Evolutionary analysis of protein diversification: model improvement, assessment and application to real data

Proteins have an essential role to play in virtually all biological processes, participating in an extraordinarily diverse set of chemical reactions and physical tasks. Consequently, there is interest within a wide variety of research disciplines to understand the origin and evolution of the functional properties of proteins. Two problems are ever-present in the analysis of proteins: (i) estimating the evolutionary history of the gene sequences, and (ii) learning about the processes that have shaped the protein's function over time. Models of protein evolution, applied within a formal statistical framework, have proved to be a powerful means of addressing these problems. Under this research program, several different approaches will be employed to improve the realism of protein models. Our primary focus will be modeling the structure of the amino acid rate matrix and its variability among sites within a gene. We will assess the impact of putative improvements with respect to model-based inferences such as phylogenetics and the identification of positively selected sites. Ultimately, the improved models will be employed to: (i) improve the precision and power of model-based methods for estimating the strength and direction of natural selection pressure, (ii) improve estimates of phylogenetic history obtained from genome-scale datasets, (iii) investigate adaptive divergence between pathogenic bacteria and their non pathogenic relatives, and (iv) assess novel metabolic diversity among uncultivable microbes. Due to the central role of proteins in so many biological processes, the impact of this research is likely to extend well beyond the above applications.