What Basic Evolutionary Patterns Can Tell

ARBIZA L., H. DOPAZO & D. POSADA

Lisboa, Portugal

Congreso; IX Jornadas de Bioinformática; 2009

Comision Hispano-Lusa de Bioinformatica

TITLE: What Basic Evolutionary Patterns Can Tell AUTHORS: Leonardo Arbiza*1, Hernán Dopazo2 and David Posada1 1Department of Biochemistry, Genetics and Immunology, University of Vigo, Spain. 2Comaprative Genomics Unit, Bioinformatics Dpt., CIPF, Spain. MOTIVATION: At a genomic level, the patterns that have shaped molecular evolution are largely heterogeneous. Variations in nucleotide composition and types of substitution rates are evident ranging from the large scale, i.e. chromosomes or chromosomal regions, to the small scale, where variations occur among, and even within, the different domains and sites that constitute functionally relevant loci. The important question arising from these differences is whether they are due to chance, have a functional or organizational basis, or result from the interplay of other processes that have molded genome evolution. In order to approach these questions, probabilistic models of nucleotide substitution can be employed to characterize some of these patterns. Moreover, in this context there has been considerable research on statistical model selection, developed in part to address the general heterogeneous nature of sequence evolution. The jModelTest program (Posada 2008) implements several criteria and strategies to select best-fit models of nucletide substitution, and incorporates model averaging as a means to address model selection uncertainty. Here we are interested in using the model selection techniques implemented in jModelTest to determine the genome wide variation in evolutionary patterns, preliminarily for protein coding loci, where GO and KEGG terms are used to evaluate their possible functional association.The overall question addressed is whether loci that evolve alike have analogous functions or are exposed to similar structural constraints. If so, evolutionary resemblance could help in further classifying different or uncharacterized sequences.