Survey of Deoxyribonucleic Acid Motif Finding Algorithms

Abstract

An important task in biology is to identify binding sites in DNA for transcription factors. These binding sites are short DNA segments which are called motifs. Given a set of DNA sequences, the motif finding problem is to detect overrepresented motifs that are good candidates for being transcription factor binding sites. The current study is a survey of motif finding algorithms. The study shows that a sensible approach to detect motif is to search for statistically overrepresented motifs in the promoter region of a set of co-regulated genes. The weak point of the available motif finding algorithms is that they tend to be sensitive to the noise, i.e., the presence of upstream sequences in data set that do not contain the motif. We conclude that instead of relying on a single motif finding tool, biologists should use a few complementary tools and pursue the top few predicted motifs of each.