RSAT gives strange motif calls

Question

[rbronste]

Hi, taking some differential ChIP-seq peaks and searching for motifs via RSAT default settings and getting some strange results. For instance as Jaspar Core Vertebrates match to the first de novo motif found in my data it gives 4 different transcription factors as matching, from 4 different families (e.g. Mef2c, Foxd3 etc.) so not sure why its not giving other members of the same family since Im assuming they would be closer. In addition across many unrelated differential peak sets I scan I get the same top motif hits which sort of appear like repeats at least on the face of it. Im wounding what may be going on here, and if I should be repeat making my ChIP-seq differential peaks before running motif analysis etc? Thanks!

Answer 1

[Jaime Castro-Mondragon]

Hi

Which program are you using, I guess peak-motifs ? I am part of the developers team, if you provide the link with your results we will be happy to help you in the interpretation or fix bugs in the programs.

In the results file (html) the motif comparison shows the top 3 matches, and a link with the remaining similar motifs, as you already pointed, is almost sure that other similar motifs will be TFs for the same family.

You may check the distribution of TFBSs of the discovered motifs for each motif (plot), some of these repetitive motifs were discovered by position-analysis, which is very good program to discover over-represented motifs around the ChIP-seq summits, but it can also detect motifs that are under-represented around the summit. We already noted this behavior.