Item

Abstract

The number of genome sequences available for different pathogens has in- creased dramatically over the last couple of years. Existing traditional meth- ods for phylodynamic analysis scale poorly with the number of sequences. Therefore, efficient heuristics are needed to cope with the growing data sets available today. In this work, an approximate maximum-likelihood framework for phy- lodynamic analysis is developed. Its main purpose has been to estimate divergence times in large sequence alignments of rapidly evolving organ- isms. In addition, it provides a functionality to estimate ancestral states, infer evolution models, re-root trees to maximize temporal signals, and es- timate molecular clock phylogenies and population size histories. The run time for most of the developed algorithms scales linearly with dataset size. The basic application fields for the framework are studies for epidemiology and pathogen evolution, including dating cross-species transmissions, dat- ing introductions into geographic regions, and studying the time course of pathogen population sizes. In the second part of this work, I present an inference scheme for evo- lutionary models with substitution rate heterogeneity among sites. These types of models can not only result in a better approximation of the phylo- genetic reconstruction, but also predict the evolutionary forces acting along protein or DNA sequences.