Entering edit mode
6.0 years ago
Lucy
▴
150
Hi all,
I am performing Homer findMotifsGenome.pl to identify motifs enriched in a subset of ATAC-seq peaks (~2500 peaks) vs. a control set of ATAC-seq peaks (~4000 peaks). I ran the command using either the binomial distribution to score motifs or hypergeometric enrichment scoring.
Changing the scoring approach resulted in very different motifs being identified as enriched. Which method should I be using in my case, and which results can I trust more?
Best wishes,
Lucy
Did you read the manual?
Thank you - yes I have read this. Based on this, I thought the hypergeometric sounded appropriate. With binomial scoring, Homer identified 27 enriched motifs, compared to 3 using hypergeometric. The reason that I am unsure of the best approach is that elsewhere in the Homer documentation, it says to use the hypergeometric if the number of background sequences < number of target sequences, which is not the case for me.