Phylogenomic inference of protein molecular function.

With the explosion in sequence data, accurate prediction of protein function has become a vital task in prioritizing experimental investigation. While computationally efficient methods for homology-based function prediction have been developed to make this approach feasible in high-throughput mode, it is not without its dangers. Biological processes such as gene duplication, domain shuffling, and speciation produce families of related genes whose gene products can have vastly different molecular functions. Standard sequence-comparison approaches may not discriminate effectively among these candidate homologs, leading to errors in database annotations. In this unit, we describe phylogenomic approaches to reduce the error rate in function prediction. Phylogenomic inference of protein molecular function consists of a series of subtasks. Once a cluster of homologs is identified, a multiple sequence alignment and phylogenetic tree are constructed. Finally, the phylogenetic tree is overlaid with experimental data culled for the members of the family, and changes in biochemical function can be traced along the evolutionary tree.