Abstract: We develop a procedure called RiPE (Retrieval-induced Phylogeny
Environment) that automatically performs an evolutionary analysis of a protein
(sub)family, (i) by retrieving the relevant sequences via a homology search,
(ii) by using the search report to construct the alignment using only
homologous subsequences (taking into account their neighborhood with a low
chance of homology), (iii) by realigning, and (iv) by generating phylogenetic
trees based on the alignment. In a first implementation of our scheme, we start
with the available proteome data of model organisms, perform a PSI-BLAST
search, use MView to convert hits into a multiple alignment, and perform
realignment and tree building. As a test case, we have investigated the human
ABC transporters of the subfamily G, starting with the five known human ABCG
transporters. Our method retrieved homologous sequences not previously
analyzed, generating a tree that is more plausible and better supported than
previously published trees. The RiPE 0.1 prototype is available at the RiPE
website, http://ifg-izkf.uni-muenster.de/fuellen/RiPE/ripe.html.
