Mining the semantics of genome super-blocks to infer ancestral architectures

The study of evolutionary mechanisms is made more and more accurate by the increase in the number of fully sequenced genomes. One of the main problems is to reconstruct plausible ancestral genome architectures based on the comparison of contemporary genomes. Current methods have largely focused on finding complete architectures for ancestral genomes, and, due to the computational difficulty of the problem, stop after a small number of equivalent minimal solutions have been found. Recent results suggest, however, that the set of minimum complete architectures is very large and heterogeneous. In fact these solutions are collections of conserved blocks, freely rearranged. In this paper, we identify these conserved super-blocks, using a new method of analysis of ancestral architectures that reconciles both breakpoint and rearrangement analyses, as well as respects biological constraints. The resulting algorithms permit the first reliable reconstruction of plausible ancestral architectures for several non-WGD yeasts simultaneously, a problem hitherto intractable due to the extensive map reshuffling of these species. See online Supplementary Material at www.liebertonline.com.