Affiliations: Russ College of Engineering and Technology, Ohio
University, Athens, Ohio, USA | Bioinformatics Laboratory, School of Electrical Engineering and Computer
Science, Ohio University, Athens, Ohio, USA
Abstract: In this paper we describe some utilizing conditions of a recently
published tool that offers two basic functions for the classical problem of
discovering motifs in a set of promoter sequences. For the first it is assumed
that not necessarily all of the sequences possess a common motif of given
length l. In this case, CHECKPROMOTER allows an exact
identification of maximal subsets of related promoters. The purpose of this
program is to recognize putatively co-regulated genes. The second, CHECKMOTIF,
solves the problem of checking if the given promoters have a common motif. It
uses a fast approximation algorithm for which we were able to derive
non-trivial low performance bounds (defined as the ratio of Hamming distance of
the obtained solution to that of a theoretically best solution) for the
computed outputs. Both programs use a novel weighted Hamming distance paradigm
for evaluating the similarity of sets of l-mers, and we are
able to compute performance bounds for the proposed motifs. A set of At
promoters were used as a benchmark for a comparative test against five known
tools. It could be verified that SiteSeeker significantly outperformed these
tools.
