[BioC] MEDIPS error extending reads

[Stephen Turner]

Thanks Lukas. That solved that problem. I now have another problem.
I'm trying to look at regions of interest as before. After I've called
MEDIPS.meth(), I then try to run:

dmr_rois <- MEDIPS.selectROIs(results=dmr_all, rois=roi_file,
columns=c("counts"), summarize=TRUE)

where roi_file is created using read.csv() on this file of ROIs:
https://gist.github.com/stephenturner/5346095

I'm now getting the error:

Error in data.frame(..., check.names = FALSE) :
  arguments imply differing number of rows: 2602, 0

I'm guessing this has something to do with my ROI file, but I'm having
trouble figuring out what to do. I have 2611 ROIs, not 2602. I've
double checked that I have no duplicates - these both return 2611:

length(unique(roi_file$ID))
length(unique(with(roi_file, paste(chr, start, stop))))

There may be some overlapping regions, but I'm unsure how to best
handle this. Could this be causing the error?

Thanks,

Stephen

On Mon, Apr 8, 2013 at 3:58 PM, Lukas Chavez
<lukas.chavez.mailings at googlemail.com> wrote:
>
> Hi Stephen,
>
> I was able to reproduce the error. It appears that you have unmapped reads
> in your bam file which we always filter out during conversion of sam to bam
> files using samtools view -S -F 4 -b -o out.bam in.sam (single end sequening
> data). In case you filter your bam files accordingly, the error message
> should disappear.  However, I will add an additional flag
> scanBamFlag(isUnmappedQuery = F) to the according scanBam() function in dev.
> Therefore, unmapped reads in a given bam file should be ignored in the
> future (MEDIPS version >=1.11.1).
>
> Lukas