Hello,

I am using DiffBind for my MeRIPSeq (mRNA methylation) analysis.

I have been using it to extract a consensus list of peaks. I am running:

> diffConsensus = dba.peakset(diff,consensus = DBA_CONDITION,minOverlap = 0.5)

> diffConsensus = dba(diffConsensus,diffConsensus$masks$Consensus)

> write.table(consensusTable,"DiffConsensus.txt",sep="\t",row.names = FALSE)

And i obtain this table:

What are the numbers in column 5,6 and 7 ?
I have gone through the vignettes and tutorials and I cannot find an explanation about the number calculated during the operation.

Thanks

Hello-

The values in the final three columns are normalized peak scores (occupancy scores). When a peakset is first loaded, all of the scores (taken from the default or specified column in the peakset file) are normalized to a 0...1 scale. This score is used when a binding matrix is constructed. If the given peakset does not include any peaks that overlap with an interval in the binding matrix, it receives a score of -1.

Whenever peaks are merged (for example, when you called dba.peakset() with consensus=DBA_CONDITION), if there are multiple overlapping peaks, the highest score will be retained.

You are correct that the documentation does not explain this clearly. There are hints in the man page for dba.peakset() and the Loading Peaksets section of the vignette, but I'll look at adding a more clear explanation in one place.

Cheers-

Rory