Quantifying ChIPseq signal for one sample

[kajot]

Hey Everybody,

I am trying to rank genes according to their phosphorylated CTD Pol II occupancy as a rough estimate of ongoing transcription.

I am using ENCODE Pol2 ChIPseq data, which means I have access to called peaks, raw and wiggle files. I would like to ask you if it's generally a good strategy to quantify signal from bigWig file using bigWigAverageOverBed UCSC tool for all the human genes in given cell line. I avoided using called narrow peaks in this case, as I believe this would approximate the polymerase stalling better than the actual transcription.

To assess this approach, I am correlating this quantified signal to RNAseq RPKM values from the same cell line and get Spearman correlation coefficient of around 0.62.

Is there a better way to rank genes by their ChIPseq signal in a single sample?


Thanks for any suggestions,
K.

[dpryan]

Something along those lines should work. We normally use computeMatrix from deepTools for this sort of purpose, but if you only want a single value then bigWigAverageOverBed should give equivalent results.

[kajot]

As always very helpful, Devon, I owe you few beers already

Thanks for the tip!