How to meme groups of sequences to find a motif that is present in at least one sequence in each group?

Question

[Soumitra Pal]

I have a specific usage scenario for MEME. I have a set of sequences that are divided into multiple subsets (or groups). For example, each sequence is an exon for a particular gene that has orthologs in multiple species. Exons from a fixed species form a group.

I would like to find motifs that are present in at least one sequence in each group. In the example, we need to find a motif that is present in at least one exon in each of the species.

Could you please let me know how to invoke meme (from the command line or website) in this special scenario?

Answer 1

[Soumitra Pal]

I got the following reply from Charles Grant (maintainer of MEME-Suite):

Hi Soumitra,

Tim Bailey (the author of MEME) has offered a suggestion that didn’t occur to me:

You could concatenate the exons for a single species into one sequence, using N’s to mark the boundaries. You could then combine those sequences into a single file, and run with MEME using the OOPS (one occurrence per sequence) model. Any motifs that MEME found would occur in at least exon per species.

MEME may not be able to generate great results with this, because the sequences are likely to be quite long, but it may work. In the next few months we’ll be releasing a new tool that might cope better with long sequences.

Charles