Reconstruction of Ancestral Genome Subject to Whole Genome Duplication, Speciation, Rearrangement and Loss

Whole genome duplication (WGD) is a rare evolutionary event that has played a dramatic role in the diversification of most eucaryotic lineages. Given a set of species known to have evolved from a common ancestor through one or many rounds of WGD together with a set of genome rearrangements, and a phylogenetic tree for these species, the goal is to infer the pre-duplicated ancestral genomes. We use a two step approach: (1) Compute a score for each possible ancestral adjacency at each internal node of the phylogeny; (2) Combine adjacencies to form ancestral chromosomes. The main contribution of our method is the computation of a rigorous score for each potential ancestral adjacency (a, b), reflecting the maximum number of times a and b can be adjacent in the whole phylogeny, for any setting of ancestral genomes. We first apply our method on simulated datasets and show a high accuracy for adjacency prediction. We then infer the pre-duplicated ancestor of a set of 11 yeast species and compare it to a manually assembled ancestral genome obtained by Gordon et al. (2009).