SCO2, LYN, and FGR as Mitochondria-Associated Immune-Inflammatory Candidate Signatures Across PMOS and MCI Datasets

Background: Polyendocrine metabolic ovarian syndrome (PMOS), previously known as polycystic ovary syndrome (PCOS), is a prevalent endocrine disorder increasingly associated with mild cognitive impairment (MCI). Mitochondrial dysfunction and immune dysregulation may underlie this association; however, shared molecular features between these conditions remain systematically uninvestigated. Methods: We used PMOS (granulosa cell, n = 16) and MCI (peripheral blood, n = 243) to identify differentially expressed genes (DEGs) intersecting mitochondrial-related genes (MRGs), supported by enrichment and prioritized by machine-learning with cross-validation. Boxplots in PMOS and MCI confirmed directionality. Single-cell RNA sequencing (scRNA-seq) identified cell types and communication. In vitro experiments assessed SCO2, LYN, and FGR in cell lines. Results: We identified 16 mitochondrial-related genes shared by PMOS and MCI datasets. Enrichment implicated NF-κB and TNF signaling in PMOS/MCI immunometabolic features. SCO2, LYN, and FGR were upregulated across datasets, although LYN was discordant in granulosa comparison. Single-cell sequencing identified five subtypes. Pseudotime showed SCO2 and LYN upregulation in early/late stages, whereas FGR was undetectable. Secondary annotation identified plasmacytoid dendritic cells, classical monocytes, and pro-B cells; pro-B cells predominated in controls (47.5%) and classical monocytes in disease (40.5%). In vitro validation showed mRNA/protein upregulation of SCO2, LYN, and FGR in PMOS/MCI models versus control. Conclusions: This study identified SCO2, LYN, and FGR as mitochondria-associated immune-inflammatory signatures in PMOS and MCI datasets. Differential expression, feature-ranking models, pathway enrichment, transcriptional deconvolution, single-cell contextualization, and cell assessment support were utilized. Without PMOS–MCI cohorts and functional perturbation experiments, findings remain hypothesis-generating, unvalidated biomarkers or causal mechanisms.