A collaborative team across 11 research institutions is applying advanced machine learning techniques to an extensive dataset of Alzheimer’s disease biomarkers and cognitive data to address the high failure rate of Alzheimer’s disease clinical trials since 2003. The project is part of the Artificial Intelligence for Alzheimer’s Disease Initiative, which aims to identify key features within the genome, biomarkers, and cognitive information to address fundamental barriers to prevention and drug discovery. Researchers believe that Alzheimer’s disease is a multifactorial disease that requires a personalized therapeutic approach, a strategy that has been successful in cancer but remains largely unexplored in neurodegeneration. The researchers suggest that precision medicine involves developing patient-specific treatment solutions based on genetic profiles and the availability of existing drug options, and that this effort will build AI/ML-enabled databases that facilitate scalable evaluation of potential treatments.
High failure rates increase the need for AD precision medicine
Since 2003, almost all Alzheimer’s disease clinical trials have failed, with about 99 percent not providing statistically significant benefit to patients. This statistic highlights the critical need for a different approach to treatment development. Researchers now believe that the high failure rate is due to significant heterogeneity among trial participants, suggesting that Alzheimer’s disease is not a single disease but a collection of symptoms with diverse underlying causes and clinical manifestations. This recognition has led to an increased focus on precision medicine. Although this strategy has been successfully implemented in oncology, it has remained largely unexplored in neurodegenerative diseases. A collaborative team of 11 research institutions aims to address this gap by analyzing a wide range of genomic, biomarker, and cognitive data to pinpoint the characteristics essential for disease prevention and finding effective drugs.
Jun, director of the AI4AD Initiative’s Genome-Guided Drug Discovery Core, is leading efforts to identify network-based signatures to prioritize genetic variants and repurpose existing drugs. Jun’s lab is developing an “AI/ML-enabled database” by profiling characteristics of predictors, signatures, biomarkers, and outcomes (PreSiBO). This facilitates scalable evaluation of potential treatments and enhances collaboration between researchers. According to the research team, this expanded effort will strengthen the readiness for AI/ML applications in precision medicine for the prevention and treatment of Alzheimer’s disease.
PreSiBO database enables drug reuse using AI/ML
The failure rate of clinical trials for Alzheimer’s disease has increased to approximately 99% since 2003, driving a shift toward leveraging existing data through artificial intelligence and machine learning approaches to identify potential therapeutic interventions. Central to this effort is the development of PreSiBO, an AI/ML-enabled database designed to profile predictors, signatures, biomarkers, and outcome characteristics for both targets and existing drugs associated with specific genome-guided patient subgroups. Recent research from Dr. Jun’s lab focuses on applying AI/ML to target prioritization and drug repurposing, demonstrating the increasing emphasis on precision medicine concepts in Alzheimer’s disease research. The creation of PreSiBO aims to increase the readiness of AI/ML applications for both the prevention and treatment of Alzheimer’s disease, providing a critical resource to accelerate the development of patient-driven treatment solutions.
Precision medicine is about developing patient-driven treatment solutions, depending primarily on genetic profiles and the availability of existing drug options.
