Tweaking Findpeaks Output

Question

[darcangelo.elisa]

Hey everyone,

I am analysing Chip-seq data with MACS and FindPeaksb (I am dealing with a transcription factor). Being new to the world of Chip-seq, I am puzzled to see how much of a difference in term of Peak numbers there is between the two. Specifically, my question is the following: While with MACS there are fewer peaks in my 24hr sample compared to my 0hr sample (as expected!!), with FindPeaks the opposite is the case. This makes me think that maybe I am not removing the background, but even when adding the control flag and selecting only for differentially enriched peaks, the problem persists. So I was wondering what flags do people commonly use for FindPeaks when looking at transcription factors? Or, more in general, how do people define a peak? Any suggestions will be greatly appreciated!!

Thank you!

Regards,

Elisa

[Steve Lianoglou]

Not directly answering your question but perhaps you might try a third method (I've heard good things about SPP) and see if you can identify the outlier?

Answer 1

[apfejes]

Hey - just a quick reply. It's absolutely impossible to know what you've done with find peaks from your description. In general, findpeaks shouldn't give you excessive peaks... but I'm not sure whats up. Perhaps you can describe what parameters you've done, or explain at least what the peaks look like that are appearing, or even what the control is...

Good luck otherwise! Anthony Fejes