How to merge indels

[weishwu]

I'm calling variants from 10 samples using one reference fasta for all. The samples are from the same strain. One sample is WT and the others were under different stress selection. The purpose is to find out whether the stress brought any mutations and what they are.

In order to do this, I need to merge the variant calling results from all samples. For example, if a position is 0/0 in the WT sample, but 1/1 in one of the stressed samples this would be of our interest.

Ideally I would want to do a joint calling from all samples together. But Snippy can only call each sample individually and it only outputs variants so all 0/0 sites are lost. I know I can use snippy-core to merge SNPs to get those 0/0 sites back, but it does not handle indels.

My questions are: (1) Is it possible to let snippy output all callable sites including 0/0? Or (2) Is it possible to let snippy-core output all variants including indels?

Any other solution?

Thanks,

[tseemann]

(1) you could try passing --report-monomorphic to freebayes using the --fbopt parameter. it might work

(2) it could be possible with lots of code changes BUT indels are much trickier to align with each other. if not handled correctly you will get false indels. the proper way to do this is joint SNP calling unfortunately.

How many isolates do you have in the sets?

[b-tierney]

Hi -- has there been any progress on this re: joint variant calling or merging indels? It seems like Freebayes could handle joint calling under the hood, but then again this is mostly out of my depth.

I'm in the same position as the original poster -- Really love snippy, would love to use its output in a GWAS but need to look across more than just SNPs and also have 0/0 calls.