Hello,

I've got a set of binding site intervals in a DBA object and use dba.count() to count the reads in those peaks. According to the manual, the default score=DBA_SCORE_TMM_MINUS_FULL. But it also mentions "Note that all raw read counts are maintained for use by dba.analyze".

So, the question is, if I then use dba.peakset(bRetrieve=T, DBA_DATA_FRAME) to get a dataframe, will that dataframe contain the TMM_MINUS_FULL scores or the raw read counts?

Thanks,

James

Hi James-

When you call dba.peakset() with bRetrieve=TRUE, you will get back the binding matrix using the score set in dba.count(). So in this case, you will get the TMM normalized scores (with control reads subtracted, and using the full library sizes instead of just the reads in peaks).

However you can quickly and easily change the read scores using dba.count() with peaks=NULL. So if you have already done the counting using the default TMM score, you can get the raw unaltered read counts as follows:

> myDBA <- dba.count(myDBA, peaks=NULL, score=DBA_SCORE_READS)
> rawreads <- dba.peakset(myDBA, bRetrieve=TRUE, DataType=DBA_DATA_FRAME)

Likewise, to get RPKM scores:

> myDBA <- dba.count(myDBA, peaks=NULL, score=DBA_SCORE_RPKM)
> rpkm <- dba.peakset(myDBA, bRetrieve=TRUE, DataType=DBA_DATA_FRAME)

Cheers-

Rory