On the Discovery of Significant Motifs in Genomic Sequences

In this thesis we study the statistical properties of some families of motifs of the same length. We develop a method for the approximation of the average number of frequent motifs in the family in random texts with independent characters. We give a bound on the approximation error and show that this bound is loose in practice. We develop a test which verifies whether the number of frequent motifs can be approximated to a Poisson distribution