Discovery of novel microglial homeostasis modulators through machine learning

Abstract

Identifying molecules capable of reducing microglial inflammation has been a major goal in neurodegeneration research, as dysregulated inflammation is a hallmark of most neurodegenerative diseases, and microglia, the brain’s tissue-resident macrophages, play a large role in initiating this inflammation. The traditional approach of drug discovery through screening thousands of compounds is both costly and time-consuming. Therefore, inspired by a study by Smer-Barreto et al. (2023), we utilized data from a previous drug screening conducted by our lab to develop a machine-learning model that can identify new candidate drugs from online databases. Using this approach, we identified 36 promising compounds that may have anti-inflammatory eLects on microglia and are performing experimental validations on them. Should the lab results return positive, this proof-of-concept study will demonstrate the validity of machine learning-assisted drug screening in inflammation research and facilitate the development of more eLicient screening methods. Furthermore, the validated hits will be added to the repertoire of neurodegenerative therapeutics and help us study the mechanisms governing microglial homeostasis.