ISMB 2008 ISCB

16th Annual
International Conference
Intelligent Systems
for Molecular Biology


Metro Toronto Convention Centre (South Building)
Toronto, Canada


 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

















Accepted Posters
Category 'Q'- Structure and Function Prediction'
Poster Q01
Prediction of COLLAGEN MODEL PEPTIDES Structure with Genetic Algorithms and Solvation Potential
Luis Scott- UFABC
No additional authors
Short Abstract: The work reported in this paper present the use of Genetic Algorithms (GA) with distinct field forces and rotamer library dependent of backbone to predict the tertiary structure of peptides as collagen model . We discuss an improved version in which the backbone and side chain were relaxed.
Long Abstract: Click Here

Poster Q02
Functional representation of enzymes by specific peptides
David Horn- Tel Aviv University
Vered Kunik (Tel Aviv University, School of Computer Science); Yasmine Meroz (Tel Aviv University, School of Physics); Zach Solan (Tel Aviv University, School of Physics); Ben Sandbank (Tel Aviv University, School of Computer Science); Uri Weingart (Tel Aviv University, School of Physics); Eytan Ruppin (Tel Aviv University, School of Computer Science);
Short Abstract: None On File
Long Abstract: Click Here

Poster Q03
A Target-Structure-Based Hybridization Model for Prediction of MicroRNA:Target Interactions
Ye Ding- New York State Department of Health
Dang Long (Wadsworth Center, Genetics Disorders); Chi Yu Chan (Wadsworth Center, Genetic Disorders); Rosalind Lee (UMASS Medical School, Molecular Medicine); Peter Williams (Dartmouth Medical School, Genetics); Victor Ambros (UMASS Medical School, Molecular Medicine);
Short Abstract: None On File
Long Abstract: Click Here

Poster Q05
High Quality Visualization of Molecular Skin Surface
Matthieu Chavent- CNRS
Bruno Levy (INRIA, LORIA); Bernard Maigret (CNRS, LORIA);
Short Abstract: MetaMol is a new program that generates high-quality 3D representations in interactive time. In contrast with existing software that discretize the surface with triangles or grids, our program is based on a GPU-accelerated ray-casting algorithm that directly uses the piecewise-defined algebraic equation of the Molecular Skin Surface.
Long Abstract: Click Here

Poster Q06
Automated Protein Subfamily Identification and Classification
Kimmen Sjolander- University of California Berkeley
Duncan Brown (Merck & Co., Inc, RNA Therapeutics); Nandini Krishnamurthy (Genentech, Bioinformatics);
Short Abstract: None On File
Long Abstract: Click Here

Poster Q07
Accuracy of structure-based sequence alignment of automatic methods
Byungkook (BK) Lee- NCI/NIH
Changhoon Kim (NCI/NIH, LMB/CCR);
Short Abstract: None On File
Long Abstract: Click Here

Poster Q08
Cell cycle kinases predicted from conserved biophysical properties
Kazimierz Wrzeszczynski- Columbia University
Burkhard Rost (Columbia University, Center for Computational Biology and Bioinformatics);
Short Abstract: Machine learning techniques can be employed to classify functionally related proteins when homology-transfer as well as sequence and structure motifs fail. We identify functionally significant residues in cell cycle proteins and then incorporated their biophysical features into a SVM classifier to identify and differentiate cell cycle kinases from other proteins.
Long Abstract: Click Here

Poster Q09
Computer-based screening of functional conformers of proteins
Gabriel del Rio- Universidad Nacional Autónoma de México, Instituto de Fisiología Celular
Héctor Marlosti Montiel Molina (UNAM, Biochemistry); César Millán-Pacheco (UAEM, Biochemistry and Molecular Biology); Nina Pastor (UAEM, Biochemistry and Molecular Biology);
Short Abstract: None On File
Long Abstract: Click Here

Poster Q10
Prediction of zinc-binding sites in proteins from sequence
Nanjiang Shu- Stockholm University
No additional authors
Short Abstract: None On File
Long Abstract: Click Here

Poster Q11
Prediction of RNA-protein interactions using SVMs
Ruth Spriggs- University of Sussex
Yoichi Murakami (Institute of Protein Research, Osaka University, Research Centre for Structural and Functional Proteomics); Susan Jones (University of Sussex, Chemistry and Biochemistry); Haruki Nakamura (Institute of Protein Research, Osaka University, Research Centre for Structural and Functional Proteomics);
Short Abstract: Support Vector Machines are trained to distinguish between RNA-binding and non-RNA-binding residues in a protein sequence, as the first step in a method using patterns of predicted binding residues to estimate whether a protein binds RNA. No previously published techniques using machine learning predict both if and where RNA binds.
Long Abstract: Click Here

Poster Q12
Predicting protein functional sites with a position-specific MINER
Dukka KC- University of North Carolina at Charlotte
Dennis Livesay (University of North Carolina at Charlotte, Bioinformatics/CS);
Short Abstract: Position-specific MINER (psMINER), our newest protein functional site prediction algorithm, combines phylogenetic motif (PM) and column conservation into a hybrid prediction scheme. The approach results in statistically significant improvements over traditional conservation-based methods. Moreover, the hybrid approach allows PMs to be used in a position-specific way.
Long Abstract: Click Here

Poster Q13
Rule analyses for KMSKS motif in aminoacyl-tRNA synthetase by its substitution to purely random sequences
Shunsuke Kamijo- The Universityu of Tokyo
Akihiko Fujii (The University of Tokyo, Institute of Industrial Science); Kenichi Wakabayashi (The University of Tokyo, Institute of Industrial Science); Kenji Onodera (The University of Tokyo, Institute of Industrial Science); Takatsugu Kobayashi (RIKEN, Genomic Science Center); Kensaku Sakamoto (RIKEN, Genomic Science Center);
Short Abstract: KMSKS loop is a motif for ATP binding in aminoacyl-tRNA synthetase.Although amino-acid sequences are well conserved around 'KMSKS' over species, they should have constraints from the revolutionally initial sequence.We then synthesized ramdon sequences substituting KMSKS, and searched rules for the loop to keep activity with ATP.
Long Abstract: Click Here

Poster Q14
Ensembles of Secondary Structure Predictors for Sequence-Based Beta-Residue Pair Prediction
Kanaka Durga Kedarisetti- University of Alberta
Lukasz Kurgan (University of Alberta, Electrical&Computer Engineering); Scott Dick (University of Alberta, Electrical&Computer Engineering);
Short Abstract: We address sequence-based prediction of beta-residue pairs that form intra-chain beta-sheets. Our meta-predictor, which employed five modern secondary structure predictors and the BETAPRO method [Bioinformatics 2005, Suppl1:i75-84], obtained 19.4% sensitivity and 19.5% specificity for beta-residue pair prediction based on cross-validation on 872 non-redundant sequences published in 2007.
Long Abstract: Click Here

Poster Q15
Predicting configuration of homologous subunits in protein assemblies
Dmitry Korkin- University of Missouri
Wah Chiu (Baylor College of Medicine, National Center for Macromolecular Imaging, Verna and Marrs McLean Department of Biochemistry and Molecular Biology); Judith Frydman (Stanford University, Department of Biological Sciences and BioX Program ); Andrej Sali (University of California at San Francisco, Department of Bioengineering and Therapeutic Sciences, Department of Pharmaceutical Chemistry, and California Institute for Quantitative Biosciences (QB3));
Short Abstract: Hetero-oligomers of homologous subunits generally result from evolutionary divergence of the homo-oligomers driven by functional specialization of the individual subunits. Here, we introduce a method for determining positions of the homologous proteins in a protein assembly, based on comparative modeling of the assembly and its interfaces followed by model assessment.
Long Abstract: Click Here

Poster Q16
A novel statistical method to predict functional regions of a protein
Wataru Nemoto- Advanced Institute of
Hiroyuki Toh (Kyushu University, Division of Bioinformatics);
Short Abstract: We developed a novel method to predict functional regions of a protein by using sequence and structure. Advantage of this method is that we can collect sequences for multiple sequence alignment automatically and objectively. We will discuss the benefits and the pitfalls of our method.
Long Abstract: Click Here

Poster Q17
i-SITE: Energy-based Method for Predicting Ligand-binding Sites on Protein Structures
Mizuki MORITA- The University of Tokyo
Tohru Terada (The University of Tokyo, Agricultural Bioinformatics Research Unit); Shugo Nakamura (The University of Tokyo, Department of Biotechnology); Kentaro Shimizu (The University of Tokyo, Department of Biotechnology);
Short Abstract: We have developed a method for predicting the ligand-binding sites on protein structures. It is a simple energy-based method and delivers high performance with apo protein structures. We could also improve the accuracy of prediction with re-ranking techniques by amino acid conservation scores.
Long Abstract: Click Here

Poster Q18
Analysis and Prediction of Curvature in the Leucine-Rich Repeat Proteins
Katherine Hindle- University of Manchester
Jordi Bella (University of Manchester, Faculty of Life Sciences); Simon Lovell (University of Manchester, Faculty of Life Sciences);
Short Abstract: We present a novel approach to the prediction of global structural characteristics in an important super-family of proteins, the leucine-rich repeat (LRR) containing proteins. We have developed a protocol to accurately predict the curvature of LRR proteins from sequence alone, an important step in the 3D modelling of these proteins.
Long Abstract: Click Here

Poster Q19
Focused docking: a computational approach to improve small-molecule docking into protein structures
Dario Ghersi- Mount Sinai School of Medicine
No additional authors
Short Abstract: A computational protocol that combines protein binding sites detection and docking is
presented here and evaluated on a set of 77 cases. The comparison with blind docking
shows that our protocol achieves a higher rate of binding site detection, more accurate
results and requires significantly less computational time.
Long Abstract: Click Here

Poster Q20
XIOS: A Graph Theoretical Approach to Finding RNA Structural Patterns
Michael Gribskov- Purdue University
Aditi Gupta (Purdue University, Biological Sciences); Kejie Li (Purdue University, Biological Sciences); Reazur Rahman (Purdue University, Biological Sciences);
Short Abstract: XIOS graphs describe RNA structures and their ensemble of suboptimal structures and pseudoknots. The maximal approximately isomorphous subgraphs of a set of XIOS graphs correspond to RNA structural motifs. This approach does not depend on sequence alignment and can be applied even when sequences have no detectable similarity.
Long Abstract: Click Here

Poster Q21
Fast Dynamics Perturbation Analysis for Prediction of Protein Functional Sites
Judith Cohn- Los Alamos National Laboratory
Dengming Ming (Nanjing University, School of Life Sciences); Michael Wall (Los Alamos National Laboratory, Computer, Computational, and Statistical Sciences Division; Bioscience Division; Center for Non-Linear Studies);
Short Abstract: We present a fast version of an algorithm that uses protein dynamics to predict functional sites, and apply it to 50,000 SCOP domains. The results recover much of the known information about binding sites and catalytic sites in SCOP domains, and suggest experimentally testable hypotheses concerning protein function.

Long Abstract: Click Here

Poster Q22
Why do proteins fold faster in vivo than in vitro?
Graham Wood- Macquarie University
David Fisher (Macquarie University, Statistics); Zelda Zabinsky (University of Washington, Industrial Engineering); Jonathan Ellis (Macquarie University, Statistics);
Short Abstract: Folding of a protein within the cell to its native state is known to be faster than the folding of a denatured (fully-extended) protein to its native state. Simple two-dimensional HP models are used to mimic these two processes; the average number of folding moves are compared, using computer simulation.
Long Abstract: Click Here

Poster Q23
A novel method considering physical binding domain for extracting protein complexes
Yosuke Ozawa- Keio University
Rintaro Saito (Keio University, Faculty of Environment and Information Studies); Shigeo Fujimori (Keio University, Graduate School of Science and Technology); Hiroshi Yanagawa (Keio University, Graduate School of Science and Technology); Etsuko Miyamoto-Sato (Keio University, Graduate School of Science and Technology); Masaru Tomita (Keio University, Graduate School of Media and Governance);
Short Abstract: Although some methods extract protein complexes from protein interaction network based on graph theory, the extracted complexes includes false positives.
We introduce a method, which considers binding domain to extract complexes. Using publicly available sets of human PPIs and DDIs, we succeeded in improving the accuracy compared to existing methods.
Long Abstract: Click Here

Poster Q24
Simulation of isolated Voltage Sensor Domains' Function
Christine Schwaiger- Center for Biomembrane Research (CBR)
Erik Lindahl (Center for Biomembrane Research (CBR), Institute of Biochemistry and Biophysics);
Short Abstract: Through simulation of isolated Voltage-Sensing Domains (VSDs) and in particular, Voltage-Sensor-Only Proteins (VSOPs), we aim to understand ion channels’ gating motion, the phosphatase activation of Ci-VSP and mVSOP’s/Hv1’s proton current. Additionally we try to detect the sequence-structure-function diversity of VSOPs by sequence-based predictions, homology modeling and MD-tools such as GROMACS.
Long Abstract: Click Here

Poster Q25
Molecular modeling of structural interaction between PfHslU and PfHslV subunits of P. falciparum and identification of interface residues: Implications for targeting protein-protein interface.
SANGEETHA SUBRAMANIAM- International center for genetic engineering and biotechnology
Dinesh Gupta (International center for genetic engineering and biotechnol, STRUCTURAL AND COMPUTATIONAL BIOLOGY); Asif Mohammed (International center for genetic engineering and biotechnology, Malaria reserach group);
Short Abstract: The present work discusses the results of predicted protein-protein interface of PfHslUV complex, the proteolytic machinery and a novel drug target of P. falciparum. PfHslUV model shares significant interface characteristic features with the prokaryotic homologues. Protein-protein interface can be an attractive target to disrupt the proteolytic machinery in P. falciparum.
Long Abstract: Click Here

Poster Q26
The organisation and analysis of coiled-coil structures
Efrosini Moutevelis- University of Bristol
Oliver D. Testa (University of Bristol, School of Chemistry); Derek N. Woolfson (University of Bristol, School of Chemistry);
Short Abstract: Coiled coils are protein-structure domains comprising 2 or more α-helices packed together via interlacing of side chains known as knobs-into-holes packing. Tools for structural and sequence analysis of coiled coils and a structural web server, CC+, are described. Additionally, we present a hierarchical ‚ÄúPeriodic Table‚ÄĚ of coiled-coil structures.
Long Abstract: Click Here

Poster Q27
Contact prediction for Transmembrane Proteins
Aron Hennerdal- Center for Biomembrane Research
Arne Elofsson (Center for Biomembrane Research, Department of Biochemistry and Biophysics, Stockholm University);
Short Abstract: We have developed a predictor based on support vector machines for residue-residue contacts in alpha-helical transmembrane proteins, utilizing sequence data as previously used for soluble proteins and additional data specific for membrane proteins. Our results are generally on par with methods for soluble proteins and in many cases performs better.
Long Abstract: Click Here

Poster Q28
Two-dimensional string matching techniques for protein contact maps
Robert Fraser- University of Waterloo
Alejandro Lopez-Ortiz (University of Waterloo, David R. Cheriton School of Computer Science);
Short Abstract: Contact maps are two dimensional abstract representations of protein structures. Searching for patterns in contact maps has generally used a costly naive sliding window approach. We study several approaches for two dimensional string matching to accelerate searching operations, and demonstrate experimentally the efficacy of these algorithms in our domain.
Long Abstract: Click Here

Poster Q29
PredPhospho2: Prediction of phosphorylation sites using SVMs
Jong Hun Kim- Samsung medical center
No additional authors
Short Abstract: Thanks to increased number of enrolled phosphorylation sites in public databases, we could make prediction models of eight kinase groups and twenty two kinase families while previous version of PredPhospho had models of four kinase groups and four kinase families. PredPhospho can
be applied to the functional study of proteins.
Long Abstract: Click Here

Poster Q30
Environment-Specific Substitution Tables for Membrane Proteins
Sebastian Kelm- University of Oxford
Jiye Shi (UCB-Group, ROC); Charlotte Deane (University of Oxford, Statistics);
Short Abstract: There currently exist no bioinformatics tools, which accurately predict the structure of membrane proteins. As a first step towards this goal, we are using environment-specific substitution tables for membrane proteins to study their molecular evolution.
Long Abstract: Click Here

Poster Q31
Activation Mechanism of GPCRs. In Silico Study of Cannabinoid Receptor Type 1
Angel Gonzalez- Pontifical Catholic University of Chile (PUC)
Leonardo Sepulveda (University of Illinois at Urbana-Champaign, Center for Biophysics and Computational Biology); Raul Araya-Secchi (Pontifical Catholic University of Chile (PUC) , Centre for Bioinformatics); Jose A Garate (Pontifical Catholic University of Chile (PUC) , Centre for Bioinformatics); C. David Pessoa-Mahana (Pontifical Catholic University of Chile (PUC) , Department of Pharmacy); Carlos F Lagos (Pontifical Catholic University of Chile (PUC), Centre for Bioinformatics); Tomas Perez-Acle (Pontifical Catholic University of Chile (PUC), Centre for Bioinformatics);
Short Abstract: The G-protein-coupled receptors constitute one of the largest superfamily of signaling proteins. Some of its members stand out because their functional states can be modulated by a broad spectrum of effector molecules. In this study we investigate the cannabinoid receptor ligand recognition plasticity, using a variety of computational tools.
Long Abstract: Click Here

Poster Q32
Retrieval of Profiles in a Case-Based Reasoning System for the Prediction of Protein Structure
Tony Kuo- Queen's University
Janice Glasgow (Queen's University, School of Computing); Hazem Ahmed (Queen's University, School of Computing); Sarah Rahmati (Queen's University, School of Computing);
Short Abstract: In our Case-Based Reasoning System for protein structure prediction, the retrieval of profiles considers both contact map and sequence similarity. This conserves any topological and homology information that may exist and is passed on into the later stages of adaptation and evaluation.
Long Abstract: Click Here

Poster Q33
Software for RNA 3-dimensional motif detection and backbone fit
Peter Clote- Boston College
Fabrizio Ferre (Harvard Medical School, Children's Hospital, Hematology/Oncology Department); William A. Lorenz (Boston College, Biology); Yann Ponty (Boston College, Biology);
Short Abstract: We describe two recent algorithms, available as web servers, developed by our lab for the classification of 3-dimensional RNA structural motifs and backbone atoms, currently one of the focal interests of the RNA Ontology Consortium. DIAL and LocalMove are available
http://bioinformatics.bc.edu/clotelab/DIAL
http://bioinformatics.bc.edu/clotelab/localmove
Long Abstract: Click Here

Poster Q34
Calculating Local Optima in the Turner Energy Model for RNA Secondary Structure
William Lorenz- Boston College
Peter Clote (Boston College, Biology);
Short Abstract: A novel technique for calculating the partition function of locally optimal secondary structures of a given RNA sequence is presented, using the Turner energy model. These states are shown to be relatively few in number, and are used to model kinetic folding of RNA.
Long Abstract: Click Here

Poster Q35
The knowledge-based force field CALF can fold a diverse set of short protein segments using only alpha carbon positions.
Patrick Buck- Rensselaer Polytechnic Institiute
Chris Bystroff (PI, Biology);
Short Abstract: CALF(C-ALpha based Force field) builds sequence specific statistical potentials based on the database frequencies from the hidden Markov model HMMSTR for alpha carbon virtual bond opening and dihedral angles, pairwise contacts and hydrogen bond donor/acceptor pairs, and simulates folding via Brownian dynamics.
Long Abstract: Click Here

Poster Q36
The implications of sequential protein folding in secondary structure
Rhodri Saunders- University Of Oxford
Charlotte Deane (University of Oxford, Statistics);
Short Abstract: Proteins are produced sequentially; some are known to fold sequentially and many may do so. Using simplified models we show that sequential folding is characterised by a restriction in the N-terminal structure space. We investigate how the directionality of protein production influences secondary structures in solved proteins.
Long Abstract: Click Here

Poster Q37
Pure Ab Initio Protein Structure Prediction Using Genetic Algorithms and Diversity-Keeping Strategies
Vinicius do O- Universidade de S√£o Paulo
Renato Tinós (Universidade de São Paulo, Faculdade de Filosofia, Ciências e Letras de Ribeirão Preto);
Short Abstract: In this work we describe a new approach to Ab Initio Protein Structure Prediction using Genetic Algorithms, adding Hypermutation and Random Immigrants procedures in order to increase population diversity and help alogrithm escape from local optima.
Long Abstract: Click Here

Poster Q38
PredPhospho2: Predictions of phosphorylation sites using SVMs
Jong Hun Kim- Sungkyunkwan University School of Medicine
No additional authors
Short Abstract: Thanks to increased number of enrolled phosphorylation sites in public databases, we could make prediction models of eight kinase groups and twenty two kinase families while previous version of PredPhospho had models of four kinase groups and four kinase families. PredPhospho can
be applied to the functional study of proteins.
Long Abstract: Click Here

Poster Q39
Applications of Delaunay Tessellation to the Analysis of Protein Structures
Todd Taylor- NIST
Iosif Vaisman (George Mason University, Bioinformatics and Comp Biology);
Short Abstract: Delaunay tessellation has proven extremely versatile in the analysis of protein structures. Here we relate several applications of Delaunay tessellation including secondary structure assignment, structural domain assignment, fold recognition, and the discrimination of thermophilic and mesophilic protein orthologs.

Long Abstract: Click Here

Poster Q40
Contact maps as an advisor to evaluate proposed protein structures
Mireille Gomes- Queen's University
Steve Sedfawi (Queen's University, School of Computing); Janice Glasgow (Queen's University, School of Computing);
Short Abstract: This work demonstrates how a simple attribute, specifically contact maps, may be used as part of an evaluation module for a case based reasoning system for predicting protein structures.
Long Abstract: Click Here

Poster Q41
Automated Ligand-Based Active Site Alignment: A Freely Available Extension to PyMol
Abraham Heifets- University of Toronto
Ryan Lilien (University of Toronto, Computer Science);
Short Abstract: The same ligand is likely to bind different proteins in similar, instructive ways. Our aim is to help automate comparisons of such active sites by developing freely available, easy to use visualization software. We demonstrate a proof-of-concept, PyMol-based, structure visualization tool, which utilizes ligand-based active site alignment.
Long Abstract: Click Here

Poster Q42
Predicting small ligand binding sites on proteins using low-resolution structures
Andrew Bordner- Mayo Clinic
No additional authors
Short Abstract: Non-covalently bound small ligands are important for protein function. We have developed a machine learning approach for predicting the location of such binding sites using only backbone geometry and so can utilize unrefined comparative models. Residue distributions, evolutionary conservation, and binding pocket shape contribute to the overall prediction.
Long Abstract: Click Here

Poster Q43
A framework for exploring conformational changes: Application to the transition between the open and closed states of K-channels
Angela Enosh- Tel Aviv University
Barak Raveh (The Hebrew University and Tel-Aviv university, Department of Molecular Genetics and Biotechnology ); Ora Furman-Schueler (The Hebrew University , Department of Molecular Genetics and Biotechnology ); Dan Halperin (Tel-Aviv university, School of Computer Science); Nir Ben-Tal (Tel-Aviv university, Department of Biochemistry );
Short Abstract: We present a general framework for the generation, alignment and comparison of motion pathways between two known protein conformations. The framework was implemented within the Rosetta software suite, and was used to study conformational changes in the potassium (K) channel. Our analysis suggested a safe-lock mechanism for channel-opening.
Long Abstract: Click Here

Poster Q44
Clustering Across Space and Time
Dariya Glazer- Stanford Medical School
No additional authors
Short Abstract: Over the course of a simulation molecules undergo structural perturbations which pose challenges in assessing similarity of sites across different time points. To this end, a clustering algorithm is described, that evaluates whether points located in different structures of the ensemble generated by the simulation lie in equivalent local environments.
Long Abstract: Click Here

Poster Q45
PIPA: an Integrated and Automated Pipeline for Genome-wide Protein Function Annotation
Chenggang Yu- Biotechnology HPC
Nela Zavaljevski (Biotechnology HPC Software Applications Institute, Telemedicine and Advanced Technology Research Center, US Army Medical Research and Materiel Command, ); Valmik Desai (Biotechnology HPC Software Applications Institute, Telemedicine and Advanced Technology Research Center, US Army Medical Research and Materiel Command, ); Seth Johnson (George Mason University, ); Fred Stevens (Argonne National Laboratory, Biosciences Division); Jaques Reifman (Biotechnology HPC Software Applications Institute, Telemedicine and Advanced Technology Research Center, US Army Medical Research and Materiel Command, );
Short Abstract: The Pipeline for Protein Annotation (PIPA) predicts genome-wide protein functions by integrating the results of multiple resources into consensus Gene Ontology terms. PIPA’s major components include: an algorithm to generate protein profiles, an automated ontology mapping between various ontologies, and a consensus algorithm to reconcile annotations from integrated resources.
Long Abstract: Click Here

Poster Q46
TOPS++FATCAT: fast flexible structural alignment using constraints derived from TOPS+ strings models
Mallika Veeramalai- Burnham Institute for Medical Research
Yuzhen Ye (Indiana University, School of Informatics); Adam Godzik (Burnham Institute for Medical Research, Bioinformatics and Systems Biology Research Program);
Short Abstract: TOPS++FATCAT: a fast flexible structural alignment using constraints derived from TOPS+ strings models. Intuitive topological constraints help to prune the search space involved in FATCAT comparison process. The TOPS++FATCAT provides FATCAT accuracy and insights into protein structural changes at a speed comparable to sequence alignments towards interactive structure similarity searches.
Long Abstract: Click Here

Poster Q47
Scoring Confidence Index: Statistical Evaluation of Ligand Binding Mode Predictions
Maria Zavodszky- Michigan State University
Andrew Stumpff-Kane (Michigan State University, Department of Biochemistry and Molecular Biology); David Lee (Michigan State University, Lyman Briggs College); Michael Feig (Michigan State University, Department of Biochemistry and Molecular Biology);
Short Abstract: We developed a confidence measure of scoring performance in ranking docked ligand poses without using knowledge of the correct binding mode. The false prediction rate was only 12% for complexes with high scoring confidence index, more than four times lower than for cases with low confidence index.
Long Abstract: Click Here

Poster Q49
A novel method for the detection of protein local structural motifs binding specific ligand fragments
Gabriele Ausiello- University of Rome "Tor Vergata"
Pier Federico Gherardini (University of Rome "Tor Vergata", Biology); Elena Gatti (University of Rome "Tor Verata", Biology); Manuela Helmer-Citterich (University of Rome "Tor Verata", Biology);
Short Abstract: We present an algorithm for the comparison of protein binding pockets that identifies small structural motifs binding specific ligand fragments. We applied this method to all proteins of known structure, identifying 657 motifs. Some of these are present in as many as 60 folds.
Long Abstract: Click Here

Poster Q50
CS23D: A web server for rapidly generating protein structures from chemical shift data
David Wishart- University of Alberta
David Arndt (University of Alberta, Computing Science); Mark Berjanski (University of Alberta, Computing Science); Peter Tang (University of Alberta, Computing Science); Ben Zhou (University of Alberta, Computing Science); Guohui Lin (University of Alberta, Computing Science);
Short Abstract: CS23D (Chemical Shift to 3D Structure) is a web server for rapidly generating accurate 3D protein structures using only assigned NMR chemical shifts as input. Unlike conventional NMR methods that require NOE and/or J-coupling data, CS23D uses only chemical shift information to generate a 3D structure of the protein of interest.
Long Abstract: Click Here

Poster Q51
Predicting zinc-binding sites in proteins from amino acids sequences
Nanjiang Shu- Duke University
Tuping Zhou (Stockholm University, Structural Chemistry); Sven Hovmöller (Stockholm University, Structural Chemistry);
Short Abstract: Zinc is one of the most common transition metals that exist in all organisms and it plays key roles in a variety of biological processes. We have developed a new method for predicting zinc-binding sites in proteins from amino acid sequences. Our method outperforms the best previously published methods.
Long Abstract: Click Here

Poster Q52
Protein Secondary Structure Prediction based on physical properties of amino acids in multiple sequence alignments
Saraswathi Sundararajan- Laurence H. Baker Center for Bioinformatics and Biological Statistics,IAState, USA
Robert L Jernigan (Laurence H. Baker Center for Bioinformatics and Biological Statistics,IAState, USA, Biochemistry, Biophysics, and Molecular Biology (BBMB));
Short Abstract: To improve Protein Secondary Structure Prediction based on physical properties of amino acids in multiple sequence alignments (MSA), global and local aggregate values of the properties are compared in order to discern patterns which will enable the interpolation of possible compositions of amino acids for each residue in an MSA.
Long Abstract: Click Here

Poster Q53
Dossier of Secondary Structure Elemnts Amino Acid Properties
Goran Neshich- EMBRAPA - CNPTIA
Ivan Mazoni (Embrapa, Laboratorio de Biologia Computacional); Luis Cesar Borro (Embrapa, laboratorio de Biologia Computacional); Daniel Alvaranga (Embrapa, laboratorio de Biologia Computacional); Pablo Lira Cecilio (Embrapa, laboratorio de Biologia Computacional); Jose Gilberto Jardine (Embrapa, laboratorio de Biologia Computacional); Adauto Mancini (Embrapa, laboratorio de Biologia Computacional);
Short Abstract: Investigating secondary structure elements might shade some light on the process of protein folding. We present here our data on secondary structure elements (SSE) taken from the whole PDB and analyzed in terms of the structure descriptors, previously stored in the STING_RDB.
Long Abstract: Click Here

Poster Q54
Analyzing Coiled Coil Proteins With Support Vector Machines to Design New Anti-Viral Drugs
Ingrid Abfalter- Johannes Kepler University Linz
Carsten Mahrenholz (Charité Medical School, Institute of Medical Immunolog); Ulrich Bodenhofer (Johannes Kepler University Linz, Institute of Bioinformatics); Sepp Hochreiter (Johannes Kepler University Linz, Institute of Bioinformatics);
Short Abstract: We use support vector machines and statistical methods to extract important amino acid sequence patterns that determine the olgimerization state of coiled coil proteins. These patterns are then used for a design tool that suggests alternative binding partners to viral fusion proteins.
Long Abstract: Click Here

Poster Q55
Prediction of Drosophila gene functions using Support Vector Machines
Nicholas Mitsakakis- University of Toronto
Zaheer Razak (The University of Toronto, Mississauga, Dept of Cell and Systems Biology and the Canadian Drosophila Microarray Centre); Michael D. Escobar (University of Toronto, Dept of Public Health Sciences); J. Timothy Westwood (The University of Toronto, Mississauga, Dept of Cell and Systems Biology and the Canadian Drosophila Microarray Centre);
Short Abstract: We have developed a method for the prediction of novel Gene Ontology Biological Processes (GO-BP) for the large fraction of un-annotated Drosophila genes. A series of Support Vector Machines (SVM) were trained using microarray expression data for annotated genes and the trained systems used for the prediction of un-annotated genes.
Long Abstract: Click Here

Poster Q56
Evolution of structural/functional features in base-excision repair proteins using structural and sequence conservation between sites
Ramiro Barrantes-Reynolds- University of Vermont
Susan W. Wallace (University of Vermont, Microbiology and Molecular Genetics); Ian Odell (University of Vermont, Microbiology and Molecular Genetics); Jeffrey P. Bond (University of Vermont, Microbiology and Molecular Genetics);
Short Abstract: We used sequences and structures of base excision repair enzymes to infer changes in the structural and functional roles of amino acids during evolution.
Long Abstract: Click Here

Poster Q57
Lightweight comparison of RNAs based on exact sequence-structure matches
Steffen Heyne- Albert-Ludwigs-University Freiburg
Sebastian Will (Albert-Ludwigs-University Freiburg, Intitute of Computer Science); Michael Beckstette (Albert-Ludwigs-University Freiburg, Intitute of Computer Science); Rolf Backofen (Albert-Ludwigs-University Freiburg, Intitute of Computer Science);
Short Abstract: Specific functions of RNA molecules are often associated to different motifs in RNA structures. We present a new pairwise RNA sequence-structure comparison method maintaining identical substructures. The results of our experiments are in good agreement with existing alignment-based methods, but were obtained in a fraction of running time.
Long Abstract: Click Here

Poster Q58
Automated Protein Function Prediction Using Extended Similarity Group (ESG) of Sequences.
Meghana Chitale- Purdue University
Troy Hawkins (Purdue University, Biological Sciences); Changsoon Park (Chung-Ang University, Statistics); Daisuke Kihara (Purdue University, Biological Science, Computer Science);
Short Abstract: Extended Similarity Group (ESG) method annotates query sequences with Gene Ontology terms by assigning probability to each annotation computed based on iterative PSI-BLAST searches. By performing iterative PSI-BLAST searches beginning from the query sequence we can cover the domain of sequences homologous to the query sequence in the sequence space.
Long Abstract: Click Here

Poster Q59
Combining Predictions of Protein Structure and Protein-RNA Interaction to Model The Structure of the Human Telomerase Complex
Ben Lewis- Iowa State University
Deepak Reyon (Iowa State University, Bioinformatics and Computational Biology); Jae-Hyung Lee (Iowa State University, Bioinformatics and Computational Biology); Vasant Honovar (Iowa State University, Computer Science); Drena Dobbs (Iowa State University, Genetics, Developmental, and Cell Biology); Andrzej Kloczkowski (Iowa State University, Plant Sciences Institute); Mateusz Kurcinski (Iowa State University, L.H. Baker Center for Bioinformatics and Biological Statistics); Robert Jernigan (Iowa State University, L.H. Baker Center for Bioinformatics and Biological Statistics); Andrzej Kolinski (University of Warsaw, Labratory of Theory of Biopolymers);
Short Abstract: Telomerase is a ribonucleoprotein enzyme pivotal in cellular senescence. Despite its importance, high-resolution structures of the enzyme with its RNA component have not been published. This study uses machine learning predictions of RNA-binding residues, along with advanced template-based protein structure prediction, to develop a model for the human telomerase complex.
Long Abstract: Click Here

Poster Q60
Transmembrane Topology and Signal Peptide Prediction using Dynamic Bayesian Networks
Sheila Reynolds- University of Washington
Lukas Kall (University of Washington, Genome Sciences); Michael Riffle (University of Washington, Biochemistry); Jeff Bilmes (University of Washington, Electrical Engineering); William Noble (University of Washington, Genome Sciences);
Short Abstract: This work presents a state-of-the-art combined transmembrane
topology and signal peptide predictor which takes full advantage of the
flexibility allowed in the dynamic Bayesian network framework.
Each topology prediction includes a set of confidence scores and predictions for all
6.3 million proteins in the Yeast Resource Center database are now publicly available.
Long Abstract: Click Here



Accepted Posters
View Posters By Category
Search Posters:
Poster Number Matches
Last Name
Co-Authors Contains
Title
Abstract Contains