MetaPhOrs: orthologs and paralogs accross 829 genomes, based on multiple phylogenetic evidence
We have launched MetaPhOrs, a meta-method to predict orthology and paralogy from multiple phylogenetic evidence. To maximize reliability and coverage, we retrieved gene trees from PhylomeDB, Ensembl, TreeFam and Orthogroups. In addition we built maximum likelihood trees for orthologous gene families in OrthoMCL, COG, and EggNOG. The resulting ~700.000 gene family trees provide information on orthology and paralogy relationships for over 300 million protein pairs over 829 fully-sequenced genomes. Taking advantage of the (partial) overlap between trees from different sources we compute a consistency score which, together with the number of independent sources used, is a good proxy for the reliability of the predictions.