Statistical Methods for Molecular Evolution

Statistical issues in evolutionary inference using aligned molecular DNA sequence data will be the focus of research. Five methodological areas of importance will be developed. In addition, working with colleagues in biology and biochemistry, methods will be applied to obtain a better understanding of the evolution of early single-celled organisms as well as how processes of selection work in pathogens. 1. Tree Inference: Understanding what the evolutionary relationships are between organisms is an important step in understanding their biology. We will derive new statistical tests of whether there is significant evidence in favour of a particular evolutionary relationship. Long term plans include developing methods that will improve Bayesian statistical inference of evolutionary relationships. 2. Positive Selection: This occurs when organisms adapt to changing environmental conditions and detecting it is of importance, for instance, in understanding how the pathogens that humans evolve resistance. We will develop tests for positive selection that better adjust for additional sources of uncertainty. It is common that selection pressure will vary across organisms and genes. Long term plans include developing models that allow varying selection pressure across gene positions and groups of organisms. 3. Protein Evolution Models: Evolution of the proteins that perform the functions of organisms is a complex process. Understanding such evolutionary processes is crucial to inference about the relationships between organisms and of interest in itself. We will develop sophisticated models that incorporate the three-dimensional structure of proteins. Long term goals include modeling the processes by which organisms gain and lose genes jointly with the evolution of those genes. 4. Distance Methods: Distance methods are a class of methods for evolutionary inference that are commonly used to infer evolutionary relationships when there are large numbers of organisms. Our work on this class of methods will be extended and we will study the statistical properties of this class of methods. 5. Diagnostics: Evolutionary inference can be adversely affected by unusual organisms or evolutionary behaviour at positions in genes. We will develop methods to detect (groups of) organisms and hot spots in genes that are unusual and/or have a large influence on inferences. Knowing what these organisms or hotspots are can be of interest in itself. Analysis after removing them can raise interesting alternative evolutionary scenarios. Application of the methods will help us to better understand the mode and tempo of evolution of early single-celled organisms as well as how positive selection on specific amino acid changes can lead to observable changes in organisms. Methods will be much more broadly applicable, however, and will be of interest to the large number of researchers interested in evolutionary biology. As has been past practice, the software developed will be publicly available.