Decoding Dementia Mechanisms: Identification of Key Oligodendrocyte-Associated Genes through Integrative Bioinformatics and Machine Learning.

Introduction: This study aims to elucidate the mechanisms underlying Dementia using bioinformatics analysis and machine learning algorithms, to identify novel therapeutic targets for its clinical management.

Methods: Gene expression datasets related to dementia were sourced from the GEO database. Differentially expressed genes (DEGs) were identified using R, and key module genes were determined through the Weighted Gene Co-expression Network Analysis (WGCNA) method. Oligodendrocyte (OL) related targets were retrieved from the GeneCards database. The intersecting genes from DEGs, WGCNA, and OL were analyzed using Gene Ontology and the Kyoto Encyclopedia of Genes and Genomes. Subsequently, three machine learning algorithms were employed to pinpoint core genes associated with OL in dementia. The CIBERSORT algorithm was used to evaluate the abundance of 22 immune cell types and their correlation with Dementia-related immune infiltration. Validation was carried out via quantitative reverse transcription polymerase chain reaction (RT-qPCR).

Results: Through bioinformatics and machine learning techniques, six core OL genes associated with Dementia were identified, notably C1QA, CD163, and TGFB2, which showed elevated expression in Dementia. Immune cell infiltration analysis indicated that several immune cell types may contribute to Dementia's pathogenesis, and RT-qPCR results corroborated the bioinformatics findings.

Discussion: The discovered genes may contribute to dementia pathogenesis through oligodendrocyte dysfunction and neuroimmune interactions. Notably, TGFB2 and complement-related genes (C1QA, CD163) suggest involvement in both myelination defects and neuroinflammation, highlighting their therapeutic potential.

Conclusion: The six feature genes: TGFB2, C1QA, CD163, ACTG1, WIF1, and OPALIN are significantly linked to Dementia.