Machine Learning Model Predicts Drugs to Prevent Harmful Heart Scarring
-
University of Virginia scientists developed a new machine learning approach to identify drugs that minimize harmful scarring after heart injury.
-
The model found a promising drug, pirfenidone, already FDA-approved for pulmonary fibrosis, that prevents harmful heart scarring.
-
The model also predicted that the experimental drug WH4023 could target contractile fibers in heart fibroblasts to prevent scarring.
-
This "logic-based mechanistic machine learning" predicts how drugs affect fibroblast behavior and scarring, not just the drugs themselves.
-
Researchers plan to test pirfenidone and WH4023 in animal models to verify they prevent heart scarring as predicted.