ding-lab/msisensor

msi fails with exit code 139 when encountering contigs not in bam/bai

genuinekale opened this issue · 5 comments

genuinekale commented

When running msi using a microsatellite list generated by MSIsensor scan on UCSC hg19, msi fails (Segmentation fault (core dumped)) when hitting contigs such as chrM, not present in the bam/bai used as input. If these are manually removed from the microsatellite list, the issue is resolved.

stevekm commented

I am having the same issue. Running the command

msi -d hg19_microsatellites.list -n HapMap-B17-1267.dd.ra.rc.bam -t SeraCare-1to1-Positive.dd.ra.rc.bam -e targets.bed -o msisensor -l 1 -q 1 -b 1

on the demo dataset here:

I get the same error:

loading bed regions ...
loading homopolymer and microsatellite sites ...
.command.sh: line 6: 27044 Segmentation fault      (core dumped) msisensor msi -d hg19_microsatellites.list -n HapMap-B17-1267.dd.ra.rc.bam -t SeraCare-1to1-Positive.dd.ra.rc.bam -e targets.bed -o msisensor -l 1 -q 1 -b ${NSLOTS:-1}

...

Command exit status:
  139

Is there a solution to this?

stevekm commented

Also I tried subsetting the microsatellite list for only the chromosomes that are present in the .bam files, but the same error still occurs.

update: after more testing, I figured out that the full UCSC hg19 reference microsatellite list was fine, the issue seems to have been caused instead by the -e regions .bed file, which contains some regions for chromosomes which were not covered in the .bam files.

update2: after even more testing, the files started returning this error again, it seems this was not the solution after all?

Beifang commented

bug fixed in missensor v0.3.

brucemoran commented

I hit this with v0.5, Singularity container here.

Tested and found gzipped exome.bed was the issue for future reference.

guodudou commented

I also have the same problem with v0.5, can you please specify how to fix? Thanks!