Simulating and reconstructing genome evolution in vertebrates
Genomes evolve through a combination of forces that are difficult to study and conceptualise. Yet the field is maturing to a point where we can now start to build models of genome evolution that capture our current understanding of chromosome and genome evolution. We have built such a model of genome evolution that accounts for chromosome rearrangements (fusions and fissions, inversions, translocations) and gene events (appearances, duplications, deletions), and we have estimated the rates at which these changes might have taken place in real evolution. We have implemented this model and these estimated rates in a computer simulator called Magsimus, and we have simulated the evolution of the chicken, opossum, dog, mouse and human genomes from a common Amniote ancestor. Using this approach, we show that it is possible to rapidly converge to optimal rates that realistically reproduce properties of modern genomes. Finally, equipped with Magsimus as a realistic simulator of genome evolution, we benchmarked an algorithm called AGORA that reconstructs ancestral vertebrate genomes. These developments provide a new framework to test evolutionary hypotheses and evaluate how incomplete our understanding of genome evolution may be.
This is joint work with Joseph Lucas, Lucas Tittmann and Matthieu Muffato.