Trouble with bedtools groupby - not summarizing by columns

Question

[yarmda]

I'm trying to use bedtools intersect and group by to get a count of reads by region they overlap. This should be straightforward, but the output I'm getting is just a number (presumably, the count of all reads that overlapped any region).

Process:

bamToBed mybam.bam > mybam.bed
sort -k1,1 -k2,2n -k3,3n mybam.bed > sort.mybam.bed
bedtools intersect -a sort.mybam.bed -b regions.bed -wa -wb | groupBy -g 1-3 -c 8 -o count > groups.bed

As far as I can tell from the documentation, that should do just what I need. I receive no errors when running this, but no columns print in the output. I understand that the input files must be sorted on the columns to be output, but I think I've done that.

Any thoughts?

[Pierre Lindenbaum]

I'm trying to use bedtools intersect and group by to get a count of reads by region they overlap

use GATK DepthOfCoverage

Answer 1

[dariober]

What version of bedtools are you using? In v2.26.0 there are some issues with groupby and the one you describe may be similar to #435. If this is the case try downgrading to 2.25.

By the way, for what you need you could use intersectbed with -c option: For each entry in A, report the number of hits in B. I.e. like: intersectBed -c -a regions.bed -b sort.mybam.bed. Also consider using the -sorted option for speed.

[yarmda]

Thanks! I didn't even see the -c option for intersectBed. And I do have v2.26 installed, which may have been the issue.

Answer 2

[chycheng2012]

I had same problem. By using v2.25, I got expected result

Answer 3

[Eijy Nagai]

Yes, I got exactly the same problem today. Seems that Bedtools version v.2.26.0 still have the problem with groupby function. I tested with v.2.25 and worked.