Final Report Summary - PHYGENOM (Models of genome evolution, phylogenomics and the tree of life)
Understanding the evolution of all living organisms is one of the fundamental challenges in biology. The phylogenetic analysis of whole genomes has already proven very useful to decipher not only their history, but also their organization and function. The general aim of PHYGENOM was the development of better methodologies for the analysis of phylogenomic data and their application to different biological problems. Under this grant we implemented a Bayesian supertree approach for the estimation of species trees from multiple gene families that considers incomplete lineage sorting, gene duplication and loss and the effect of horizontal gene transfer. We also developed a model selection framework for phylogenomic data, that implements several greedy and heuristic algorithms to find optimal partition schemes for multilocus sequence alignments, and to assign specific models to the different partitions. In addition, we also developed high performance computing versions of our programs jModelTest and ProtTest that can be executed in parallel in multicore desktops and clusters in order to process large genomic data sets. Furthermore, we implemented several models for the simulation of genomic data that allowed us to compare different phylogenomic strategies. Here we found that simpler approaches worked very well, especially when the estimated gene trees were more or less good approximations of the true history. In general, we showed that we can recover reliable species trees from genomic data using approximate models of genome evolution in a reasonable timeframe. Moreover, we have also gathered different phylogenomic data sets for various groups of eukaryotes and we are concluding their analysis. Importantly, we developed several of open access bioinformatic tools that are freely available to the scientific community. Some of these tools are offered as web servers that can run on our dedicated computational cluster.