Closed bopohdr closed 5 years ago
Because of the significant computational burden of gVCF merging and joint genotyping, I'd suggest running it first and then applying straightforward filters like you mention downstream, using bcftools, Apache Spark, or the like. This way you don't have to re-run joint genotyping just to fiddle with those filters. However it is possible to reconfigure GLnexus to increase required_dp which sounds like what you want. You can customize the configuration by copying the unifier_config
and genotyper_config
sections into my_config.yml and passing --config my_config.yml
to GLnexus.
Thank You. Yes, that is what I want. Is it possible to execute somehow also online on DNAnexus ?
Closing this as the discussion has moved to #161 - thanks
GLnexus as output genotypes have mostly 4 options: 0/0 - reference hom 0/1 1/1 - alt hom ./. - no reads Although if the specific locus is sequenced by only 1 read , sample have genotype as 0/0 or 1/1 . Is it somehow possible to specify a minimal quality / depth to call as ./. genotype (e.g. all locuses with DP<10) ?