bam/bigwig compare ChIP/input

[Elena Gómez-Díaz]

Hi all,

I have some concerns related to the substraction of the input from the ChIP signal. Whether you calculate the difference or the log2 ratio, my issue come in using this normalized file for downstream analysis for example for comparing between samples or conditions. So the question is should I convert negative values (i.e. means that the signal in the input was greater than the ChIP) in zeros?

thank you

elena

[Friederike Dündar]

Hi Elena,

sorry for the delayed response, but things are a bit busy here and there's no simple answer to your question.

The main question is: why do you feel that you should set the background to zero? Do you feel that it is interfering with your downstream analyses?

When I did the comparisons and clusterings for our eLife paper (http://dx.doi.org/10.7554/eLife.02024.006), I felt that the clustering was sometimes "mislead" or confounded by the variation in background signals, and since I wanted to assess the similarity between the enrichments, I indeed set all values below 0 in the matrices obtained from computeMatrix to 0 and subsequently rank-transformed each matrix. This way I eliminated the variation in the background, but I also was able to compare signals between different samples that showed very different dynamic ranges.

That being said, I also think that it may be worthwhile to investigate the reasons for the large variability in background signals (that I'm assuming you have). Perhaps the normalization scaling factor is not optimal or perhaps your ChIP did not work as perfect as expected. Did you use bamFingerprint to assess the nature of your ChIP signal? And if you do a heatmap for randomly chosen regions for which you wouldn't expect a signal, do you see a more or less flat line around 0? If that's not the case, you should play around with the bamCompare parameters (esp. the scaling) until that's the case.

Not sure if that helps at all, just keep on asking if things are unclear :)

Best,

Friederike

--
You received this message because you are subscribed to the Google Groups "deepTools" group.
To unsubscribe from this group and stop receiving emails from it, send an email to deeptools+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.