A 3-Approximation Algorithm for the Multiple Spliced Alignment Problem and Its Application to the Gene Prediction Task

Abstract

The Spliced Alignment Problem is a well-known problem in Bioinformatics with application to the gene prediction task. This prob- lem consists in finding an ordered subset of non-overlapping substrings of a subject sequence g that best fits a target sequence t.Inthiswork we present an approximation algorithm for a variant of the Spliced Align- ment Problem, called Multiple Spliced Alignment Problem,that involves more than one target sequence. Under a metric, this algorithm is proved to be a 3-approximation for the problem and its good practical results compare to those obtained by four heuristics already developed for the Multiple Spliced Alignment Problem. Keywords: