How to find peaks with specific TFs identified by HOMER

Question

[b8921019]

Hi, I have attempted to utilize HOMER to identify peaks associated with specific transcription factors (TFs). I employed "findMotifsGenome.pl" and "annotatePeaks.pl" to extract sequences linked to these TFs. In theory, the results obtained from both methods should be identical. However, I obtained 567 regions from findMotifsGenome.pl and 2070 regions from annotatePeaks.pl. I find the last column of the annotation result quite perplexing, and I haven't come across any explanations regarding the discrepancies between these two methods. Could someone provide guidance on how I should interpret these differences? Thank you.

Answer 1

[rfran010]

It seems like the output differs slightly as to what is reported on each line, http://homer.ucsd.edu/homer/ngs/peakMotifs.html under "Finding Instance of Specific Motifs".

It appears that "findMotifsGenome.pl" may only output regions that contain TF motif instances. While "annotatePeaks.pl" will output all the regions in your regions file. I would expect that column 21 of "annotatePeaks.pl" output only shows data for 567 regions (or less if the motif occurs multiple times within the same region).

I think this explanation of the last column of the annotation results might help?

Motif Instances have the following format:

<distance from center of region>(<sequence>,<strand>,<conservation>)

i.e -29(TAAATCAACA,+,0.00)