A Faster Algorithm for Motif Finding in Sequences from ChIP-Seq Data

Motif finding in nucleotide sequences for the discovery of over-represented transcription factor binding sites is a very challenging problem, both from the computational and the experimental points of view. Transcription factors in fact recognize very weakly conserved sequence elements, that in typical applications are very hard to discriminate against random sequence similarities. Recent advances in technology like ChIP-Seq can generate better datasets to be investigated, in which the degree of conservation of binding sites is higher: on the other hand, the size itself of the datasets has posed new challenges for the design of efficient algorithms able to produce results in reasonable time. In this work we present an updated version of our algorithm Weeder, in which time and space requirements are significantly reduced and, moreover, also the accuracy of the results is notably improved.