Even in apparently “orphan” proteins, there often remains traces of sequence identity that can yield wonders thanks to modern tools, and to biological insight. In fact, you can very well identify structural and functional domains below 10% sequence identity, even in the complete absence of a template with known structure. I will show some striking examples, in particular of how to identify highly distant homologs using a combination of very powerful new tools (profile-profile comparison : HHpred, HHblits, csblast) and of “contextual” information (e.g. taxonomy and gene order). I will also show how to rapidly and robustly predict other sequence features, such as transmembrane segments, cysteine bridges... that give invaluable information on the function of your proteins. All these tools are available in user-friendly, web-based version, so you needn’t be afraid any more of sequence analyses.

Contact : ibs.seminaires@ibs.fr