Basic Science and Pathogenesis

Area of Science:

• Neuroscience
• Genetics
• Computational Biology

Background:

• Minority stress may accelerate biological aging and worsen neurodegeneration, especially in transgender individuals on gender-affirming hormone therapy (GAHT).
• The link between minority stress, epigenetic aging, and Alzheimer's disease (AD) progression remains under-explored.
• This study uses an AI-driven approach to investigate these relationships.

Purpose of the Study:

• To explore the impact of minority stress on epigenetic aging and Alzheimer's disease (AD) progression.
• To develop and validate an AI-driven framework for analyzing these complex interactions.
• To identify potential protective factors, such as social support.

Main Methods:

• Integrated publicly available DNA methylation, neuroimaging, and socio-demographic data.
• Employed synthetic data augmentation for modeling GAHT exposure and minority stress.
• Utilized a Transformer-based Variational Autoencoder for data harmonization and ensemble Random Forests/Temporal Graph Convolutional Networks (T-GCN) for analysis.

Main Results:

• The T-GCN model demonstrated strong predictive performance for AD progression (AUC: 0.86).
• Minority stress was linked to an average epigenetic age acceleration of 6.2 years, with significant changes in FKBP5 and NR3C1 genes.
• Accelerated epigenetic aging correlated with reduced hippocampal volume and increased amyloid-β deposition; social support reduced epigenetic aging by 21%.

Conclusions:

• A novel AI framework was developed to understand minority stress-induced epigenetic aging and its effects on AD.
• Advanced computational techniques were used to address gaps in inclusive AD research.
• Findings contribute to the development of precision healthcare strategies for AD.