version of reference panel
Opened this issue · 2 comments
Hello, @WansonChoi ,
It's very kind of you create this method for HLA imputation. I noticed that this software automaticlly transformed the imputed targeted data to hg18, but current data are mostly based on hg19 and hg38, how should i do if i'm willing to construct my own reference panel in hg19 and impute based on it?
Waiting for your kindly reply.
Hi, Thank you for your interest in CookHLA.
To make a custom reference panel, We need (1) HLA type data, (2) SNP data, and (3) IMGT database. There are practically two ways. If you are fine with using the fixed v3.7.0 IMGT database, then you can use the MakeReference module of SNP2HLA(https://software.broadinstitute.org/mpg/snp2hla/makereference_manual.html). If you want to use the latest or specific version of the IMGT database, you have to use HATK(https://github.com/WansonChoi/HATK#3-usage-example).
The latter way requires more post-processings and handling of errors. If you want the latter, try HATK first and let me know whether it succeeded. Then I'll let you know how to phase its output and what more post-processings are needed.
I have used SNP2HLA to make a HG38 reference panel. However when I use python -m MakeGeneticMap althoug I have indicated that my data is HG38 it still tries to uplift my data. Any suggestions?