Research directions
Overall
Most of our current research activities center on one theme, extracting biological knowledge relevant to cancer research from raw data produced by the genome projects, especially from the human genome working draft. At a more practical level, we develop sequence analysis algorithms, computer programs and biological databases, and apply these tools to particular problems. We also perform more fundamental research on genome and protein structure and evolution.
In response to the recent developments in large microarray-based gene expression data sets, we have recently started several collaborations with other groups in the Lausanne area to acquire the know-how and the tools for exploiting microarray-based gene expression data. We also made a modest first move in the direction of systems biology by using computer simulations to interpret the results from immunological experiments
Protein sequence analysis
Proteins that are similar in sequence often serve similar biological functions. This is why the Institute has been working on a method called "generalized profile" that serves for the identification and characterization of divergent protein homology domains. The method is implemented in a software package called
pftools [1], which is available from our FTP site Initially, we used generalized profiles mainly for the discovery of novel protein homology domains. A recent example is the apoptosis-related CARD (caspase recruitment) domain. Another important application is the identification of previously unnoticed members of a known domain family
Gene discovery in databases
The addition of a frame-shift error-tolerant algorithm for searching DNA sequences with a protein profile to the pftools package, allowed us to scan the large EST sequence collections, made of short mRNA sequence fragments containing many errors (single-base insertions and deletions breaking the reading frame of a coding region). We discovered many interesting new genes related to apoptosis and cell cycle regulation However, with the latest increase in EST sequence data this method became too slow
Development of new, yet hypothetical protein databases and search for motifs
In order to exploit the information contained in EST and genomic databases, the institute is producing databases of the likely coding regions of those DNA sequences. The development of a new concept was needed to extract potential coding regions from error-prone EST sequences. In conjunction with a gene prediction program (Genscan) the new
ESTscan [2] programme made possible the creation of both an EST based database (
TrEST) [3] and a genome based database (
TrGen) [4] which are distributed from our server
Using the above mentioned concept of (generalized) profiles on the hypothetical protein databases made possible the development of a new powerful gene discovery instrument: "
Hits" [5]. This tool provides access, via a web interface, to an updated list of pre-computed hits of all our current profile collections on the protein databases. "Hits" enable biologists to discover similarities between proteins at a complex motif level rather than at the usual amino acid level. Moreover, the search of
metamotifs [6] (combination of motifs) can further improve this new technique
Bioinformatics of gene regulation
With the advent of complete genome sequences and microarray-based global gene expression profiles, the analysis of gene regulatory regions has become a major focus in computational sequence analysis. The Institute is in a good position to conduct research in this area as we have developed the
Eukaryotic Promoter Database (EPD) [7] for many years. EPD remains one of the fundamental databases in the study of the gene regulatory regions by comparative sequence analysis
Another aspect of the research in gene regulation using bioinformatics tools is based on a collaboration wiht a laboratory of the University of Lausanne. It involves the modelling of the binding site of transcription factors. New algorithms were needed to correlate the data, obtained from oligonucleotide - transcription factor binding experiments with the profiles parameters representing the binding region
Useful Links
[1] prtools: http://www.isrec.isb-sib.ch/ftp-server/pftools/pft2.2
[2] ESTScan: http://www.ch.embnet.org/software/ESTScan_help.html
[3] TrEST: ftp://ftp.ch.embnet.org/pub/databases/trest
[4] TrGen: ftp://ftp.ch.embnet.org/pub/databases/trgen
[5] Hits: http://hits.isb-sib.ch/>Hits
[6] metamotif: http://www.isrec.isb-sib.ch/ftp-server/mmsearch
[7] EPD: http://www.epd.isb-sib.ch
Competences
Développement d'algorithmes
Analyse de séquence d'ADN et de protéines
Edition d'un journal électronique "Cancer Immunity"
http://www.cancerimmunity.org
