Motivation: Current methods that annotate conserved transcription factor binding sites in an alignment of two regulatory regions perform the alignment and annotation step separately and combine the results in the end. If the site descriptions are weak or the sequence similarity is low, the local gap structure of the alignment poses a problem in detecting the conserved sites. It is therefore desirable to have an approach that is able to simultaneously consider the alignment as well as possibly matching site locations.
Results: With SimAnn we have developed a tool that serves exactly this purpose. By combining the annotation step and the alignment of the two sequences into one algorithm, it detects conserved sites more clearly. It has the additional advantage that all parameters are calculated based on statistical considerations. This allows for its successful application with any binding site model of interest. We present the algorithm and the approach for parameter selection and compare its performance with that of other, non-simultaneous methods on both simulated and real data.
Availability: A command-line based C++ implementation of SimAnn is available from the authors upon request. In addition, we provide Perl scripts for calculating the input parameters based on statistical considerations.