search

SchulzLab / STARE

TF analysis from epigenetic and Hi-C data
MIT License
16 stars 2 forks source link

TFBS enrichment without promoter region #3

Closed Andi964 closed 1 year ago

Andi964 commented 1 year ago

Hello,

according to your publication: Does Stare consider +-2500 bp around the TSS in its TFBS enrichment analysis? If yes, (how) can this mode be deactivated to investigate only the enrichment in Enhancers?

Best Andreas

DennisHeck commented 1 year ago

Hello!

STARE does not consider the whole ±2500bp around the TSS for the TF affinities, but enhancer peaks that overlap with this ±2500bp window. Those peaks overlapping the promoter window are always considered, independent on whether they were linked to that gene in a preceding ABC-scoring step or not. We don't have an option to turn that off, as we found those peaks to be especially informative, and because the contact frequency in the ABC-score in that case is the contact of the region with itself. A workaround would be to remove any peaks from the candidate enhancer file that overlaps a promoter window, but then again that could remove potential enhancers of other genes.

Best wishes, Dennis

Andi964 commented 1 year ago

Many thanks for your quick answer! But in this case...I assume that your method would map the enhancers to those TSS since they are sooooo close?

Best Andreas

DennisHeck commented 1 year ago

That's very likely, yes. We didn't specifically quantify that, but at least for expressed genes the promoter should be relatively active and thus would end up with a high ABC-score. We saw however, that explicitly including all peaks in promoter range improved our expression prediction. So there are some genes where they were not initially included. Best, Dennis