AI in the Identification of Genetic Biomarkers for Alzheimer's Disease: A Meta-Analysis of Computational Approaches

Background: Alzheimer’s disease (AD) is a progressive neurodegenerative disorder with increasing global burden, necessitating early diagnosis for timely intervention. Traditional genetic research methods often fall short in identifying complex biomarkers due to the nonlinear and high-dimensional nature of genetic data. The emerging application of artificial intelligence (AI) offers potential to overcome these limitations, yet a comprehensive synthesis of AI-driven genetic biomarker discovery remains lacking. Objective: This study aims to systematically review and narratively synthesize evidence on the application of AI techniques—particularly machine learning (ML) and deep learning (DL)—in identifying genetic biomarkers associated with Alzheimer’s disease, evaluating their diagnostic performance and potential clinical utility. Methods: A systematic review was conducted according to PRISMA guidelines. Five studies (n = 870) published between 2019 and 2022 were included, comprising observational, clinical trial, and computational designs. Inclusion criteria required peer-reviewed studies using AI to analyze human genetic data for AD biomarkers. Data extraction captured AI models used, biomarker targets, and diagnostic metrics. Ethical approval was not required as secondary data were analyzed, adhering to the Declaration of Helsinki. A narrative synthesis approach was applied due to methodological heterogeneity; statistical summaries were performed using R. Results: Deep learning algorithms demonstrated the highest diagnostic performance (AUC: 0.89–0.92; sensitivity: 85–90%; specificity: 82–88%), identifying both established (APOE, MAPT, TREM2) and novel biomarkers. Sample size and AI model type significantly influenced performance. No pooled effect sizes were calculated due to study heterogeneity. Conclusion: AI, particularly deep learning, exhibits superior potential in identifying genetic biomarkers for Alzheimer’s disease, enabling more accurate, early, and personalized diagnosis. These findings support AI’s integration into clinical genomics and pave the way for data-driven precision medicine in neurodegenerative disease management.