How to find motif among overlapping binding regions of a TF activated by two different ligands

[Sid Das]

Hi,

I am trying to find the motif among common regions (from ChIP sequencing data) bound by estrogen receptor after activation by two different ligands that seem to activate common and different biological mechanisms using RNA seq. So I was planning to motif analysis to find common and different motifs using the common and unique regions bound by them. I was using the MEME suite but it seems like for ChIP data, the ideal thing is to use around 500 bp around peak summit using MEME-ChIP tool.

 I was wondering what would be the right approach to motif finding under these circumstances?

[cegrant]

MEME-ChIP is a good tool when  you expect motifs to occur in the central region of the input sequences, as is typical for ChIP-Seq data. If you are using other protocols that might not result in centrally enriched motifs you might want to use XStreme instead. XStreme is similar to MEME-ChIP, but uses SEA to look for motif enrichment rather than Centrimo. SEA does not assume that motifs are centrally located in the input sequences.

Using 500bp sequences sound ideal. The main thing to avoid is having 1000 sequences between 100 and 500bp in length, and 10 sequences that are 100-500 kb in length. The long sequences will skew the statistics so that assessments of statistical significance will be unreliable.