Finding differences in motif enrichment between ATAC peak samples?

Question

[a.rex]

I have performed ATAC-seq on two different biological samples. I want to compare peaks between these files, as well as enrichment for particular motifs.

Firstly, I have mapped my PE reads, and then removed duplicates. I now have a bam file, that I use deeptools BamCoverage to normalise to RPKM, and the output is bedgraph. I do these for both my samples that I wish to compare.

Now, I want to compare the motifs that are enriched in the peaks between the two files. Is the following a sound methodology:

• Run the resultant RPKM bedgraph files through Homer's findmotifsgenome.pl. Can the output knownResults.html and and homer de novo html, be compared between the two samples to check for differences in motif enrichment?

[a.rex]

I would have to convert from bdg to bed if I wanted to use homer

Answer 1

[James Ashmore]

Take a look at the recently released diffTF (https://difftf.readthedocs.io/en/latest/index.html) - it will allow you to infer differential transcription factor binding and motif enrichment.

Answer 2

[danvoronov]

You can also use knownResults.txt file from the findMotifsGenome.pl so this could be used for comparisons.

Alternatively try RGT motif analysis tool for differential motif analysis, although I don't think it can do de-novo predictions. http://www.regulatory-genomics.org/motif-analysis/introduction/

Dan