MultiQC silently skips large samtools stats files due to log_filesize_limit default (50MB)

[Leilanasd]

Hi everyone,

I noticed that when running taxprofiler with "--perform_longread_hostremoval",
some samples are missing from the Samtools Stats section in the MultiQC report.
There is no error or warning they just do not appear.
I found the cause: MultiQC has a default "log_filesize_limit"
of 50MB. Samtools stats files from high-coverage Nanopore samples can exceed this, in my case one sample produced a 68MB stats file which was silently skipped.
You can see it in the MultiQC debug log: skipped_module_specific_max_filesize: 1
And confirmed by checking the MultiQC source directly:

import multiqc.config as c
print(c.log_filesize_limit)  # 50000000

To fix:
add this to "assets/multiqc_config.yml":

log_filesize_limit: 500000000

I tested this with 5 samples (including the problematic one) and all appear
correctly in the report after the fix.

This is likely to affect other nf-core pipelines using samtools
stats with high-coverage data too.

Happy to open a PR with the fix if helpful!

[sofstam]

Hi! @LilyAnderssonLee and I have noticed this before for the nanopore data and you are welcome to open a PR with the fix