Extended Data Fig. 7: Polygenicity of PGS to protein associations.

Linkage disequilibrium (LD) blocks contributing to each PGS to protein association in polygenicity tests. Briefly, each PGS was partitioned into 1,703 approximately independent LD blocks⁵⁴ then tested for association with each protein in linear regression adjusting for age, sex, 10 genotype PCs, sample measurement batch, and time between blood draw and sample processing in 3,087 INTERVAL participants. Full summary statistics including exact two-sided P-values for these tests are detailed in Supplementary Data 3,e. Next, to obtain the set of LD blocks contributing to each PGS to protein association, LD blocks were sequentially removed from the PGS in ascending order by association P-value (two-sided) until the association between resulting PGS and protein levels were attenuated (two-sided P > 0.05). Full summary statistics including exact two-sided P-values for these tests are detailed in Supplementary Data 3,f. The polygenicity of PGS to protein association (% of genome) shown on the left (and in Fig. 1c) was subsequently computed based on the sum of lengths of all contributing LD blocks (in base pairs) as a proportion of the genome. Here, associations (−log₁₀ two-sided P-values) between protein levels and LD blocks contributing to the PGS to protein association are shown. Regions in white contain LD blocks that did not contribute to the PGS to protein association. PGS to protein associations listed in red are those explained by pQTLs (cis and/or trans) rather than polygenic.