Mitochondrial dysfunction has been implicated in the pathogenesis of major depressive disorder (MDD); however, the causal contributions of specific mitochondrial genes across regulatory layers remain unclear.

We integrated genome-wide association study summary statistics from the Psychiatric Genomics Consortium and FinnGen with quantitative-trait-locus (QTL) datasets for DNA methylation, gene expression (eQTL), and protein abundance. Mitochondrial genes were annotated using the MitoCarta3.0 database. Summary-based Mendelian randomization and Bayesian colocalization were applied to assess causal relationships, with colocalization determined by the posterior probability of a shared causal variant (PPH4), and the false discovery rate used for multiple-testing correction. Brain-specific effects were evaluated using Genotype-Tissue Expression eQTL data. Prioritized genes were ranked based on cross-omics consistency and replication evidence.

Five mitochondrial genes were prioritized. TDRKH showed consistent associations across methylation, transcription, and protein levels, with hypermethylation at cg24503712 linked to reduced expression and a lower risk of MDD (Tier 1). METAP1D (Tier 2) demonstrated protective effects at both the transcript and protein levels. LONP1, FIS1, and SCP2 (Tier 3) exhibited consistent but complex regulatory patterns. Several signals were replicated in brain tissues, including TDRKH in the caudate and METAP1D in the cortex.

This study provides multi-omics evidence for the causal involvement of mitochondrial genes in MDD. TDRKH and METAP1D emerged as key candidates, offering promising targets for future mechanistic research and therapeutic development.