AN IMPROVED APPROXIMATION FOR ASSESSING THE STATISTICAL SIGNIFICANCE OF MOLECULAR SEQUENCE FEATURES

Sabine Mercier (1) , Dominique Cellier (2) , François Charlot (2)

1 IMT - Institut de Mathématiques de Toulouse UMR5219
2 LMRS - Laboratoire de Mathématiques Raphaël Salem

Using random walk theory, we first establish explicitly the exact distribution of the maximal partial sum of a sequence of independent and identically distributed random variables. This result allows us to obtain a new approximation of the distribution of the local score of one sequence. This approximation improves the one given par Karlin et al., which can be deduced from this new formula. We obtain a more accurate asymptotic expression with additional terms. Examples of application are given.

Mots clés: statistical significance; sequence analysis; local score; random walks

Domaines: Statistiques [math.ST]; Théorie [stat.TH]

Sabine Mercier, Dominique Cellier, François Charlot. AN IMPROVED APPROXIMATION FOR ASSESSING THE STATISTICAL SIGNIFICANCE OF MOLECULAR SEQUENCE FEATURES. Journal of Applied Probability, 2003, 40, pp.427-441. ⟨hal-00937529⟩