The polygenic indexes (PGIs) shared through the Repository are based on summary statistics from three types of sources: GWASs conducted in UK Biobank (UKB), GWASs conducted in samples of volunteer research participants from 23andMe, and other published genome-wide association studies (GWAS). In Section I below, we begin by detailing how the Repository facilitates researchers’ access to the PGIs. In Section II, we describe how the summary statistics used in our main analyses were generated, quality-controlled and meta-analysed to generate a set of files used as inputs into construction of the single-trait and multi-trait PGIs. In Section III, we define and justify the R2 criterion we used to determine which PGIs to include in the first release of the Repository. We then describe quality-control filters applied to the individual-level genotype data supplied by each Repository cohort. We conclude by describing the methods used to construct the cohort PGIs. In Section IV we state our measurement-error-corrected estimator and its standard error in terms of estimable quantities. Section V describes our estimation of ρ in the HRS, WLS and UKB. Section VI describes the data underlying Figure 1.
