Perfect sorting by reversals

Abstract

In computational biology, gene order data is often modelled as signed permutations. A classical problem in genome comparison is to detect conserved segments in a permutation, that is, genes that are co-localised in several species, indicating that they remained grouped during evolution. A second largely studied problem related to gene order data is to compute a minimum scenario of reversals that transforms a signed permutation into another. Several studies began to mix the two problems, and it was observed that their results are not always compatible: often parsimonious scenarios of reversals break conserved segments. In a recent study, Bérard, Bergeron and Chauve stated as an open question whether it was possible to design a polynomial time algorithm to decide if there exists a minimum scenario of reversals that transforms a genome into another while keeping the clusters of co-localised genes together. In this paper, we give this polynomial algorithm, and thus generalise the theoretical result of the aforementioned paper. © Springer-Verlag Berlin Heidelberg 2005.