Combining AI and AlphaFold to design better therapies for autoimmune encephalitis

Funding

As a proof-of-concept, a study could collect samples from biorepositories such as those supported by the National Institutes of Health (NIH), the ARUP Institute for Research and Innovation in Diagnostic and Precision Medicine, the Mayo Clinic Biobank, and the Mayo Clinic Platform. The aim would be to aggregate, standardize, and analyze data to build basic elements of AE research. Numerous other repositories could also provide samples, though all such initiatives need to comply with HIPAA regulations and secure patient consent to move forward. Since AE exhibits significant inter-patient variability that demands personalized treatment strategies, AI can be used to integrate AlphaFold-derived structural data with multi-omics profiles (genomics, proteomics, and transcriptomics) to stratify patients and predict therapeutic responses. Random forest or gradient boosting models can correlate protein structural features—such as mutation-induced conformational shifts—with clinical outcomes, identifying biomarkers of disease severity or treatment resistance.