Normalization of INDELs: required or should be avoided

[davidyuyuan]

Normalization of INDELs: required or should be avoided

[holtjma]

Hello,

HiPhase does not require it to run successfully (e.g. most of the TRGT calls have not been through a traditional normalization step). However, the tools we currently support should normalize the variants by default (excluding TRGT).

If you're using other variant calling software, then I probably recommend normalizing the variants. For a single VCF file, I doubt this would have a major impact. But if you have 2 or more input VCFs to phase, then normalization starts to become more important for detecting identical and/or overlapping calls between the two VCFs.

Matt

[davidyuyuan]

Thanks, Matt.
I am using bcftools mpileup and bcftools call to generate a VCF file for 1 sample at a time. Its norm command does correct a good number of INDELs.
Thank you for the clarification. I will normalize the raw VCFs before sending them to HiPhase.
David

[holtjma]

Closing for now, but feel free to re-open if there are follow ups!