--OT and --OB suggestions by mbias produce "Invalid bounds string"

[DaGaMs]

I ran MethylDackel mbias on a WGBS file (mapped with bwameth.py) and got:

>$ MethylDackel mbias -@ 4 -q 5 -r chr7 hg38.fa GSM4505863_md.bam GSM4505863_mbias_chr7
Suggested inclusion options: --OT 2,0,0,115 --OB 0,124,11,0

Now I just copy-pasted this to MethylDackel extract, but I get a warning Invalid bounds string, 2,0,0,115:

>$ MethylDackel extract --OT 2,0,0,115 --OB 0,124,11,0 -@ 4 --mergeContext -q 5 -o GSM4505863 hg38.fa GSM4505863_md.bam                                                                                                     
Invalid bounds string, 2,0,0,115
Invalid bounds string, 0,124,11,0

I'm guessing this is because the handling of the 0 doesn't work in version 0.5.1, from all I can tell. changing the 0's to 1/150 "fixed" it. The same problem affects -nOT and -nOB, which also can't handle the 0, which for those arguments is particularly problematic. There must be some bug in the parser logic somewhere...

[dpryan79]

That's weird, thanks for reporting this. I'll try to track down this bug.

[dpryan79]

@DaGaMs Can you try this with version 0.5.2? I haven't changed any of the parsing for this in a number of releases, so I wonder if this is due to having linked against the wrong htslib version in 0.5.1.

[DaGaMs]

I recompiled 0.5.2 with gcc-8.3.0 against htslib-1.11.0. I still get the same error:

MethylDackel extract -@ 2 -q 5 -l hg38_main.bed.gz --cytosine_report --OT 8,124,2,0 --OB 2,120,0,104 -o merged_701503  ref.fa ../Alignments/merged_701503_md.bam

Invalid bounds string, 8,124,2,0
Invalid bounds string, 2,120,0,104
[...]

[DaGaMs]

I just tried with 0.6.0 and I still get Invalid bounds string with any pattern that contains a 0

[jganbat-fn]

@DaGaMs and @dpryan79
So the error comes from errno != 0 somehow errno was not 0 when it first parses the argument. Maybe before the star of the program it should be initialized to 0, however doing some quick research says that is not a very prudent thing to do 😅

[voudiou]

Hi, any news about this ?
I just got the same problem. Without 0 it runs fine, but with 0, I got the same Invalid bounds string error.
The version is 0.6.1, and HTSlib 1.16
Thanks !

[jganbat-fn]

Hi, any news about this ? I just got the same problem. Without 0 it runs fine, but with 0, I got the same Invalid bounds string error. The version is 0.6.1, and HTSlib 1.16 Thanks !

I just re-initialized the errno variable at the beginning of parameter parsing and compiled it, i think that is safe since it is at the beginning.