Differential binding analysis

Question

[marina-orlova]

Hello

Could you please tell me if there is any way to make differential binding analysis with chip-chip and chip-seq data (one sample is from chip-chip experiment (.cel files and gff3 file with peaks) and the other is .bam from ChiP-seq (.bam files and gff3 with peaks)).

[dariober]

I have no experience with chip-on-chip, but comparing two sample (and two conditions?) with one technique is already difficult to interpret as you cannot know to what extent the differences you see are due to the conditions or to the within condition variability. Comparing two samples from two techniques is even worse...!

Answer 1

[seidel]

There's no robust way that I'm aware of. As dariober hinted at, there are too many caveats to consider when comparing peaks across two very different technologies. However, depending on your ultimate objective, you could perform an ad hoc analysis that could tell you a few things. You could use bedtools or the IRanges capabilities of R to tell you how well the peaks overlap between the two experiments - that's the most basic level of differential analysis. You could then use those same tools to define the common regions called as peaks between both platforms, and compare the distributions of scores (i.e. each might be a ratio over background) under the common areas to see how comparable they are. To quantify differences you might have to normalize them to each other, or transform them to a similar scale (i.e. z-score, etc.), or examine them by rank - a lot depends on what the data looks like, and what you're really trying to get out of it.

[marina-orlova]

Thank you a lot for your answer, could you please advise how I can normalize chip to input ratio to compare them to each other (please advise what to do in general terms)? The first image is peaks of ChiP-seq and the other is peaks of ChiP-ChiP

Answer 2

[vivekbhr]

To compare two conditions in ChIP-Seq, you can use the DiffBind package in Bioconductor ..