search

BCCDC-PHL / auto-cpo

Automation of Genomic Analyses for Carbapenemase-Producing Organisms (CPOs)
GNU General Public License v3.0
0 stars 0 forks source link

Add QC checks #9

Closed dfornika closed 1 year ago

dfornika commented 1 year ago

The overall analysis may fail when an analysis includes a low-coverage sample. This typically occurs in the BCCDC-PHL/routine-assembly pipeline. We should identify samples that are likely to fail in the routine-assembly pipeline due to low coverage and exclude them from that analysis.

dfornika commented 1 year ago

The BCCDC-PHL/routine-assembly pipelne has been updated to be more robust against low-coverage samples crashing that pipeline. The BCCDC-PHL/mlst-nf pipeline should get similar updates.

But we should also introduce some quality checks in this automation tool that will filter out low-quality samples that are unlikely to produce meaningful results from downstream analyses. We can encode specific QC filtering logic, with configurable thresholds.

We should also produce a QC summary with PASS/WARN/FAIL for each check.