Intro-to-ChIPseq

Intro-to-ChIPseq copied to clipboard

It seems like the score they ask us to use when filtering reproducible peaks does not work out when I tried manually computing it. Found this with a current consult and have posted an issue:

https://github.com/nboley/idr/issues/43

keep it as note, bring bedtools back in check with shannan about the plot she generates

Hi @mistrm82 ,

I have been using MACS2 to call chipseq peaks with a lenient pvalue 0.1 on true replicates, then using IDR as instructed in the HBC tutorial. However, when I run awk '{if($5 >= 540) print $0}' sample_IDR_results | wc -l There are 0 peaks. But based on column 11 (global IDR value?) there are 90k peaks with column 11 <0.05. I'm confused which value I should use to get significant peaks. The sample_IDR_results file has ~100k peaks. Do you have any advice on this? Thank you!

Hi @laijen000 , I am not sure if perhaps IDR has changed the format of the output since we last updated our materials. The idr docs state that column 5 is the score that we should threshold on: https://github.com/nboley/idr#output

But maybe contacting them, directly to see if this is no longer the case?