Comparative ChIP-Seq

[RockChalkJayhawk]

Hey all,

Does anyone have a script that I can use to correlate 2 ChIP-Seq experiments?

I want to compare the correlations before and after I write a program to normalize the data to total tag counts.

For example,

In File1.txt I have:
chr1 1000 2000 23
chr1 2000 3000 15

In File2.txt I have:
chr1 1000 2000 10
chr2 5000 6000 2

where the first column is the chromID, col2 is the start, col3 is the end, and col4 is the number of hits in that bin.

I want to build an XY scatterplot between the two samples, but I have to match both the chromID and the start site for the X and Y coordinates. Some scores are in File1.txt but not File2.txt and vice versa.

Ideally the file would look like:
chr1 1000 2000 23 10
chr1 2000 3000 15 0
chr2 5000 6000 0 2 Where col4 is the count from File1.txt and col 5 is the count from File2.txt

Any ideas???

[apfejes]

In a way, I have this in FindPeaks 4.0.x - it's just not as simple as you've laid it out here.

I don't do a "binned" approach either - and the file formats require you to have access to the raw data - not just peak calls. it does create png scatterplots, though.

If that's not too far from what you're looking for, i can point you in the right direction.

[RockChalkJayhawk]

In a way, I have this in FindPeaks 4.0.x - it's just not as simple as you've laid it out here.

I don't do a "binned" approach either - and the file formats require you to have access to the raw data - not just peak calls. it does create png scatterplots, though.

If that's not too far from what you're looking for, i can point you in the right direction.

Thanks Anthony. I'll take a look.

[SoftGenetics]

NextGENe has a comparator for what you wish to do

[simonandrews]

SeqMonk will allow you to do this kind of thing (but again, working from the raw data not your binned summary). You can create a series of bins, quantitate within each one (even adjusting for different total read counts) and then either export a matched dataset, or filter out bins of interest based on their quantitated counts. You can also plot out the distribution of counts across all bins or compare distributions with a BoxWhisker plot.