How to handle multi-genome build in gtc2vcf?

[rajwanir]

In most newer chips (GSA, confluence, and even some Omni series), Illumina has manifests available on single genome build so you wouldn't observe this issue. But for older or custom chips (such as Consortium-OncoArray), you would likely run into this issue. In my view, liftover coordinate to single genome at this stage (this could even be hg19) is the most straightforward solution so everything downstream is simple and clear. Even if you don't lift at this stage, at imputation or other downstream work, this would have to be either lifted or filtered out.

For example for Consortium-OncoArray, here is the distribution of markers per genome build:

genomeBuild	markers
36	1191
36.2	1115
37	8718
37.1	522039
37.2	84
no build	484
total	533631

Workaround:
A possible option for gtc-to-bcf workflow for handling such manifests is to error out at config validation and let user know that is is a multi-genome manifest either liftover manifest or use gtc-to-ped workflow.