Capture homozygous reference INDELs based on individual positions

[BinglanLi]

We have received a request to improve how the VCF Preprocessor recognizes homozygous reference INDELs. See the description and discussion of the issue on a previous issue thread. The idea is that if all genomic positions for an INDEL variant are homozygous reference, then, this should be sufficient evidence to deduce that no genetic variation is observed and the individual should be represented as homozygous reference for the INDEL variant.

This improvement has more use for basepair-resolution gVCF files that are generated by the GATK HaplotypeCaller.

This feature is currently waiting on the bcftools team to address a bug in merging INDEL records, bcftools issue #2163.

We will continue developing the feature once the bug is resolved.

[moe1619]

bcftools seems to have closed this issue
samtools/bcftools#2163

[BinglanLi]

I appreciate the reminder. The issue is closed but the problem still lingers, e.g., case 2 in the samtools/bcftools#2163. If the recommendation from the bcftools team works for your data, that is great. On the other hand, this feature is still under development and testing.