Bioinformatic analysis and experimental validation of mitochondria-related genes as novel biomarkers for moyamoya disease.

Moyamoya disease (MMD) is a rare chronic cerebrovascular disorder, yet the role of mitochondria in its pathogenesis remains unclear, hindering progress in understanding pathological mechanisms and identifying therapeutic targets. However, mitochondria-related genes (MiRGs) have not been systematically investigated in MMD, and their pathogenic relevance remains unclear. This study aimed to identify MMD-associated MiRGs, evaluate their diagnostic potential through integrative bioinformatics and machine learning, and validate key findings experimentally. We retrieved two MMD transcriptomic datasets (GSE157628, GSE141025) from the Gene Expression Omnibus (GEO) and obtained MiRGs from MitoCarta 3.0. Three machine learning methods were used to screen hub genes, with validation via an independent cohort (GSE141025) and reverse transcription-quantitative polymerase chain reaction (RT-qPCR). Three hub MiRGs (SLC25A20, METAP1D, and SPIRE1) were confirmed to be associated with MMD, as supported by transcriptomic analysis, machine learning, and RT-qPCR validation. As the first comprehensive study to systematically explore MiRGs in MMD-distinct from prior work focused on mitochondrial function or clinical observations-our findings provide potential diagnostic biomarkers for early MMD detection and guidance for developing therapies targeting mitochondrial function restoration.