Description succincte
(Anglais)
|
The Protein Structure Bioinformatics group is mainly focusing on the development of methods for protein structure modeling and their application to biological problems. The protein structure bioinformatics group was established at the Biozentrum Basel in November 2001. During 2002, our main focus was to build up a work group, to start a teaching program in Basel, to put an IT infrastructure in place and to establish technical and scientific collaborations. The infrastructure for SWISS-MODEL server and the DeepView server, that was previously provided by GlaxoWellcome Geneva,needed to be established in Basel to allow uninterrupted service via the ExPASy web server.
|
Résumé des résultats (Abstract)
(Anglais)
|
IT-Infrastructure : Basel Computational Biology Center
We established a collaboration between the Friedrich Miescher Institute (FMI) and the Biozentrum in the field of information technology. This partnership, named Basel Computational Biology Center (BC)2, includes the organization of meetings and conferences, sharing bioinformatics IT infrastructure via a glass fiber connection and provides synergies in system administration and user support. With financial aid of Novartis, we were able to install a 48-CPU Beowulf Linux Cluster at the Biozentrum. The BC2 infrastructure includes Oracle databases and application servers.
BC2 Basel Computational Biology Conference 2003
In 2003, we have organized the first Basel Computational Biology Conference: "Life Sciences meet IT". About 170 international visitors from IT, biotech companies, pharma industry, universities, and research institutions attended this two-day conference, covering a broad range of aspects at the interface between life sciences research and IT. Additionaly, a yearly plenary lecture series featuring internally well known scientists was initiated. In 2003, the first lecture was given by Ron Davies (Stanford Genome Center). These activities will be continued in the future to sharpen the internationally visible scientific profile of the Swiss Institute of Bioinformatics.
SWISS-MODEL Server : Comparative Protein Structure Modeling
Our group develops, maintains and supports the SWISS-MODEL server, the most widely used web-based homology modeling server worldwide. The server is accessible directly from the SIB's ExPASy server, or as links from withing individual Swiss-Prot entries.
Each year, SWISS-MODEL processes more than 120'000 requests for 3-dimensional protein models. To ensure and continuously evaluate the quality of the modeling results, SWISS-MODEL was the first homology modeling server to join the EVA project (Rost et. al. Rockefeeler University USA) for automated large scale evaluation of protein prediction servers. The refinement of automated homology modelin techniques is a major focus or our group.
SWISS-MODEL Repository
We are developing the "SWISS-MODEL Repository" as an Oracle-based database of annotated comparative protein structure models. It contains the results of automated large-scale modeling for sequences from the SwissProt/TrEMBL database using the SWISS-MODEL pipeline. In 2003, we were able to develop a modified computational SWISS-MODEL pipeline adapted to large-scale modeling. The automated modeling workflow was mapped to the model database to allow effective control fo the modeling uptdate cacles by the database.
The SWISS-MODEL Repository is directly cross referenced from Swiss-Prot entries on ExPASy. We are currently working on the integration of structural models with other biochemical knowledge databases, e.g. GermOnline. Collaboration with the Swiss-Prot group at the EBI has been established to incorporate structural models in the InterPro database as part of the TEMBLOR project supported by the European Community.
DeepView - Swiss-PdbViewer (spdbv)
Furthermore, we collaborate with Nicolas Guex (GlaxoSmithKline) in the development of DeepView - the Swiss PdvViewer, a widely used tool to display and analyze macromolecular structures. We have established a network service which allos users to automatically import protein structures, modeling templates, Swiss-Prot sequences from within DeepView. Also, this service provides additional computational services for DeepView users, e.g. BLAST similarity searches vs. Swiss-Prot or ExPDB or PROSITE definition lookup. In 2003, this new service has been widely accepted by the scientific community and is now accessed several hundred times per week.
Reserach Project: Structural analysis of non-synonymous SNPs.
One major research project in our group is the structure-based anaylsis of sequence variations to provide insight in the molecular basis of disease-causing mutations. Our preliminary studies on mutations of human phenylalanine hydroxlase causing phenylketonuria showed that a computational approach based on protein structure models could significantly contribute to an objective quantification of mutational effects.
|
