Transcription Factor Enrichment Analysis Using SEA of Meme-Suite

Question

[ilovesuperheroes1993]

I would like to run a transcription factor (TF) enrichment analysis using SEA of meme-suite. I have around 2000 query sequences in which I would like to find the enrichment. I have two questions regarding this process:

1. Should I use shuffled query sequences as control or is it better to provide random DNA sequences as control? To select random sequences, I have generated random genomic coordinates of lengths identical to my query sequences. Is this method correct?
2. As for the TF motifs, I have used the mononucleotide models (full) from HOCOMOCO database. I wanted to use the dinucleotide models, which are supposed to be more accurate, but only the mononucleotide models have the motifs in meme format.

Any inputs would be appreciated. Thanks

[Malcolm.Cook]

what process produced the 2000 sequences in the first place?

[ilovesuperheroes1993]

The 2000 query sequences are ChIP Sequencing peaks. I didn't mention it in my original post as it did not seem important to what I was asking.

[Malcolm.Cook]

with more info about the upstream experiment and analysis that generated these peaks one might suggest that you

• compare "good" (high-scoring) peaks v "bad" (low-scoring) ones.
• compare peaks that are differential in your experiment v peaks that don't change
• compare sequence under peaks against the same peaks "shifted" some number of bases in the genome (ensuring length profile is preserved)