Computational and mathematical models for the evolution of genomic architecture

Orateur : Carole KNIBBE (Université Lyon 1 - LIRIS - INRIA "Beagle" team)

Even though numerous genome sequences are now available, the evolutionary mechanisms that determine the evolution of genes and genomes are still debated. While comparative genomics provides us with complete snapshots of extant genomes, they tell us almost nothing about their dynamics, which can only be inferred by phylogenetic methods. An important problem is that phylogenetic methods rely on assumptions about evolution and natural selection that can hardly be tested, and which could yield to biases in the inferred dynamics. Here, I will present an alternative approach, closer to experimental evolution, where we develop mathematical models and simulations of genome evolution independently of a specific phylogenetic method, in order to test whether the classical assumptions about genome evolution hold. I will focus on the dynamics of gene number and gene density and show that counter-intuitive behaviors can result from simple setups. For example, I will show that the spontaneous dynamics of genome size naturally prevents it from growing infinitely, even if the duplication rate is twice the deletion rate and natural selection directly favors the longest genomes.

Télécharger dans mon agenda