New approximate P-value of gapped local sequence alignments

Abstract

We propose a new method to approximate the significativity of gapped local sequence alignments. We focus on short sequences for which standard methods are known to be less accurate since they have been developed under asymptotics. Our approach combines an approximate distribution of ungapped local score of two sequences and a special scoring scheme that allows the insertion of gaps. For a positive integer h, the scoring scheme is defined on h-tuples of the components of the sequences and corresponds to the gapped global score. The influence of h and the accuracy of the p-value are numerically studied. To cite this article: A.M. Fayyaz et al., C. R. Acad. Sci. Paris, Ser. I 346 (2008).