Array-PGS vs. WGS-PGS

Here are the directories and file used for the PGS analyses in the paper:

• All_of_Us: Contains the main scripts used for PGS computation on the All of Us platform. All individaul-level information has been removed.

• PRS-CS: Contains scripts for running PRS-CS on a local machine. The outpus were uploaded back onto the All of Us platform to compute PGS.

• ldblk: Contains codes for generating the whole-genome LD matrix in TGP + HGDP EUR.

• LDSC: Contains codes for calculating heritability for the 10 Pan-UKBB tratis.

• PGS_Catalog: Contains codes for evaluating PGS-Catalog models in UKBB cohort.

• Simulation: Contains codes for simulating various trait settings based on real genotypes.

• Fine_mapping: Contains codes for identifying likely causal SNPs using SuSiE.

• phenotype_list.txt: Lists 10 Pan-UKBB GWAS traits and their sample sizes that were used in PRS-CS and LDSC.

Please refer to the README in each folder for more detailed descriptions.

Current publication:

Causal variant capture in genotype discovery approaches drives polygenic prediction performance across traits and populations

Yi-Sian Lin, Taotao Tan, Ying Wang, Bogdan Pasaniuc, Alicia R. Martin, Elizabeth G. Atkinson

bioRxiv 2025.03.18.644029; doi: https://doi.org/10.1101/2025.03.18.644029