Chip-Seq analysis of biological duplicates

Question

[avinashsrivastava5]

I have done ChIP-seq for a transcription factor in two biological duplicates. Peaks were called using GEM on both the replicates. the first replicate gave around 48, 000 peaks while the second replicate gave 1, 80, 000 peaks. FRIP for these replicate was 5% and 43% respectively. Interestingly, 93% of peaks in the first replicate were common with the second replicate. Right now, I am planning to take the intersection of the two peak set for further analysis. I am not sure if it will create a problem while publishing. Please suggest if I need to do more any more statistical analysis before going for further analysis.

P.S. MEME analysis of intersect peaks shows enrichment of 2 motifs. Thanks

Answer 1

[GouthamAtla]

Since you have only two biological replicates, why don't you run IDR analysis and take the consistently reproducible replicates across replicates which was recommended by Encode ?

For this you need to call peaks using macs2 ( or any other program ) using a very low stringency like p-value of 0.1 ( not q value ), and then feed those peaks to IDR analysis to get the reproducible peaks.

[avinashsrivastava5]

Hii, Thanks for comments ! I have done the IDR analysis. 18,000 of our 45,000 common peaks could pass the IDR of 1%.

Is it OK. ?