Question: Homer motif analysis: "Very close % of target vs % of Background sequences with Motif" & "SeqBias at 1st rank"
0
gravatar for chiefcat
19 months ago by
chiefcat90
chiefcat90 wrote:

Hi everyone. I'm trying to find the motifs bound by my ChIP-ed protein using Homer. I don't have knowledge regarding NGS analysis. The analysis was done by my labmate who understand programming languages but lack of experience in analysing ChIP-seq.

When going through the result list, I found that there are 2 obvious issues happened.

1) The % of target vs % of Background sequences with Motif are very close throughout the list. When this happened, does it mean that the motifs found are actually not/ very lowly enriched in my ChIP-ed sample?

2) The "SeqBias: CG bias" appears as the 1st ranked motif in my list. I know this indicates that the analysis parameter is not appropriate as stated in http://homer.ucsd.edu/homer/motif/practicalTips.html. This is probably related to the normalization to either by total GC-content or CpG-content if I understand the tutorial correctly. Should I just adding "-gc" to the motif finding command to solve this problem? If having this problem on my list, do other results on the list still valid?

Thanks very much!

enter image description here

motif chip-seq homer • 1.3k views
ADD COMMENTlink modified 19 months ago by ildem40 • written 19 months ago by chiefcat90
0
gravatar for ildem
19 months ago by
ildem40
IGH - Montpellier
ildem40 wrote:

Hi, I usually get a decent enrichment over background, sometimes upto 3 fold, but at least a 1.5 fold would be expected. Can you give us your command line? What does your bedfile (input) look line? Could it be that you are searching for a motif in too large of an area? Usually you search for the motif within +/-100 bp of the peak summit, and +/-250bp for the "co-binding" Tfs. If you search for a motif in large areas (i.e. entire peak) you might end up with high background.

Also, Have you removed the "blacklisted" regions from your Chip-seq peaks to get rid of the background regions that are called peaks by every single ChIP experiment (type blacklisted regions and your genome in google).

As Homer says, I would try to match the background to your Chipseq peaks. Are they mostly promoters? then generate a background file for promoters etc.. by default homer is using randomly picked regions as background.

ADD COMMENTlink written 19 months ago by ildem40
Please log in to add an answer.

Help
Access

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
Powered by Biostar version 2.3.0
Traffic: 1873 users visited in the last hour