We established the quantitative accuracy of our approach by several independent methods. First, to address sources of technical variation in our measurements, we established methods to normalize and filter the protein data as described in the Methods. Second, we observed that 13 pairs of antibodies targeting different epitopes for the same proteins resulted in significantly correlated measurements for their intended targets across all individuals (p < 0.05) (Figure S6). Third, we observed a strong preservation of interindividual rank order for the same protein quantified by both RPPAs and MWAs (example illustrated in Figure S2). Fourth, we verified that the antibodies faithfully reported on the levels of their intended targets by performing siRNA knockdown of 18 proteins, 15 of which were randomly selected and 3 of which were pQTL targets. Of the 18, 17 exhibited a significant reduction in protein levels relative to a scrambled control in at least one YRI LCL at one time point (Table S6).