LD clump GRCh38 GWAS results

Question

The vignette of R package ieugwasr describes a plink based wrapper function for LD clumping GWAS data using the 1000 genomes reference dataset (http://fileserve.mrcieu.ac.uk/ld/1kg.v3.tgz). I would like LD clump GRCh38 based GWAS data. I am under the impression that the previous LD reference dataset is based on GRCh37. Therefore, I kindly ask:

• Is 1kg.v3.tgz based on GRCh37?
• Can you use GRCh37 based reference to LD clump GRCh38 GWAS results?
• Is there a GRCh38 reference version of 1000 genomes that is split to different super populations (EUR etc)?

Answer 1

Is 1kg.v3.tgz based on GRCh37?

Is there a GRCh38 reference version of 1000 genomes that is split to different super populations (EUR etc)?

Yes, the genome build for 1000 Genomes phase 3 release is in GRCh37. Luckily ISGR has reanalyzed their original sequencing data based on GRCh38. The data is provided both at the ISGR website https://www.internationalgenome.org/data-portal/data-collection/30x-grch38 and it is also made available by Christopher at PLINK2 website https://www.cog-genomics.org/plink/2.0/resources. However I suspect the data was not split into super populations.

Can you use GRCh37 based reference to LD clump GRCh38 GWAS results?

I wouldn't recommend doing so. Some variants identified in GRCh37 has been merged into others in GRCh38. Some variants are assigned different rsIDs. The position for the variant may also differ in different genome builds. Therefore clumping GRCh38 GWAS results with GRCh37 based reference may lead to fewer variants correctly mapped, which may influence your furthur analysis.

Hope you find it helpful. Best regards.