hg38 version of both baseline annotation and LD scores

JZ

unread,

May 1, 2019, 3:45:01 PM5/1/19

to ldsc_users

Hello, I am new to LDSC, and I am wondering for both baseline annotation (bed files) and LD scores (https://data.broadinstitute.org/alkesgroup/LDSCORE/), are they hg19 or hg38?

I want to do cell type-specific analysis on hg38, so I want to confirm first. Thank you very much.

Jory Schossau

unread,

Jun 26, 2019, 3:20:40 PM6/26/19

to ldsc_users

I believe the provided baseline is hg19. I'm finishing a workflow for hg38 baselining of arbitrary 1kgenome subsets (including the usual EUR) if you are interested. I could probably just share my baseline with you so you don't have to generate it.

Weifang Liu

unread,

Jun 29, 2019, 4:42:36 AM6/29/19

to ldsc_users

Hello,

I am also doing cell type-specific analysis on hg38. Could you share your baseline model? It would be very helpful. Thank you very much!!

Jory Schossau

unread,

Jul 9, 2019, 4:47:50 PM7/9/19

to ldsc_users

Sure! I'm making sure it all looks good and I'm looking into where to host it. The plink files are about 20 Gigabytes.

Maybe 1000 Genomes would like to host this...

Jory Schossau

unread,

Jul 31, 2019, 4:54:26 PM7/31/19

to ldsc_users

I'm still putting a bow on my workflow for arbitrary population sets and baselines, but a couple weeks after I emailed Broad and a few other organizations about hosting my hg38 baseline, Alkes Group got to work and uploaded an hg38 baseline.