Integrative multi-omics and machine learning analysis identifies candidate biomarkers associated with mitochondrial quality control in major depressive disorder

Abstract

Background Major depressive disorder (MDD) possesses a complex pathogenesis, with abnormal mitochondrial quality control (MQC) proposed as a potential mechanism involved in the pathological process. Methods This study integrated two microarray expression profiling datasets with a single-nucleus RNA sequencing (snRNA-seq) dataset from the human prefrontal cortex (PFC). Candidate genes were identified by intersecting differentially expressed genes (DEGs) from the training set with MQC-associated module genes identified through WGCNA. Ten machine learning algorithms ranked MQC-associated candidate biomarkers, followed by preliminary mRNA-level verification using PFC tissues from chronic restraint stress (CRS) rats. Additionally, MQC-related gene set activity was computationally inferred at the single-cell level to examine cell-type-specific transcriptional alterations associated with MDD. Results The application of ten machine learning algorithms highlighted DCHS1 and HS3ST2 as candidate biomarkers linked to MQC-related transcriptional alterations. Gene set enrichment analysis (GSEA) indicated associations of these genes with oxidative phosphorylation and cytokine-cytokine receptor interaction pathways. In CRS rats, DCHS1 mRNA expression decreased, while HS3ST2 mRNA expression increased, aligning with bioinformatic findings. Among the 18 annotated cell types in the snRNA-seq dataset, computationally inferred MQRG activity significantly decreased in eight cell types, including several excitatory and inhibitory neuronal subtypes. Inhib_GRIK1 neurons exhibited cell-type-specific expression differences in DCHS1 and HS3ST2. Conclusion DCHS1 and HS3ST2 may serve as candidate biomarkers associated with MQC-related transcriptional alterations in the PFC of patients with MDD. MQRG activity demonstrated marked cell-type heterogeneity and reduction across multiple PFC cell populations. These findings provide preliminary, hypothesis-generating evidence for the link between MQC-related transcriptional dysregulation and MDD; however, further functional experiments are necessary to ascertain whether DCHS1 and HS3ST2 directly regulate mitochondrial quality control.