Replication of a GWAS signal near HLA-DQA2 with AML using a disease-only cohort and external population-based controls

Acute myeloid leukemia (AML) is the most common type of acute leukemia in adults. Genome-wide association studies have identified 4 common inherited variants associated with AML risk, but these findings have not yet been confirmed in many independent data sets. Here, we performed a replication study with 567 AML cases from the Leucegene cohort and 1865 controls from the population-based cohort CARTaGENE (CaG). Because genotypes were generated using different technologies in the 2 data sets (eg, low- vs high-coverage whole-genome sequencing), we applied stringent quality-control filters to minimize type 1 errors. We showed, using data reduction methods (eg, principal component analysis and uniform manifold approximation and projection), that our approach successfully integrated the Leucegene and CaG genetic data. We replicated the association between cytogenetically normal AML and rs3916765, a variant located near <i>HLA-DQA2</i> (odds ratio, 1.96; 95% confidence interval, 1.26-3.06; <i>P</i> = .003). The effect size of this association was stronger when we restricted the analyses to patients with AML with <i>NPM1</i> mutations (odds ratios of >2.25). We also found that rs3916765 is a whole-blood expression quantitative trait locus for <i>HLA-DOB</i> (<i>P</i> = 1.05 × 10^-14) and <i>HLA-DQA2</i> (<i>P</i> = 2.23 × 10^-4). Our results confirm that a common genetic variant at the HLA locus associates with AML risk and the expression of HLA class 2 genes, providing new opportunities to improve disease prognosis and treatment.