Motif Analysis searching for NFKB/CEBPB

Question

[ilyco54321]

I am trying to perform motif analysis on the promoters of a set of human genes known to be pro-inflammatory, many Nfkb and CEBPb targets. I tried several tools, MEMECHIP (and other MEME tools), Homer, RSAT and I don't see any Nfkb or CEBPB enrichment.

Then I tried a set of peaks obtained with CEBPb ChIP-seq and again i did not obtain any enrichment for CEBPB. I also tried a short (< 100) gene list known to be canonical Nfkb targets and again, failed to get any RELA related enrichment. Many SP1-3 come up so then I tried enhancer peaks (defined separately) and again failed to get any inflammation-related motif enrichment.

I also tried different representations of promoters/enhancers e.g. TSS +- 500, 1500, 2500 or just nucleosome-free regions in the promoter/enhancers. Please offer any advice on how you do motif enrichment analysis.

Thank you in advance!

[ATpoint]

Please share some command lines. With only words it is diffcult to see what exactly you did so far.

Answer 1

[manuel.belmadani]

It's possible your motifs aren't enriched enough for the tools to find them in the input data you're looking for. Many of those were designed to work on ChIP-Seq datasets which would gives you at least a few thousand sequences of about 100bp long (sometimes more, sometimes less) from various locations in the genome. Does your input look anything like that? If not, it might not be possible for those tools to detect them, say if your search sequences are much longer and you have fewer input sequences. Those tool work well because the sequence is short, and there's many examples to mine.

If you don't have something like ChIP-Seq for your targets of interest, I would look across the genome using the motifs you're interested in. Homer has a tool to do that: http://homer.ucsd.edu/homer/motif/genomeWideMotifScan.html

Then, see where your hits map, and see if they appear where you would expect (with respect to the genes list you're interested in.) See here on how to create your own motifs for Homer: http://homer.ucsd.edu/homer/motif/creatingCustomMotifs.html You could probably get the data you need from JASPAR or CIS-BP. Keep in mind though that it's not because you have a match for a motif somewhere in the genome that it's a functional, or even interesting site. You would have to look at more data to identify which match might be interesting (ATAC-Seq, Conservation, maybe something like BHEST or GREAT).