ga4gh/quality-control-wgs metrics_definitions.md - Mean insert size

[justinjj24]

Description: The mean insert size of short paired-end sequencing high quality, non duplicated, properly paired reads, primary alignments, achieving a mapping quality of 0 or greater on GRCh38 assembly.

Implementation details: In the NPM-sample-QC reference implementation it is computed using GATK Picard’s CollectInsertSizeMetrics, reporting the MEAN_INSERT_SIZE field. Only the non duplicated, properly paired reads mapped on GRCh38 assembly with MAPQ > 0 are considered.

Good to include the comment given below or remove the high quality from the description? The high quality reads also not part of the description in the proposed GHIF WGS QC document!

• NPM-sample-qc implementation metric Mean insert size doesn't select high quality reads using picard-collectinsertsizemetrics.
• It is critcal that the (BAM/CRAM) alignment files generated only from the illumina PF reads.

[justinjj24]

state 'No minimum mapping-quality is imposed' or 'No filter on mapping qualiy' is applied, instead of MAPQ > 0 are considered?

A mapping quality of zero in bwa means that the read maps to multiple locations with the same quality and that the mapper has picked one of these positions at random.

This process uses non duplicate reads only.

[mhebrard]

Fix typo, clarify MQ criteria & exclude duplicate

• Description: The mean insert size of short paired-end sequencing high quality, non duplicated reads, primary alignments, mapped on GRCh38 assembly. No minimum mapping quality is imposed. It is critical that the (BAM/CRAM) alignment files be readily marked for duplicated reads
• Implementation details: In the NPM-sample-QC reference implementation it is computed using GATK Picard’s CollectInsertSizeMetrics, reporting the MEAN_INSERT_SIZE field. Only the non duplicated reads mapped on GRCh38 assembly are considered. No mapping qualiy is applied.