Hello Biostars,
Can anyone tell me how to give narrow & broad peak parameters in 'macs2 bdgdiff' program, as we can mention in macs2 callpeak with '--call-summits' for narrow & '--broad' for broad peaks.?
Thanks in advance. :)
There are no parameters like that for bdgdiff, which just looks at the raw bedgraph pileups to determine differentially bound regions. Personally, I've found it doesn't do all that great a job of it either.
If you want to see which peaks called have significantly different signal between two files, I'd recommend MAnorm which will take two peak files and two reads files, derive a consensus peak set, and tell you which of the peaks are significantly different between the two samples. Using this method, you could feed in the summits or broad peaks as your peak files.
I'll be a bit more technical here.
bdgdiff first generate three tracks from the 4 input tracks, which contains the log-likelihoods of t1 > t2, t1=t2, and t1 < t2. The three output files are in fact peaks called by bdgpeakcall--the subfunction for calling narrow peaks--from these three tracks. So, bdgdiff is not suitable for calling differential broad peaks.
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
Yes, MAnorm sounds better. The MAnorm's manual says that it gives M & A values too. Can I say A is somewhat equivalent to fold change in macs2?
The M value is more akin to fold change, though it's a better idea to consider it as a measure of the magnitude of the change between the two samples. Significance is great, but a significant yet small change between the two samples may be biologically irrelevant - using the magnitude of the difference helps eliminate false positives seen using the p-value alone. You can use the p-value output and the M value (positive for peaks enriched in sample 1, negative for peaks enriched in sample 2) to define your different lists of differentially bound peaks.