Modular Evolution of Genes and Proteins

September 21th, 2009
2 pm
LIRMM - Room 3.23

Erich Bornberg-Bauer, Institute for Evolution and Biodiversity, University of Muenster, Germany
ieb.uni-muenster.de/ebb
ebb@wwu.de

Domains are evolutionary conserved units of proteins and well characterised by Hidden Markov Models in various databases. Most proteins consist of two or more domain, giving rise to a variety of combinations of domain arrangements. While most domains occur in few combinations, some co-occur with many different ones. We investigated and characterized this phenomenon of skewed versatility by introducing a new measure which is independent from domain frequency. We also investigated genetic mechanisms of domain rearrangements and found that fusion and domain loss at the ends are way more frequent than e.g. recombination or retrotransposition. Furthermore we utilise these findings to develop recursive algorithms for rapid identification of identical or close to identical domain arrangements thus giving rise to methods which help identify homologous proteins by functionally meaningful signatures.

References: J Weiner et al. BMC Evol Biol. 8:285 2008
AD Moore et al. Trends Biochem Sci 33:444 2008
F Beaussart et al. Bioinformatics 23:1834 2007
J Weiner et al. Mol. Biol. Evol. 23:734 2006
J Weiner et al. Bioinformatics 21:932 2005
J Weiner et al. FEBS J 273:2037 2006