long-read-pipelines
long-read-pipelines copied to clipboard
Have a depth check for input/output bam
See discussion in #111
Changed ticket to reflect a more general scenario: that when a BAM is sharded, the output/input might now be informative enough. By informative, we mean that either the bam size is suspiciously small, or coverage is unlikely to be suited for downstream analysis purposes. In the first scenario, the sharding code should check that. In the latter, which is application-dependent, it's up to the workflow/task themselves to check up front, and ultimately may only be possible after we get more experience and collect more metrics.