Approximating the Minimum Breakpoint Linearization Problem for Genetic Maps without Gene Strandedness

The study of genetic map linearization leads to a combinatorial hard problem, called the minimum breakpoint linearization (MBL) problem. It is aimed at finding a linearization of a partial order which attains the minimum breakpoint distance to a reference total order. The approximation algorithms previously developed for the MBL problem are only applicable to genetic maps in which genes or markers are represented as signed integers. However, current genetic mapping techniques generally do not specify gene strandedness so that genes can only be represented as unsigned integers. In this paper, we study the MBL problem in the latter more realistic case. An approximation algorithm is thus developed, which achieves a ratio of (m2 + 2m − 1) and runs in O(n7) time, where m is the number of genetic maps used to construct the input partial order and n the total number of distinct genes in these maps. Index terms — Comparative genomics, partial order, breakpoint distance, feedback vertex set.