Closed anoronh4 closed 2 years ago
we need to hard-fail a sample even before it is included in a cohort, when it is a tumor sample and meets the tumor coverage fail condition. therefore, we will need to pass tumor/normal info to the SampleRunMultiQC in order to correctly assess the sample.
Alternatively, we can repeat coverage statistics in SomaticRunMultiQC
and use tumor/normal coverage stats there.
So our decision is to repeat coverage statistics in SomaticRunMultiQC
and use tumor/normal coverage stats there. It's done by #923
The reason is we want to keep our ability to generate sample level QC report without pairing file or tumor/normal information.
QC results will be display in PM and investigator trackers, so it will be necessary to revise any QC summaries to make sure they are ready for those views, and also to make the statuses easily retrievable (simple table instead of embedded in the html file) for other systems.
We might also consider to update QC summary statuses to a centralized location such as a flat text file or database file (such as an sqlite3 file). This file could contain the sample id, QC status (pass/warn/fail), QC reason (empty unless warn or fail) and the date of the last update.