Prediction of Conserved Motif Networks based on Comparative Genomics

Discovery of regulatory motifs in unaligned DNA sequences remains a fundamental problem in computational biology. Most motif finding algorithms perform well in finding over-represented substrings in a small collection of sequences. This requires some prior knowledge about the functional relations among genes in a study in order to select a good collection of promoters. Microarray gene profiling and Chromotin-IP are often used to provide a list of sequences that may be enriched for binding sites of some transcription factors (TFs). However, any limitation in such experiments will propagate to the motif finding step, preventing us from finding motifs for many TFs, as well as from getting a global view of motif network in the whole genome. To date, over 200 yeast TFs have been assayed, but less than 100 have confident motif predictions.