RNA_Seq dataset of manuscript entitle Multi-omics Mendelian randomization integrating GWAS, eQTL, mQTL and pQTL data prioritizes mitochondrial gene FXN as a hypothesis‑generating candidate for nephrolithiasis

Background: Mitochondrial dysfunction is linked to urolithiasis, but causal genetic drivers remain unclear. We integrated multi-omics data using Mendelian randomization to identify mitochondrial-related genes causally associated with urolithiasis. Methods: We obtained mitochondrial methylation (mQTL), gene expression (eQTL), and protein abundance (pQTL) from respective quantitative trait locus (QTL) studies along with GWAS summary data for nephrolithiasis, ureterolithiasis and bladder calculus from the Million Veteran Program (discovery), with replication in FinnGen and UK Biobank. Summary-data-based Mendelian randomization (SMR) and colocalization were applied to infer causality. Results: Integrated analysis identified FXN as the leading candidate for nephrolithiasis. Genetically elevated circulating FXN protein was inversely associated with nephrolithiasis risk (OR 0.69, 95% CI 0.56–0.84). This protective effect was supported at the epigenetic level: FXN methylation at cg14656297 and cg13974534 correlated with lower nephrolithiasis risk. For ureterolithiasis, higher GRHPR protein levels were protective (OR 0.81, 95% CI 0.71–0.91). In a mouse kidney stone model, Fxn expression and frataxin protein levels were decreased, providing correlative in vivo support consistent with human genetic findings. Conclusion: This multi-omics MR study links mitochondrial genes, particularly FXN, to urolithiasis risk. Although colocalization evidence was weak (PP.H4 = 0.0239) and replication in independent cohorts was not statistically significant, the multi‑omics consistency across methylation, expression, and protein levels prioritizes FXN as a hypothesis‑generating candidate for further investigation. Because all QTL data are blood‑ or plasma‑derived, this study provides blood/plasma QTL‑based genetic prioritization rather than kidney‑specific causal inference.