National Grid Seminar Series
Jointly organized with HP Singapore

The Vital-IT Projects & Its Developments
Prof. Ernest Feytmans
(Swiss Institute of Bioinformatics)

HPC Needs in Proteomics
Prof. Ron Appel
(Swiss Institute of Bioinformatics & GeneBio)

22 October 2004 (Friday)
0915 hours @ The Big One
Institute for Infocomm Research
21 Heng Mui Keng Terrace, Singapore 119613

Abstract for “The Vital-IT Projects & Its Developments”

The Vital-IT project is an innovative joint venture between academic and industrial partners, whose shared goal is to develop high-performance computing tools and infrastructure to support the needs of the life sciences. The academic partners are the Ludwig Institute for Cancer Research, the Universities of Lausanne, Geneva, Basel and the Swiss Federal Institute of Technology (EPFL). The industrial partners are Hewlett Packard Company and Intel Corporation.

A scalable and evolving cluster of high-performance servers based on the Intel Itanium® family chips was chosen. A continuing partnership with Intel Corporation and Hewlett Packard Company ensures that the hardware is always kept up to the latest version of the chip. Initially, a development cluster with 12 processors and a production cluster with 64 processors have been deployed. The production cluster has an aggregate performance of about 700 GFlops. The clusters are interconnected through a high-bandwidth network, and have access to 2 terabytes of online storage. Significant upgrades of the hardware are planned throughout the lifetime of the project, which should boost the performance well over two-fold.

A dedicated team of software engineers is adapting to the Vital-IT hardware a wide palette of programs that have proven their value to the life science community. This includes software for sequence analysis, statistical analysis, molecular modeling, database searching, comparative genomics, etc. The goal in each case is to ensure that the architecture of the Itanium processor is used optimally, and that tasks are distributed efficiently among the servers in the cluster. Software developed by researchers at the SIB and the Vital-IT partner institutions is given priority, for example pftools (P. Bucher), T-COFFEE (C. Notredame), ALDentE and Popitam (PIG group).

The software and infrastructure developed in the project have to be immediately useful to the community. In some cases, this is done by providing more powerful resources behind the existing Web servers of the SIB, ExPASy and EMBnet-CH. In other cases, Vital-IT develops dedicated working environments, in particular for the technology platforms (DNA chips, proteomics) of its partner institutions. Early projects submitted to Vital-IT include a statistical analysis of the contribution of the expression of multiple genes on cancer phenotypes (Asa Wirapati, ISREC), the computation of the binding free energy of TCR/MHC/peptide complexes (Olivier Michielin, LICR), the identification of retro-transposed genes in vertebrate genomes (Henrik Kaessmann, UniL), and the rapid clustering of local sequence alignments for the classification of protein families (Marco Pagni, SIB). It is expected that a steady flow of research projects will continue to use the high-performance computing environment and the software consulting services provided by the Vital-IT group.

Biodata of Ernest Feytmans

Ernest Feytmans graduated in Cell Biology (1968) and in Statistics (1971) at the Catholic University of Louvain, Belgium. He received his Ph.D. in 1973 for a thesis in Cell Biology with Jacques Berthet and Christian de Duve and he carried out a postdoctoral fellowship with Christian de Duve at the Rockefeller University. He was professor at the Catholic University of Santiago in Chile until 1975 and then full professor at Namur University, Belgium, heading the Department of Biology and the Unit of Structural Molecular Biology. He was also Chancellor from 1982 to 1987 and Dean of the Faculty of Sciences from 1991 to 1997. He was then appointed at the University of the West Indies as Dean of the Faculty of Medical Sciences and Head of the Biochemistry Unit. His major research interests have focused more recently on the prediction of protein 3D-structures using atomic mean force potentials and he developed a new methodology to assess the quality of protein 3D structures. Since 2001, he is the Director and General Manager of the Swiss Institute of Bioinformatics, member of its executive board and chairman of its Group Leader’s Council. He is also Affiliated Scientist of the ISREC (Swiss Cancer Research Institute) and Professor of the Faculty of Biology and Medicine of the University of Lausanne. He published more than 100 articles in peer reviewed journals and he is member of several scientific advisory boards in Switzerland and abroad and member of scientific editorial boards.

Abstract for “HPC Needs in Proteomics”

Proteomics aims at deciphering the proteome, the complement of the genome, with the goal of increasing the understanding of biological processes, as well as improving and speeding up the development of drugs by discovering disease biomarkers and drug targets. The major elements of proteome analysis are powerful protein separation techniques such as liquid chromatography (LC) and two-dimensional electrophoresis (2-DE) associated to enzymatic processing and mass spectrometry. These techniques have requested considerable efforts in the development of dedicated bioinformatics. Proteomics and proteome informatics research has been extensively developed for twenty years in Geneva, both at the Swiss Institute of Bioinformatics (SIB) and at Geneva Bioinformatics (GeneBio) SA, providing thousands of researchers worldwide with comprehensive and state-of-the-art software tools and databases for proteome analysis. In particular, the Proteome Informatics Group at SIB has been working, in part in collaboration with GeneBio that acts as its worldwide distributor, on various aspects of bioinformatics for proteomics, in particular on the image analysis of 2-DE and LC-MS data and on protein identification and characterization from MS (PMF) and MS-MS (PFF) data, as well as on proteome databases. More and more though, proteomics researchers are faced with the problem of having to analyze huge amounts of data, typically thousands or millions of mass spectra in one experiment, revealing the need for high performance computing facilities. We shall present past experience we have had in Geneva in high-throughput proteome informatics, as well as current and future needs.

Biodata for Ron D. Appel

Ron D. Appel graduated in Computer Sciences at the University of Geneva. He received his Ph.D. in 1987 at the same university for a thesis on the application of artificial intelligence and machine learning to the analysis of two-dimensional protein maps. After a postdoctoral fellowship at Harvard School of Public Health’s Clinical Decision Making Group, he set up his own group at the Numerical Imaging Unit of Geneva University Hospital, carrying out work on bioinformatics applied to proteomics and in particular in computer analysis of two-dimensional gel electrophoresis and protein identification from mass spectrometric data.

In 1998 Ron Appel co-founded the Swiss Institute of Bioinformatics (SIB), a non-profit academic foundation where he serves as a member of the Council. His research group at the SIB, known as the Proteome Informatics Group, is performing research and development work on differential analysis of proteomes and automation of proteome analysis.

Ron Appel is professor of bioinformatics at the University of Geneva associated to the Department of Computer Science of the Faculty of Sciences and the Department of Structural Biology and Bioinformatics of the Faculty of Medicine.

He is also scientific co-founder of two biotechnology companies in Geneva: one in bioinformatics – Geneva Bioinformatics (GeneBio) - and the other one in high-throughput proteomics – GeneProt. Prof Appel is currently the chairman of the Board of GeneBio and member of the Scientific Advisory Board of GeneProt. He is also a member of the Council and of the Executive Board of the Health On the Net Foundation (HON).

Ron Appel is author or co-author of more than 60 articles in journals with editorial policy, 20 books or book chapters and 200 other articles or abstracts. He serves on the editorial board of several scientific journals, notably as senior editor of PROTEOMICS.

---
This seminar series is organized by the National Grid Office located at 21 Heng Mui Keng Terrace, Singapore 119613.

Admission is Free. All are Welcome.
Please email to Miss Vasugi (email: vasu@ngp.org.sg) to register a place by 1000 hours on 21 October 2004.