convenient tools to collect target mapping metrics

Question

[J.F.Jiang]

Hi all,

I want to collect some metrics for my NGS panel.

To evaluate the capture efficiency of the probes, I want to calculate the on target rate from reads and base.

For reads, bedtools is applied to get the rate, however, reads can be calculate twice if they span two probes. Any method to avoid this?

To calculate the rate from base, I first get the depth file from bam using samtools. The sum of all these depth is listed as M. Then probe base depth was extracted from the depth file among those targets., which is listed as N. Base on target rate is calculated as N/M. Any quick guide for more efficient tool?

Thanks. Junfeng

[GenoMax]

Picard tools would be one place to start.

[J.F.Jiang]

I've tried PICARD to get the HSMetrics from bam file. However, it is very interesting that it filter those reads not uniq mapped to the target, as PF_BASES_ALIGNED.

However, ON_BAIT_BASES and ON_TARGET_BASES are quite different, although same interval file is offered.

For example, 90% bases are uniquely mapped to the genome, 60% bases are mapped to the BAIT, 12% bases are mapped near to the BAIT.

But only 36.8% bases are mapped to the target.

I am wondering why so different results are obtained since the same interval file is provided.