sambama depth base in v0.5.8 doesn't work with a regions file

[roryk]

v0.5.8:

./sambamba depth base -L small_MIG.bed align/NA12878-1/NA12878-1-sort.bam | head -3
Processing reference #1 (1)
REF POS COV A   C   G   T   DEL REFSKIP SAMPLE

v0.5.6

sambamba depth base -L small_MIG.bed align/NA12878-1/NA12878-1-sort.bam | head -3
Processing reference #1 (1)
REF POS COV A   C   G   T   DEL REFSKIP SAMPLE
1   69296   1   0   1   0   0   0   0   NA12878-1
1   69297   1   1   0   0   0   0   0   NA12878-1

[lomereiter]

Hi Rory,
Please check the base qualities in a BAM viewer. The bases used to be counted incorrectly, namely, all bases in a region would be counted for every read that has at least one good base in the region (see 15f75bb)

[roryk]

Hi Artem,

Hmm-- I don't think that is it, I get 0 counts even if I run a ton of regions. I dissected out one region that looks like it should be generating counts:

v0.5.8

sambamba depth base -L test.bed test.bam | wc -l
1

v0.5.6

sambamba depth base -L test.bed test.bam | wc -l
149

http://dl.dropbox.com/u/2822886/Screenshots/e6lk6pbwr4gy.png is a screenshot of the alignments in the region 1:979489-979637.

I put up the test files here:

https://dl.dropboxusercontent.com/u/2822886/sambamba-depth-test-files.tar

[lomereiter]

Ok, I'll look into this.

[roryk]

Thanks! You've been great about responding to issues, I really appreciate it.

[lomereiter]

This bug was introduced by my awful fix of another bug (6d1f5a3), BED parser should be completely rewritten once I find time for that.

[roryk]

haha-- I know that pain. Thanks for the quick fix Artem.