Variance quantitative trait loci analysis for expression of coding genes in DLPFC

Polygenic diseases have complex genetic mechanisms. Previous studies primarily identified disease-related susceptibility loci through genome-wide association studies (GWAS) and elucidated the molecular mechanisms by which genetic variations affect disease susceptibility by integrating data from expression quantitative trait loci (eQTL) and other sources. The effect sizes of genetic variations discovered in GWAS are mostly low, and the ability of methods like polygenic risk scores (PRS) to explain phenotypic variation is also limited; for schizophrenia, this value typically does not exceed 0.2. There is a significant gap between the explanatory power of PRS for phenotypes and the heritability of the disease, a phenomenon known as "missing heritability." Gene-environment interactions are one explanation for this missing heritability. In recent years, an increasing number of studies have aimed to enhance the explanatory power for complex polygenic diseases by exploring gene-environment interactions. The complexity of polygenic diseases is also an important characteristic. Unlike previous GWAS or eQTL studies that focus on the effects of genetic variations on phenotypes or gene expression (such as increasing or decreasing disease risk, or promoting or inhibiting gene expression), variance quantitative trait loci (vQTL) focus on variance to explore genetic variations related to complexity. This complexity may arise from the environmental factors. The effects of genetic variations on phenotypes or gene expression depend on context; in different contexts, the effects of genetic variations may be gained, lost, or even reversed. The cumulative result of environmental complexity weakens the effects of variations, leading to their intricate impacts. Therefore, vQTL reflect genetic-environment interactions and have the potential to detect loci whose effects are obscured by environmental complexity. This study utilizes data from BrainSeq Phase I, which includes 414 DLPFC samples, and conducts cis-vQTL analysis for coding genes using OSCA. Additionally, we divided the samples intro different groups according to their age, sex, phenotype, and ethnicity to explore the commonalities or differences in vQTL under different grouping conditions, thereby evaluating the impact of grouping factors on complexity. To reduce the impact of outliers on the analysis results, rank-based inverse normal transformation (RINT) was implemented in this study, thereby generating a total of 18 vQTL results.