Unraveling the role of protein folding-related genes in depression through multi-omics mendelian randomization

Abstract

Depression has a substantial genetic component. Protein folding and proteostasis pathways have been implicated in depression-related biology, but the genetically supported associations between specific protein folding-related genes and major depressive disorder (MDD) remain insufficiently understood. We applied summary-data-based Mendelian randomization (SMR) to integrate MDD genome-wide association study data with blood-derived cis-methylation quantitative trait loci (mQTLs), cis-expression quantitative trait loci (eQTLs), and cis-protein quantitative trait loci (pQTLs) summary datasets. Candidate signals were evaluated using single-SNP and multi-SNP SMR, heterogeneity in dependent instruments (HEIDI) testing, colocalization analysis, false discovery rate (FDR) correction, supportive FinnGen depression analyses, bipolar disorder cross-disorder analyses, and exploratory Gene Expression Omnibus (GEO) brain-region expression analyses. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were further performed to characterize functional context and biological modules associated with prioritized candidate genes. In the discovery cohort, 86 methylation sites, 15 genes, and 3 proteins passed the prespecified SMR screening criteria. After FDR correction, 11 methylation sites from 5 genes remained FDR-significant, whereas expression and protein-level findings did not survive multiple-testing correction and were nominal or suggestive signals. Colocalization analyses supported shared genetic signals for 10 methylation sites, 4 eQTL-associated genes, and 2 proteins, of which 4, 2, and 1 signals, respectively, showed strong colocalization evidence (PP.H4 > 0.8). Cross-omic analyses provided genetically supported evidence for potential regulatory relationships between FNIP2 methylation-related signals and gene expression, and between PDIA3 expression and protein abundance. These molecular traits also showed exploratory associations with MDD risk across corresponding omic layers. Enrichment implicated proteostasis-related functional modules, including endoplasmic reticulum stress, unfolded protein response, chaperone-mediated folding, protein processing, quality control, and antigen presentation. Supportive cohort and brain tissue-related analyses remained exploratory. This study prioritized protein folding-related candidate genes potentially involved in MDD susceptibility, particularly PDIA3 and FNIP2 . These findings were genetically supported and hypothesis-generating, and require independent replication and functional validation.