both local conformational changes of side chains and loops at the protein-protein interface and global conformational relaxation of the protein partners. We have developed a docking approach that is based on energy minimization of translational and rotational degrees of freedom of protein partners and on a reduced protein representation allowing efficient search for docking minima. A multicopy approach is used to select the most favourable side-chain or loop conformation at the protein-protein interface during the docking process. To approximately account for possible global conformational adaptation, a method has been developed that allows to relax the protein structure in pre-calculated flexible degrees of freedom (soft modes) during docking. Such flexible modes can, for example, be obtained from molecular dynamics simulations or on the level of a reduced protein representation by employing an energy function that depends on the local protein density. Application of the approaches to test systems and targets of the CAPRI (Critical Assessment of PRedicted Interactions) docking challenge will be presented.

Convergent evolution of receptors for protein import into mitochondria

Dozent	Prof. Paul Gooley, Department of Biochemistry and Molecular Biology and Bio21, University of Melbourne, Australia
Ort	ZBH, Raum 16
Zeit	21. Apr 2005 16:00

The Translocase of the Outer Mitochondrial Membrane (TOM) receptor is an essential multisubunit machine for importing proteins into the mitochondria. We have determined the solution NMR structure of the cytosolic ligand-binding domain of the `master' receptor Tom20 from plants which shows two novel tetratricopeptide repeats. Remarkably, the plant protein shows no detectable sequence similarity with the corresponding animal sequence, but structural features can be seen when one sequence is read backwards. Based on the sequence and structure information, we can pose some interesting hypotheses regarding protein functions and the split of the evolutionary lineage.

Bayesian inference applied to macromolecular structure determination by NMR

Dozent	Dr. Michael Habeck, Unité de Bio-Informatique Structurale, Institut Pasteur, Paris, France
Ort	ZBH, Raum 16
Zeit	14. Apr 2005 16:00

Abstract: Calculation of macromolecular structures from nuclear magnetic resonance (NMR) data involves subjective decision-making: NMR structures are not fully determined by experimental data but strongly depend on personal choices in data treatment and parameter settings. This makes an objective judgement of the uniqueness and quality of NMR structures impossible. We use Bayesian inference to resolve this fundamental issue. Our inferences are derived from a probability density that represents the unknown structure and its precision. Other unknowns, such as the precision of the measurements, that are required to infer the coordinates but hitherto had to be chosen empirically, are also determined by this probability density. We find that our approach is practically feasible by Markov Chain Monte Carlo (MCMC) techniques. For the first time, we are thus able to obtain NMR structures in an objective way. Our method leads to a gain in structural quality; it is completely general and not limited to NMR data.

Unified sequence and structure alignment

Dozent	Prof. Dr. Jens Kleinjung, Faculty of Sciences, Free University of Amsterdam, Amsterdam, The Netherlands
Ort	ZBH, Raum 16
Zeit	18. Feb 2005 15:00

Computer Vision: some approaches to model-based tracking of humans

Dozent	Paul Reuter, Karlsruhe
Ort	ZBH, Raum 16
Zeit	27. Jan 2005 16:00

Structural Biology of Mtb - a new route to anti-TB drugs?

Dozent	Dr. Manfred S. Weiss, EMBL Hamburg Outstation
Ort	ZBH, Raum 16
Zeit	13. Jan 2005 16:00

Since the WHO has declared tuberculosis a global health emergency in 1993, the need for new drugs against the disease has become more and more apparent. In a large project, which involves many research groups around the whole world, a structure-based approach is used to identify and validate new lead compounds which may have the potential of being developed into new drugs. Particularly, the enzymes of the Leu- and Lys-biosynthetic pathways appear to be promising targets. Some of the recent structural results will be presented as well as first hints towards the identification of inhibitors.

Rational crystallogenesis using mesoscopic protein models

Dozent	Prof. Thomas Huber, The University of Queensland, Brisbane, Australia
Ort	ZBH, Raum 16
Zeit	16. Dec 2004 16:00

Biological macromolecules need to be brought into a well ordered crystalline form in preparation for structural elucidation through X-ray diffraction analysis. Reliable prediction of crystallisation conditions has proven elusive - an unsurprising circumstance given the apparently fickle nature of crystal growth, widespread ignorance of its causes and lack of any systematic recording of crystallisation data. Trial-and-error methodologies based on “crystal screens” are commonly featured by crystallogenesis designs, often unencumbered by planning or logic. Anecdotal evidence perpetuates the notion that crystal formation is best achieved through persistence and luck, a notion reflected in the three current pillars of crystallization: exotic chemical additives, hope, and, as is increasingly common, massive scale. The focus of this talk will be on approaches towards rational crystallogenesis using mesoscopic protein models. I will introduce a simple physical model that captures many essential features important to predict conditions under which a protein is likely to crystallise. The model’s generality and performance is demonstrated on a large, statistically significant number of proteins and a total of more than 80000 consistently recorded crystallisation experiments. Further, I will show how the model was used to rational design conditions to crystallise Lysozyme over a wide range of pH. Finally, limitations are outlined by following the time honoured tradition of trying to understand things by breaking them.

Mapping Property Distributions of Molecular Surfaces (MaP): Theory, validation and extensions of a translationally and rotationally invariant 3D-QSAR-technique

Dozent	Dr. Nikolaus Stiefl, Lilly Research Laboratories (LRL), Hamburg
Ort	ZBH, Raum 16
Zeit	09. Dec 2004 16:00

During the past decades, a vast number of molecular descriptors have been developed to numerically capture the chemical properties of small molecules. Unfortunately, when modelling quantitative structure relationships (QSAR’s) of a set of compounds most of these descriptors are not easy to interpret, which in turn makes the proposal of optimised leads from computational and medicinal chemists a major problem. Even though, major breakthroughs in terms of interpretability have been achieved by grid?based 3D-QSAR techniques (e.g. Comparative Molecular Field Analysis, CoMFA) as well as other methods that allow a chemically comprehensible visualisation of important descriptors, the so-called ‘alignment’ problem is still a widely spread disadvantage of these methods. To overcome this drawback, the translationally and rotationally invariant (TRI, sometimes called ‘alignment-independent’) 3D-QSAR-technique MaP was developed. MaP is based on simple pharmacophore-type surface properties, which describe the characteristic interactions at the layer between ligand and receptor. In this talk, a brief introduction on the topic of TRI-descriptors and the chemical and statistical theory behind MaP will be given. Additionally, application examples, the structure-based validation as well as the latest 4D-extensions of MaP (xMaP) will be briefly introduced.

Minimum Cycle Bases in Graphs

Dozent	Dr. Franziska Berger, Vienna University of Technology, Institute of discrete mathematics and geometrie, Vienna, Austria
Ort	ZBH, Raum 16
Zeit	02. Dec 2004 16:00

Ring structures in molecules are important substructures for several applications in Computational Chemistry. However, typically not all rings are of interest. Hence different sets of 'meaningful' rings have been proposed over the years. The most well-known set is perhaps the 'Smallest Set of Smallest Rings' which corresponds to a minimum cycle basis of the model graph of the molecule. In the talk, we present mathematically exact and efficient algorithms for computing a minimum cycle basis of a molecular graph and introduce other interesting cycle sets.

Vortragsreihe Lebenswissenschaften - Alles Leben ist ChemieGenomprojekte: Warum Sequenzen nicht lügenPD Dr. Ute Willhöft, Zentrum für Bioinformatik Mittwoch, den 17.11.2004, 17:15-18:00Hörsaal B im Fachbereich Chemie, Martin-Luther-King-Platz 6

Dozent	Genomprojekte: Warum Sequenzen nicht lügen
Ort	Hörsaal B
Zeit	17. Nov 2004 17:15

Genomprojekte befassen sich mit der Entschlüsselung der DNA Sequenz ganzer Genome. Das Human Genomprojekt ist das bekannteste Projekt, hier wurde bereits 2001 eine Arbeitsversion publiziert. Weitere Arten sind derzeit Gegenstand von Genomprojekten, so zum Beispiel wichtig Nutzpflanzen wie Reis und Mais oder auch Genome von Säugetierarten, die Vergleiche mit dem Humangenom ermöglichen. Zunächst werden die Schlüssel-Techniken zur Sequenzierung und Assemblierung von Genomen vorgestellt, die durch das Zusammenspiel von molekularbiologischen und bioinformatischen Methoden qualitativ hochwertige Sequenzen generieren. Für die Sequenzierung wird die genomische DNA zunächst in kleine Fragmente zerlegt und mit Hochdurchsatzmethoden sequenziert (siehe Abbildung). Für die Assemblierung werden Softwareprogramme genutzt, die aus großen Datenmengen an Sequenzfragmenten die chromosomale Anordung rekonstruieren. Im letzten Teil wird exemplarisch vorgestellt, wie weitere Auswertungen der Sequenzdaten eine neue Sicht auf die Lebenswissenschaften ermöglichen. Informationen im Internet zu Genomprojekten:http://www.dhgp.de/ - Einführung in das Human-Genomprojekt (in deutsch) http://www.genomesonline.org - Liste mit allen Genomprojekten, die öffentlich zugänglich sindhttp://www.ncbi.nlm.nih.gov/Traces/trace.cgi? - Trace Archive Sequenzen und Assemblierung von Rohsequenzen http://www.ncbi.nlm.nih.gov/genome/guide/human/ - alle Daten zum Human-Genomprojekt - alle Daten zum Human-Genomprojekthttp://www.doegenomes.org- alle Daten zum Human-Genomprojekt http://www.ensembl.org/ - alle Daten zum Human-Genomprojekt

A Mass Spectrometric Approach to Protein Structure

Dozent	Dr. Stephan Wefing, CAESAR Center of Advanced European Studies and Research, Bonn
Ort	ZBH, Raum 16
Zeit	28. Oct 2004 16:00

Tests of molecular simulation methods in Gramicidin A channel

Dozent	Dr. Serdar KuyucakSchool of Physics, University of Sydney, Australia
Ort	ZBH, Raum 16
Zeit	22. Oct 2004 16:00

Conformation dynamics, metastability analysis and virtual screening

Dozent	Dr. Frank CordesKonrad-Zuse-Zentrum für Informationstechnik Berlin (ZIB), Berlin-Dahlem
Ort	ZBH, Raum 16
Zeit	21. Oct 2004 16:00

A correct statistical representation of molecular flexibility is important for several aspects in the drug design pipeline. Identification of metastable subregions within the conformational space can help to reduce the so called critical slowing down of many conformational sampling methods. Docking algorithms can use presampled information of drug like molecules instead of exhaustive search techniques. Scoring functions need terms, which evaluate or approximate the change in thermo-dynamics during the coarse of binding. Methods for conformational dynamics developed at ZIB aim at a physically detailed description of the thermodynamical distribution of molecules. The talk will introduce the concept of metastability and its use for the simulation of conformation dynamics, for docking of conformational ensembles and for the design of new scoring functions.

Evaluation and Prediction of Protein Structures using a novel technique based on Probability Density Functions

Dozent	Ch. Kevin Au-Yeung, ACGT ProGenomics AG, Halle
Ort	ZBH, Raum 16
Zeit	13. Aug 2004 12:00

The prediction of reliable protein structure models from sequence data is still an unresolved problem. Various prediction methods use a library of conformation to identify a relationship between a protein structure and a peptide chain. This thesis describes a novel method, which can be applied for the calculations of peptide chain conformations as well as for the evaluation of protein structures. Basically, a library of conformations collected from resolved protein structures is the key of this method. The database information originates from conformational data of tetrapeptides. Conformations were recorded by determining the dihedral angles of the second and third residue of a tetrapeptide. A set of dihedral angle extracted for a particular tetrapeptide is then analysed using the non-parametric method of kernel density estimation. The result is a probability density function, which is able to describe the preferential conformation of a particular tetrapeptide. Altogether the conformational knowledge of the library derives from the probability density functions for every unique tetrapeptide.

Protein Dossier and STING Report: The Data Base and User Interface of the Physial-Chemical ParametersDescribing the Protein Structure and Function

Dozent	Dr. Goran Neshich, Nucleo de BioInformatica Estrutural, Embrapa Informática Agropecuária, Campinas, Brasil
Ort	ZBH, Raum 16
Zeit	06. Aug 2004 12:00

I will try to elaborate on an answer to the following question:Is there a set of parameters which can UNIQUELY define an amino acid ensemble which coincides with the active site of a given protein?

Parallel Computing in Life Science: Prospects and Pitfalls

Dozent	Dr. Thomas Steinke, Konrad-Zuse-Zentrum für Informationstechnik Berlin (ZIB), Berlin Center for Genome Based Bioinformatics (BCB), Berlin Dahlem
Ort	ZBH Zentrum für Bioinformatik, Bundesstr. 43, Raum 16
Zeit	08. Jul 2004 16:00

The challenge of protein structure prediction, variations on an old theme

Dozent	Dr. Ernst-Walter Knapp, Free University of Berlin, Institute of Chemistry, Macromolecular Modeling and Simulation of Biomolecules, Berlin
Ort	ZBH Zentrum für Bioinformatik, Bundesstr. 43, Raum 16
Zeit	24. Jun 2004 16:00

Data preparation and feature selection for chemical data sets - Building 'general' models

Dozent	Jörg K. Wegner, Zentrum für Bioinformatik Tübingen (ZBIT), Universität Tübingen
Ort	ZBH, Bundesstr. 43, Raum 16
Zeit	03. Jun 2004 16:00

We present some basic principles of data preparation and the actual feature selection topic in 'quantitative structure activity relationship' (QSAR). Especially the high variance for unseen samples (overfitting) will be shortly introduced to grant 'general' models (hypotheses) when applying feature selection. We introduce also the relevance of features for similarity analysis and discuss critically the often used 'neighborhood principle' also known as 'structure activity relationship'. For calculating the descriptors the open source library JOELib was used, which contains all presented descriptor calculation methods. Examples for predicting the Human Intestinal Absorption (HIA) and octanol/water partition coefficients logP are presented.

Building intelligent P2P search engines with asynchronous network programming

Dozent	Thorsten Henninger
Ort	ZBH, Bundesstr. 43, Raum 16
Zeit	27. May 2004 16:00

Structure-based De Novo Design: Lessons from Developing and Applying the Program Skelgen

Dozent	Dr. Martin Stahl, F. Hoffmann - La Roche AG, Pharmaceutical Division, Molecular Design and Bioinformatics, Basel, Switzerland
Ort	ZBH, Bundesstr. 43, Raum 16
Zeit	26. Mar 2004 11:00

Biological Hydrogen Conversion by the Enzyme (NiFe) Hydrogenase

Dozent	Dr. Matthias Stein
Ort	ZBH Zentrum für Bioinformatik, Bundesstr. 43, Raum 16
Zeit	25. Mar 2004 14:00

NiFe] hydrogenases are a class of enzyme that catalyze the heterolytic splitting of molecular hydrogen. The active site consists of an unusual heterobimetallic cluster which is paramagnetic in some states of the enzyme and can thus be studied with Electron Paramagnetic Resonance (EPR) spectroscopy and related techniques. The calculation of EPR parameters from first principles allows to correlate the composition of the active site in its various redox state with experimental data. Detailed investigations of the paramagnetic states as well as subsequent studies on the reaction mechanism of the enzyme will be presented.

EASED: Extended Alternatively Spliced EST Database

Dozent	Heike Pospisil, Ph.D., Max Delbrück Center for Molecular Medicin Berlin-Buch, Department of Bioinformatics, Berlin
Ort	ZBH Zentrum für Bioinformatik, Bundesstr. 43, Raum 16
Zeit	08. Mar 2004 15:00

The EASED project is establishing a comprehensive database of alternatively spliced human mRNAs. Moreover, EASED includes useful biological information and provides the possibility to search for biologically relevant data and for candidate genes for the origin of diseases. I will present our analysis of human alternative splice forms concerning their splice motifs and their biological meaning in relation to tissue type, developmental stage notation, cancer and the distribution in certain GO categories.

Docking of Ligands into

Dozent	Dr. Andreas Kämper, Max-Planck-Institut für Informatik, Saarbrücken
Ort	ZBH, Raum 16
Zeit	26. Feb 2004 16:00

One of the main goals in the field of molecular recognition is the de novo design of new artificial (synthetic) receptors. As a neccessary prerequisite, a fast and reliable method for the theoretical prediction of the structure of the resulting ligand-receptor complexes has been developed. Our approach is based on FlexX [1,2], a standard tool in drug design for docking flexible ligands into rigid proteins. This program has been adapted to the more general problem of docking into arbitrary organic hosts. For the handling of flexible receptors, first a full conformational analysis of the receptor is performed. Then the ligand is built up by the FlexX' incremental construction algorithm [1] sequentially in all conformers of the receptor. Alternatively, it is also possible to dock the receptor around the ligand, with the ligand acting as a template (inverse docking). Finally, all generated candidate structures are ranked by an empirical scoring function [3]. The new methods have been validated on a number of well known ligand-receptor complexes, and will form the basis for the development of algorithms for receptor design. The presented methods together with a simple user interface have been implemented in the tool FlexR. FlexX has been proven to give reliable docking results for thousands of different protein-ligand complexes. The new tool FlexR on the other hand is able to handle the most important types of complexes between artificial receptors and their ligands. With a combination of both tools docking of a complete new class of drugs, the glycopeptide antibiotics(GPA), is possible. The feasibility of docking cell-wall fragments into both unmodified GPA and their mimetics is shown. FlexR has also influenced FlexX in many ways. Some of the new additions to FlexX are the handling of protein/nucleic acid complexes (like DNA polymerase) and an improved descriptions of metalloproteins (like Urease). In recent study, FlexX was applied for the virtual screening for novel Urease inhibitors. During this process all available screening compounds have been collected, prepared for docking, and more than 2.5 million compounds have been docked, using a parallel version of FlexX.

Protein-Miniaturisation: Mimicking, Understanding and Utilising an Inhibitory Reactive Site Loop

Dozent	Dr. Arnd Brauer, Department of Chemistry, Imperial College London, London, UK
Ort	ZBH Zentrum für Bioinformatik, Bundesstr. 43, R16
Zeit	19. Feb 2004 16:00

Phase4: Automatic evaluation of database search methods

Dozent	Dr. Marc Rehmsmeier, International NRW Graduate School in Bioinformatics and Genome Research, CeBiTec, University Bielefeld
Ort	ZBH Zentrum für Bioinformatik, Bundesstr. 43, Raum 16
Zeit	29. Jan 2004 16:00

Aktuelle Themen der Bioinformatik - Phylogeny: An Overview of Methods and Tools

Dozent	Dr. Ute Willhöft (ZBH)
Ort	ZBH, Raum 16
Zeit	28. Jan 2004 16:15

Modeling and Comparison of Binding Sites for Rational Drug Design

Dozent	Prof. Dr. Gerhard Klebe, Institut für Pharmazeutische Chemie, Philipps-Universität Marburg
Ort	ZBH Zentrum für Bioinformatik, Bundesstr. 43, Raum 16
Zeit	22. Jan 2004 16:00

Aktuelle Themen der Bioinformatik - Now you see it, Now you don´t: Molecular Similarity

Dozent	Prof. Dr. Matthias Rarey (ZBH)
Ort	ZBH, Raum 16
Zeit	21. Jan 2004 16:15

From micro to meso to macro: Macromolecular simulation on various length scales

Dozent	Prof. Dr. Florian Müller-Plathe, International University Bremen, Bremen
Ort	ZBH Zentrum für Bioinformatik, Bundesstr. 43
Zeit	15. Jan 2004 16:00

Aktuelle Themen der Bioinformatik - Protein sequence design? Fact or Fraud?

Dozent	Prof. Dr. Andrew Torda (ZBH)
Ort	ZBH, Raum 16
Zeit	14. Jan 2004 16:15

Aktuelle Themen der Bioinformatik - Molecular Docking: Behind the Scenes

Dozent	Prof. Dr. Matthias Rarey (ZBH)
Ort	ZBH, Raum 16
Zeit	07. Jan 2004 16:15

Aktuelle Themen der Bioinformatik - Protein Structure Prediction - a Quarter Century of Defeat and Desaster

Dozent	Prof. Dr. Andrew Torda (ZBH)
Ort	ZBH, Raum 16
Zeit	17. Dec 2003 16:15

Changes in the energetics and flexibility of proteins upon macromolecular association

Dozent	Dr. Holger Gohlke, J.W. Goethe-Universität, Fachbereich Biologie und Informatik, Frankfurt/Main
Ort	ZBH Zentrum für Bioinformatik, Bundesstr. 43, Raum 16
Zeit	11. Dec 2003 16:00

Aktuelle Themen der Bioinformatik - Force Fields without Real Physics

Dozent	Prof. Dr. Andrew Torda (ZBH)
Ort	ZBH, Raum 16
Zeit	10. Dec 2003 16:15

Virtual Multi Component Reactions (MCR) for structure based ligand design

Dozent	Christof Gerlach, Institut für Pharmazeutische Chemie, Philipps-Universität Marburg, Marburg
Ort	ZBH Zentrum für Bioinformatik, Bundesstr. 43, Raum 16
Zeit	04. Dec 2003 16:00

Aktuelle Themen der Bioinformatik - Sorting Distances: From Pancake Flipping to Genome Rearrangement

Dozent	Prof. Dr. Matthias Rarey (ZBH)
Ort	ZBH, Raum 16
Zeit	03. Dec 2003 16:15

Structure, function and inhibition of ribosomal protein biosynthesis

Dozent	Dr. Frank Schluenzen, MPG - ASMB c/o DESY, Hamburg
Ort	ZBH Zentrum für Bioinformatik, Bundesstr. 43
Zeit	27. Nov 2003 16:00

Aktuelle Themen der Bioinformatik - Repeat Analysis in Complete Genomes

Dozent	Prof. Dr. Stefan Kurtz (ZBH)
Ort	ZBH, Raum 16
Zeit	26. Nov 2003 16:15

Aktuelle Themen der Bioinformatik - Multiple Sequence Alignments: Models and Algorithms

Dozent	Prof. Dr. Stefan Kurtz (ZBH)
Ort	ZBH, Raum 16
Zeit	19. Nov 2003 16:15

Aktuelle Themen der Bioinformatik - Sequence Analyses using BLAST: a Primer

Dozent	Dr. Ute Willhöft (ZBH)
Ort	ZBH, Raum 16
Zeit	12. Nov 2003 16:15

Neural Networks as Post-Docking Filters

Dozent	Dr. Jens Sadowski, AstraZeneca RD Mölndal, Mölndal, Sweden
Ort	ZBH Zentrum für Bioinformatik, Bundesstr. 43, Raum 16
Zeit	06. Nov 2003 16:00

Aktuelle Themen der Bioinformatik - Genome Projects: Why Sequences don´t lie

Dozent	Dr. Ute Willhöft (ZBH)
Ort	ZBH, Raum 16
Zeit	05. Nov 2003 16:15

Aktuelle Themen der Bioinformatik - Systematic Dynamic Programming in Bioinformatics

Dozent	Prof. Dr. Stefan Kurtz (ZBH)
Ort	ZBH, Raum 16
Zeit	29. Oct 2003 16:15

Phylogenetic Networks

Dozent	Prof. Dr. Daniel Huson, Center for Bioinformatics Tübingen, Tübingen University
Ort	ZBH Zentrum für Bioinformatik, Bundesstr. 43
Zeit	23. Oct 2003 16:00

Intergenomic Approach towards Consensi in Mulifactorial Inheritance Loci

Dozent	Dr. Steffen MöllerUniversität Rostock, Institut für Immunologie, Proteom Center Rostock, Rostock
Ort	ZBH, Raum 16
Zeit	29. Sep 2003 11:00

Hydrophobic contact, loop entropy and secondary structure

Dozent	Dr. Bosco HoCentre de Biophysique Moléculaire Numérique, Gembloux, Belgium
Ort	ZBH, Raum 16
Zeit	26. Sep 2003 11:00

Metabolic Analysis using Petri nets - demonstrated for the Sucrose-to-Starch Pathway

Dozent	Prof. Dr. Ina KochTechnische Fachhochschule Berlin, Berlin, Germany
Ort	ZBH, Raum 16
Zeit	25. Sep 2003 16:00

1. Teil

Dozent	Dr. Hanno WildBayer AG, Wuppertal
Ort	ZBH, Raum 16
Zeit	04. Sep 2003 11:30

Extracting Knowledge from High-Throughput Screening Data: Towards the Generation of Biophore Models

Dozent	Marc Zimmermann
Ort	ZBH, Raum 16
Zeit	03. Jul 2003 16:00

CLIC1 - structural transitions in an unusual redox-regulated chloride ion channel

Dozent	Dr. Paul CurmiSchool of Physics, University of Wales,Sydney, Australia
Ort	ZBH, Raum 16
Zeit	01. Jul 2003 16:00

3D-QSAR CoMSIA Models for CDK1, CDK5, and GSK3 Inhibition by Paullones

Dozent	Dr. Thomas LemckeUniversity of Hamburg, Institute of Pharmacology, Hamburg, Germany
Ort	ZBH, Raum 16
Zeit	26. Jun 2003 16:00

Building a conformation space for proteins and furnishing it with a potential function

Dozent	Prof. G. M. CrippenCollege of Pharmacology, University of Michigan, Ann Arbor, USA
Ort	ZBH, Raum 16
Zeit	13. Jun 2003 11:00

Evaluation of different virtual screening protocols for cyclin-dependent kinase 2 (CDK2)

Dozent	Hans Briem und Andreas SteffenSchering AG, Berlin
Ort	ZBH, Raum 16
Zeit	10. Jun 2003 11:00

Molecular modeling with CAChe software

Dozent	Sebastian SosnikFQS Poland, Krakow, Poland
Ort	ZBH, Raum 16
Zeit	19. Mar 2003 16:00

Mini-Workshop:

Dozent	Dr. Dirk Repsilber, Dr. Rainer Spang
Ort	ZBH, Raum 16
Zeit	07. Feb 2003 14:15

Teil I: A not to short Introduction to Normalization and Clustering Dr. Dirk Repsilber (Universität Lübeck Institut für Medizinische Biometrie und Statistik) Teil II: Differential Gene Expression and Computational Diagnostics Dr. Rainer Spang (Max-Planck-Institut für Molekulare Genetik, Berlin)

Layout of Molecule Structural Formula under User-Defined Constraints

Dozent	Patrick FrickerFraunhofer-Institut für Algorithmen und Wissenschaftliches Rechnen (FhI-SCAI) Universität Bonn
Ort	ZBH, Raum 16
Zeit	30. Jan 2003 16:00

Dynamic Programming Algorithms for Haplotype Block Partitioning

Dozent	Prof. Michael S. WatermanUniversity of Southern California,Los Angeles, CA, USA
Ort	ZBH, Raum 16
Zeit	10. Dec 2002 16:00

Integration of data management and analysis for genome research

Dozent	Volker BrendelIowa State University,Department of ZoologyGenetics
Ort	ZBH, Raum 16
Zeit	04. Oct 2002 16:15