Improved characterization of gene-environment interactions for vitamin D through variance quantitative trait loci

Background Understanding gene-environment interaction effects influencing vitamin D status may refine nutrition and public health strategies for vitamin D deficiency. Recent methodological advances have enabled the identification of variance quantitative trait loci (vQTLs) where gene-environment interaction effects are enriched. Objectives To identify vQTLs for serum 25-hydroxy vitamin D (25OHD) concentration and characterize potential gene-environment interaction effects of vQTLs. Methods We conducted vQTL discovery for 25OHD using a newly developed quantile integral linear model in the UK Biobank individuals of European (N = 313,514), African (N = 7,800), East Asian (N = 2,146), and South Asian (N = 8,771) ancestries, respectively. We tested for interaction effects between the identified vQTL lead variants and 18 environmental, biological, or lifestyle factors, followed by multiple sensitivity analyses. Results We identified 19 independent vQTL lead variants (p-value <5x10-8) in the European ancestry population. No vQTLs were identified in the non-European ancestry populations, likely due to limited sample sizes. A total of 32 interaction effects were detected with a false discovery rate <0.05. While known gene-season of measurement interaction effects were confirmed, additional interaction effects were identified involving modifiable risk factors, including time spent outdoors and body mass index. The magnitudes of these interaction effects were consistent within each locus upon adjusting for season of measurement and other covariates. We also identified a gene-sex interaction at a vQTL that implicates DHCR7. Integrating transcript- and protein-level evidence, we found that the sex-differentiated genetic effects may act through sex-biased expression of DHCR7 isoforms in skin tissues due to alternative splicing. Conclusions Through the lens of vQTLs, we identified additional gene-environment interaction effects affecting vitamin D status apart from season of measurement. These findings may provide new insights into the etiology of vitamin D deficiency and encourage personalized prevention and management of associated diseases for at-risk individuals.