Zamin Iqbal
Zamin Iqbal
You are right, but at the level of make-prg there is no concept of sites, apart from where you drop vertical slices of matches, separating regions within which there may...
Er..wow. @bricoletc what k size did you use for collapsing?
I agree about not clustering here "ideally". Still trying to understand. On phone while bloody laptop installs updates
i think you're 100% right.
in case helpful, this is the data @bricoletc posted above here: https://github.com/rmcolq/make_prg/issues/15#issuecomment-714718997 
I'm not a GATK user, so feel free to ignore this, but my suggestion, FWIW, is to call your default thing something other than a VCF file, as it breaks...
Thanks for this detailed reply! Those positions you generally mask are really valuable info which we can use to test our own assemblies. I completely agree, for heterogeneous data you...
Thanks @AngieHinrichs ! Will digest, but it does make me think the solution for us is to make our own covid-specific faToVcf which replaces deletion (minus sign) with ref. (It...
Actually could keep faToVcf but change "-" to Ref in the input to faToVcf . This would make it v quick to test if things look better.
Also @AngieHinrichs , we'll check out and answer this question : *if you look at BAMs for a few examples where the deletion is not called (ref in VCF instead...