DKFZ-ODCF/AlignmentAndQCWorkflows

BAM integrity check job

Opened this issue · 0 comments

If two jobs are (by mistake or bug in an automation system or batch processing system -- our scenario) run during the same time their files get mixed and may be corrupt. For the statistics files this might by acceptible (they can be recalculated), but not for the BAM (e.g. if the FASTQs are deleted after some time -- which is our use case). The costs of such mistakes are more expensive than the costs of implementing and maintaining this feature.

Implement a simple job that checks the integrity of the final merged BAM file, e.g. by samtools flagstat $bam > /dev/null.