Hannes Pétur Eggertsson

> But then we would like to keep the version number the same compared to the rest of the cohort, and the tries at REF_PATH setting has not been successful...

Hi. You should only keep the "SVMODEL=AGGREGATED" records, each of them should correspond to an SV allele in your input VCF. ```sh bcftools view --include "SVMODEL='AGGREGATED'" -Oz -o gt.agg_only.vcf.gz gt.vcf.gz...

Yep I see, it would make the most sense to also restrict to only aggregated in that case. > Is there a downside to keeping only AGGREGATED ? Sometimes, for...

I see, so in your example it's considered 1 true positive and 5 false positives (or 6 false positives). I guess it makes sense to just keep one version then....

Hello, yes only diploid genotyping model is available and we don't have any guidelines for genotyping somatic SVs. You *could* discover the somatic SVs with Manta and then call them...

Hello, if you some assemblies in FASTA you can generate a VCF (not BAM) using dipcall, see: https://github.com/lh3/dipcall This is what we used in our SV paper. Best, Hannes

It has issues with determining the SV type of some record. I will have a look but for now, use the "O" option in python: ```sh $ python -O svimmer...

I would think this is happening in regions that have very high alignment depth. In earlier versions of graphtyper SV genotyping, we had a high coverage downsampling filter but later...

The `--avg_cov_by_readlen` option to subsample reads has been added to graphtyper v2.6.1.

Hey, I am not sure when I will have time to work on this feature but its good to know that it would be useful. Best, Hannes