Ancestral reconstruction under weighted maximum matching

Abstract

Ancestral genome reconstruction has attracted increasing interests from both biologists and computer scientists. It has been conducted using various evolutionary models ever since comparative genomics moved from sequence data to gene order data. We propose a Flexible Ancestral Reconstruction Model, FARM, based on the maximum likelihood and weighted maximum matching algorithms, to infer ancestral gene orders. This will accommodate various evolutionary scenarios, including not only genomic rearrangements, but also insertion/deletions (indels), segment duplications, and whole genome duplications. We evaluate this work by using various simulated evolution experiments while comparing FARM to existing methods, like InferCarsPro, GASTS and PMAG++. FARM shows significant improvement in running time and the final assembling process and, therefore, can be used in large-scale real biological data ancestral inference.