Paper detail

G-protein coupled receptor subfamily identification using phylogenetic comparison of gene and species trees

Most approaches to prediction of protein function from primary structure are based on similarity between the query sequence and sequences of known function. This approach, however, disregards the occurrence of gene duplication (paralogy) or convergent evolution of the genes. The analysis of correlated proteins that share a common domain, taking into consideration the evolutionary history of genes under study may provide a more reliable annotation of predicted proteins with unknown function. A computer program that enables real-time comparison of 'gene trees' with 'species trees' was developed. The Phylogenetic Genome Annotator (PGA) performs a profile based multiple sequence alignment of a set of sequences that share a common domain to generate a phylogenetic gene tree, which is compared to the species phylogeny inferred from aligned ribossomal RNA data. The correlated protein domains are then displayed side-by-side with the phylogeny of the corresponding species. The statistical support of gene clusters (branches) is given by the quartet puzzling method. This analysis readily discriminates paralogs from orthologs, enabling the identification of proteins originated by gene duplications and the prediction of possible functional divergence in groups of similar sequences. The tool was tested in three distinct subfamilies of the G-protein coupled receptor superfamily. In the analysed datasets, the paralogy prediction agreed with the known subfamily grouping, suggesting that subfamily divergence was facilitated by duplication events in the ancestral nodes.