Dear GATK team,
I would want to see restored the StrandAlleleCountsBySample (SAC) annotation, which appears in GATK3 but is no longer in the current code. At least for our lab, it's nearly the last thing preventing us from migrating to GATK4.
Our lab uses an ensemble genotyping pipeline, and as a result some hard thresholds are applied to calls from all callers; some of them involves a minimum number of alternate-allele reads on each strand. In GATK3, we filter HC calls using a JEXL filter over the SAC annotation. However, this annotation is removed in GATK4 without replacement, and the closest substitute, StrandBiasBySample (SB), isn't exactly the same thing since it only reports the sum of all non-ref counts.
What is more curious is that the subsetting of the SAC annotation is only implemented in GATK4 (PR #1852), which means if operations involving allele subsetting is involved (such as splitting multi-allelics), we'd need to perform the weird process of calling by GATK3 and processing by GATK4.
Thus, I would like the GATK team to port SAC to GATK4. Thanks!
Please sign in to leave a comment.