Optimization Multiple Sequence Alignment Scheme in DC-BTA

Multiple sequence alignment is a fundamental problem in computational molecular biology. This paper shows a brand new refinement strategy combining divide-and-conquer and Beam-Through alignment (DC-BTA). Optimization objective function (OF) is additively computed with new stage beam area, which is corresponding to beam area rate in [6]. The refinement is based on previous alignment result to extract heuristic start beam sources (BS). By a logical structure of P.S.M of beam status from the latest round, combined with beam weight strategy, sequences are classified into similarity class for better BS to improve refinement of later round. Some important conclusions are given on divide-and-conquer of MSA, BS collecting, and similarity degree. These conclusions promise sound and efficient optimal MSA refinements.