Evolution of Genome Organization

点击翻译按钮获取中文摘要

This is a computational biology program aiming to improve our understanding of genome evolution. The abundance of sequenced and annotated genomes present in public repositories has revealed a tremendous diversity of structures in terms of sequence, gene organization and karyotype, which is the footprint of the various mechanisms taking place during evolution: single nucleotide substitutions are at the origin of sequence diversity, duplications and other content-modifying operations are at the origin of gene family diversity and rearrangements are at the origin of gene order diversity. Each of these types of footprints has been used separately to address different questions. Sequence similarity is the main criterion used to cluster genes into families and to construct phylogenetic trees. Content-modifying operations are used to infer the orthology/paralogy relation between gene copies, which has important implications towards the annotation and functional specificity of genes. As for the diversity of gene order, it is used in comparative genomics to infer ancestral features and study the rearrangements (inversions, transpositions, translocations, fusions, fissions) taking place during evolution. In this program we seek to integrate the three types of footprints in a unified methodology for the Inference of Ancestral Genomes. This is one of the main goals of comparative genomics essential not only from an evolutionary perspective, but also for important implications of the mechanisms taking place during evolution on the genetic and physiological specificities of species, with far-reaching applications to drug discovery and personal medicine. Most existing methods for inferring ancestral genomes stand exclusively on gene order information, ignoring all available phylogenetic information on gene families, which is a serious limitation towards their applicability to genomes with multiple gene copies. The most direct benefit of an integrative method accounting for gene trees will be to avoid a drastic filtering of datasets, allowing an increase of precision of the predicted ancestral features.