Hi,

I have run Diffbind on a set of ChIP-seq samples and generated a dba.count object. Now, I want to count reads on a set of peaks (promoters/enhancers) that I pre-defined and got this error:

H3K27ac.dba<- dba(sampleSheet="my_full_diffbind.csv", scoreCol= 7, filter=80, peakFormat = "macs")

## using RPKM for normalization (reads number in the bam files)
H3K27ac_RPKM<- dba.count(H3K27ac.dba, minOverlap=2, 

                      fragmentSize = 200, bParallel = T,
                      score = DBA_SCORE_RPKM)

mypeaks<- import("my_enhancers.bed", format="BED")

mypeaks.count <- dba.count(H3K27ac_RPKM, peaks= mypeaks, bParallel = T,

+                                      fragmentSize = 200, score = DBA_SCORE_RPKM, minOverlap = 0)

Error in `$<-.data.frame`(`*tmp*`, "chrmap", value = c("chr1", "chr10",  : 

  replacement has 24 rows, data has 45728

Thanks very much.

Ming

Hi Ming-

There appears to be a bug here. We'll have to wait a short while to fix it in the new release as the current release is closed to new fixes.

I do have a workaround for you however: if you set bUseSummarizeOverlaps=TRUE when calling dba.count(), this should work (although it won't support the fragmentSize=200 parameter).

I am a little curious about what you are trying to do, specifically why you are calling dba.count() two times, and passing in the result of the first call to the second call. If you want to count based on your predefined enhancers, you can pass them into the first call to dba.count(). Also, whenever you pass in pre-defined peaks, the minOverlap parameter is ignored as it only applies when a new set of consensus peaks are being derived.

Apologies for the inconvenience, I'll post here when we are able to check in a fix.

-Rory