vcfeval --roc-subset behavior

23 views
Skip to first unread message

Glenn Carson

unread,
Mar 20, 2024, 8:20:18 PM3/20/24
to RTG Users
It looks like --roc-subset supports multiple variant types to be provided, e.g. --roc-subset hom,snp. However, so far it seems like the tool treats those as multiple separate restrictions (roc file for SNP's, and another for homozygous variants) rather than a unified restriction (roc file for homozygous SNP's). 

I wanted to sanity check if anyone was aware of a way to restrict in this way (dual restriction for SNP's that are homozygous). I wonder if --roc-expr would be a good way to do this. I could look for homozygous calls (GT: 1|1, 0|0) and restrict to single nucletide ALT's. 

Any ideas?


You can find a detailed explanation of our privacy practices here.

Len Trigg

unread,
Mar 20, 2024, 8:32:04 PM3/20/24
to Glenn Carson, RTG Users
Hi Glenn,

Yes, you should be able to use --roc-expr to do what you want.

It also turns out that we have a contributed PR that would probably achieve what you want -- essentially a flag that allows interpreting multiple roc subsets as outputting a cross product of their conditions (so if you gave --roc-subset het,hom,snp you would end up with two output roc files, one for het snps, and one for hom snps). If you're technically minded you could build that and have a play (https://github.com/RealTimeGenomics/rtg-tools/pull/13).

Cheers,
Len.

--
You received this message because you are subscribed to the Google Groups "RTG Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to rtg-users+...@realtimegenomics.com.
To view this discussion on the web visit https://groups.google.com/a/realtimegenomics.com/d/msgid/rtg-users/1bd71012-1015-44d7-bbb0-30398b1120a1n%40realtimegenomics.com.
Reply all
Reply to author
Forward
0 new messages