Empirical codon models for comparative re-sequencing data

Darwinian selection is an important source of evolutionary innovation and a major force behind the divergence of species. Consequently, a wide variety of methods have been developed to detect genes that have been subject to selection, including comparative or phylogenetic methods that utilize patterns of substitutions between species. For example, standard likelihood ratio tests for positive selection have been developed that are based on codon substitution models. However, if applied to closely related species such as primate genomes, these tests lack power, and therefore very few genes show signs of positive selection. Incorporating additional information, such as patterns of intraspecific polymorphism, promises to improve the detection of positive selection. With the emergence of new sequencing technologies these data are now available. Indeed, for several species, including Human, Drosophila and Arabidopsis, 1,000 genomes will be available soon. However, it remains unclear whether the probabilistic methodologies previously used in phylogenetics and in population genetics are suitable to analyse these data sets. The proposed project includes both a theoretical and an applied component that will provide bioinformatic tools and biological knowledge on the evolution of protein coding genes. The theoretical component will aim at developing new empirical codon models. We will develop new algorithms to estimate empirical models that take into account substantial site and lineage specific rate variation in comparative polymorphism data (i.e., sequences from several species and multiple individuals). This part of the project will therefore be geared towards implementing the model and testing it using computer simulations and empirical data sets such as the mammalian phylogeny to identify its underlying properties. The software developped will be made available to the public as open source software. In the applied component we will use the empirical codon models and their extensions to understand the evolutionary processes on the Drosophila phylogeny. Taking advantage of the 12 Drosophila genomes data and the 1000 D. melanogaster project we will first analyse the melanogaster subgroup. Furthermore sequencing of D. ananassae populations will be performed to study another clade with different substitution patterns. Tests for Darwinian selection will be performed on both clades.

Darwinian selection is an important source of evolutionary innovation and a major force behind the divergence of species. The selective forces acting upon a protein are highly informative about its biological function and evolutionary history. Consequently, a wide variety of methods have been developed to detect genes that have been subject to selection. However, if applied to closely related species such as primates, these tests lack power. With the advent of next generation sequencing (NGS) technologies genomes, both from closely related species and from multiple individuals of the species, are increasingly available. Indeed, for several species, including Human, Drosophila and Arabidopsis, 1,000 genomes are available. These large amounts of data offer a great opportunity to study speciation and the evolutionary history of populations, provided we can properly model the process of evolution within and between species simultaneously. For our FWF project, we have developed Polymorphism-aware phylogenetic Models (PoMos). A substitution is hereby modeled through a mutational event followed by a gradual fixation. Polymorphisms can either be observed in the present (tips of the phylogeny) or be ancestral (present at inner nodes). With this approach, we naturally account for incomplete lineage sorting and shared ancestral polymorphisms that is a major concern for species tree estimation. Our method can accurately and time-efficiently estimate the parameters describing evolutionary patterns for phylogenetic trees of any shape (species trees, population trees, or any combination of those). We have implemented software package called IQ-TREE-PoMo. We demonstrated that PoMo is suitable to infer large scale phylogenies from population data of primate species. The new software will enable biologists to better understand speciation events.