Using functionally important sites to identify orthologous relations

Making accurate functional predictions for genes plays an important role in the era of proteomics. The most reliable functional information is extracted from orthologs in other species when annotating an unknown gene. Here a site-based approach is proposed to predict orthologous relations. The method first identifies important sites that confer specificity of paralogs in the multiple sequence alignment of homologous proteins. It then predicts orthologous relations for unannotated proteins based on the important sites found. When applied to the bacterial transcription factor PurR/LacI family and the protein kinase AGC group family, our method was able to identify, with few false positives, the important sites that agree with those obtained from biological experiments. We also tested it on the AGC group family, the /spl alpha/-proteasome family, the glycoprotein hormone family and the growth hormone family to demonstrate its ability to predict orthologs. Compared with other prediction methods based on phylogenetic analysis or hidden Markov models, our method not only has competitive prediction accuracy, but also provides valuable biological information of important sites associated with orthologs which can be further studied in biological experiments.