Drop hom_ref calls in sample optimized tables

googlegenomics / gcp-variant-transforms

GCP Variant Transforms

Apache License 2.0

134 stars 55 forks source link

Drop hom_ref calls in sample optimized tables #689

Closed samanvp closed 3 years ago

samanvp commented 3 years ago

In order to make the size of sample optimized tables more manageable we are dropping homozygous reference calls. This means sample optimized tables will only include:

variants
no calls

samanvp commented 3 years ago

I ran multiple manual tests and confirmed the modified query is generating the correct output.

I will update docs in a follow up PR.

moschetti commented 3 years ago

Per conversations with customers, this change would have a negative impact on the workflow. As the genotype probability could change with increasing number of samples based on the statistics of that variant. Therefore rejecting this PR as the benefits would not outweigh the potential impact.