Species Selection for Phylogeny-Based Motif Detection

Detecting conserved regions in multiple species alignment is crucial when modeling orthologous entities. However, in phylogenetic analysis of entities other than genes, for instance transcription factor binding sites (TFBS), this proves to be non-trivial due to the high functional turnover and incomplete orthology even within close species, such as Drosophila clade. Having more species does not necessarily contribute to the alignment, especially when the noise that it brings to the alignment is more than the information that it adds. Picking the correct set of species might considerably improve the accurateness of the analysis; however, in the current literature there has not been much work in identifying the most informative set of species for multiple sites.