closest command missing results

Using v2.17.0. Have confirmed the issue on Ubuntu Linux and Windows 7 (compiled 
using cygwin)

Command I executed:

bedtools closest -a test.bed -b test2.bed > closest_actual.bed

I've attached the 2 input files, the output (closest_actual.bed), as well as 
the output I expected and received from previous BEDTools version 
(unfortunately I'm not sure exactly which version), filename 
closest_expected.bed.

There is one line difference between the actual and expected:
>chr1   200 201 chr1    150 201

This feature is not reported in v2.17.0, which I believe to be incorrect 
behavior. There are two features in test2.bed, chr1:175-375 and chr1:150-201, 
which overlap chr1:200-201. According to the doc, since they both completely 
cover the feature in file A, they should be tied and both reported. I tried 
using each of the -t options (all, first, last), the result was the same.

This may be a duplicate of issue #144.

Original issue reported on code.google.com by ja...@broadinstitute.org on 24 Dec 2012 at 6:22

Attachments:

Genomicus / bedtools

closest command missing results #145