To highlight specific HLA alleles or amino acid residues that confer overall PsV risk, we defined a multivariate full regression model including all the identified risk variants of HLA-C, HLA-B, HLA-A, and HLA-DQA1 (Table 1; Table S3). For HLA-C, we observed increased risk associated with HLA-C∗06:02 (OR = 3.26, 95% confidence interval [CI] = 3.02–3.52, p = 2.1 × 10−201) and HLA-C∗12:03 (OR = 1.38, 95% CI = 1.26–1.52, p = 6.5 × 10−12). For HLA-B, we observed increased risk associated with Cys67 (OR = 1.56, 95% CI = 1.45–1.67, p = 6.0 × 10−35), Met67 (OR = 1.44, 95% CI = 1.30–1.58, p = 2.6 × 10−13), and Asp9 (OR = 1.33, 95% CI = 1.21–1.45, p = 1.6 × 10−9). HLA-A Val95 (OR = 1.31, 95% CI = 1.25–1.38, p = 4.7 × 10−28) and HLA-DQα1 Arg53 (OR = 1.07, 95% CI = 1.01–1.13, p = 0.016) also demonstrated increased risk. In combination, these risk variants explained 6.7% of phenotype variance of PsV under the assumption of 2.0% of disease prevalence,2 whereas HLA-C∗06:02 alone explained only 4.9% of the variance.