svim-asm on multiple samples

[dfguan]

Hello David,

Currently I am running svim-asm on mulitple samples(species) separately, but I wonder if I can run sv calling on the samples together and generate a vcf with multiple samples. Or is there a way or tool to merge those SVs from different samples, regenerate a unified SV set?

Best,

Dengfeng.

[eldariont]

Hi Dengfeng,

thanks for your question. Unfortunately, SVIM-asm does not support multi-sample SV calling. However, you can run SVIM-asm separately for each sample (as you already did) and merge the calls later with a tool like SURVIVOR. Please have a look at a similar issue created for SVIM here.

Cheers David

[dfguan]

Hi David, thanks for your reply. I have tried SURVIOR to merge those SVs, however there are too many missing genotypes. Is there a way to genotyping those merged SVs? Cheers, Dengfeng.

[eldariont]

Hi Dengfeng,

when you run SURVIVOR it will try to combine similar calls from different samples. The genotypes of the merged calls will either reflect the genotype in the original sample VCF (if the variant is present there) or ./. if the variant is not present there. Therefor, your merged set already contains genotypes.

However, it might be a good idea to re-genotype the merged calls but SVIM-asm does not support this at the moment.

Best David

[athenasyarifa]

Hi Dengfeng and David, I'm doing a similar approach, running svim-asm using multiple assemblies and trying to merge them together. I encountered the same problem where I have a lot of missing genotypes. Would it be a good idea to do force-calling step using Sniffles with the merged vcf file? I have not tried this yet, and I suspect it might fail. So I am asking just in case you found a work-around solution on how to do this for comparative genomics?

Many thanks in advance! Cheers, Rifa