Identifying short motifs by means of extreme value analysis

摘要: The problem of detecting a binding site —a substring DNA where transcription factors attach— on long sequence requires the recognition small pattern in large background. For short sites, matching probability can display fluctuations from one putative to another. Here we use self-consistent statistical procedure that accounts correctly for deviations predict location sites. We apply it two distinct situations: a) detection sites three specific set 134 estrogen-regulated genes; b) identification, 138 possible factors, ones nine genes. In both instances, experimental findings are reproduced (when available) and number false positives is significantly reduced with respect other methods commonly employed.