Entering edit mode
4.8 years ago
NS
▴
10
I need to perform genotype imputation on my data. The build of my data is hg18, so first I want to convert it into hg19. My Plink files are divided by chromosomes. When I use liftOver to lift my coordinates to hg19, I lose chromosomes 13,14,15,21 and 22 since none of the variants within these chromosomes are updated from hg18 to 19 (due to "partially deleted in new", "split in new", etc). How can I deal with this and not lose my data for these chromosomes?
If you cannot lift them it means that the regions you aligned to are not reliable so they did not make it to the newer assemblies. It would therefore be risky to build any assumptions on these variants. The best would be to start from scratch and realign to hg38 which is the standard since it came out in 2013. hg18 is from 2006, and hg19 from 2009. A lot was improved in hg38, if possible use this one.
How about I perform imputation for these chromosomes using hg18 reference panel and then lift them to hg19 after imputation ?