Invited talk (3): 2:00 p.m. - 2:35 p.m.
Co-Evolutionary Analysis of Protein-Protein Interaction
(joint work with Tetsuya Sato)
Speaker
Hiroyuki Toh, Vice-Director, Computational Biology Research Center, Advanced Industrial Science and Technology (AIST), Japan
Abstract
Co-evolutionary approach is one of the bioinformatics strategy to investigate protein-protein interaction (PPI). Mirror tree method is one of such approaches, which has been developed independently by Valencia et al. [1] and Goh et al. [2]. The assumption behind the approach is that the phylogenetic trees of interacting proteins resemble due to the co-evolution, comparing to those of non-interacting proteins. In the method, the orthologous sequences of proteins under consideration are collected from the same set of the organisms at first. Then, a multiple alignment of the orthologous sequences is constructed for each protein. The distance matrices are calculated from the multiple alignments. In the actual application of the mirror tree method, the correlation coefficient between a pair of distance matrices is calculated as the intensity of co-evolution, instead of evaluating the similarity of the phylogenetic trees. The method is widely utilized for the computational study of PPI, due to the simplicity. It is known that the prediction by the method suffer from many false positives. We have developed two different approaches to reduce the false positives of the mirror tree predictions [3, 4]. Pazos et al. has independently developed a method to solve the problem too [5]. All of the approaches share the same idea, and the residuals after the removal of the information about the evolutionary relationships from the distance matrices are used to calculate the correlation coefficient. Both groups have succeeded in drastic reduction of false positives by the approaches. In this talk, we will introduce a novel approach for the prediction of protein-protein interaction in the context of the mirror tree approach.
REFERENCES
[1] Goh et al. (2000) J. Mol. Biol. 299, 283-293.
[2] Pazos and Valencia (2001) Protein Eng. 14, 609-614.
[3] Sato et al. (2005) Bioinformatics 21,3482-3489.
[4] Sato et al. (2006) Bioinformatics 22, 2488-2492.
[5] Pazos et al. (2005) J. Mol. Biol. 352, 1002-1015.