Closed ghuls closed 3 months ago
Hi, thanks for the feedback and checking the input files. The input files should be fine and the issue seems a htslib (bgzf_flush) error. After google search, I found three related issues on Github indicating that the error could be caused by insufficient disk space, given the large size of the input BAM file. You may see the three issues for detailed discussion: issue-1, issue-2, and issue-3.
You may check your disk and try increasing the quota. Alternatively, you can reduce the number of threads (-p
option) in cellsnp-lite to save memory in case there is a out-of-memory problem, as this Biostars page indicates.
It was indeed a quota problem and I checked the quota before reporting this error. Another user filled up the quota and deleted some files when they saw it, so when I checked the quota, there was space again (more than 1 TB) , so I didn't expect disk space issues.
Input files seem to be OK:
Are the
bgzf_flush
errors from writing partialcellSNP.base.vcf.gz.0
,cellSNP.base.vcf.gz.1
files?So far the same command worked on other BAM files. Only
PB-002-pool_5.sorted.bam
andPB-002-pool_7.sorted.bam
gave those errors (PB-002-pool_9.sorted.bam
andPB-002-pool_10.sorted.bam
are still running). By chance (or not), those are also the biggest files. Any idea what could be going on? Like an integer overflow.