Researchers from the University of Virginia have developed a new machine learning approach to identify potential drugs that can minimize harmful scarring after a heart attack or other injuries. This cutting-edge computer model has already found a promising candidate that could prevent scarring in a unique way compared to previous drugs. The researchers believe that their machine learning tool has the potential to predict and explain the effects of drugs for various diseases.
According to Anders R. Nelson, a computational biologist and former student in UVA’s lab, machine learning helps to reduce the complexity of treating diseases like heart disease, metabolic disease, and cancer. It enables scientists to identify the crucial factors contributing to the disease and better understand how drugs can modify diseased cells.
The researchers combined a computer model based on decades of human knowledge with machine learning to study the effects of drugs on cells called fibroblasts. Fibroblasts play a role in repairing the heart after an injury and can cause harmful scarring in the process. By using machine learning and human learning together, the team aimed to find drugs that would prevent scarring and improve patient outcomes.
Previous attempts to identify drugs targeting fibroblasts were limited in scope, and understanding how these drugs work remained unclear. To address this, the UVA researchers developed a new approach called logic-based mechanistic machine learning. This approach not only predicts drugs but also predicts how they affect fibroblast behaviors.
The researchers tested 13 promising drugs on human fibroblasts and used the data to train the machine learning model. The model successfully predicted how the drug pirfenidone, approved by the FDA for idiopathic pulmonary fibrosis, suppresses contractile fibers in fibroblasts, which stiffen the heart. It also predicted how another experimental drug, the Src inhibitor WH4023, could target a different type of contractile fiber.
While further research is required to confirm the effectiveness of these drugs in animal models and human patients, the results suggest that mechanistic machine learning has the potential to revolutionize the development of new treatments. The researchers hope to use this approach not only for heart injury but also for other diseases.
The findings of this study have been published in the scientific journal PNAS (Proceedings of the National Academy of Sciences). The research team included scientists from UVA’s Department of Biomedical Engineering. The study was supported by several grants from the National Institutes of Health.
In conclusion, the University of Virginia scientists have made significant progress in using machine learning to identify potential drugs for minimizing harmful scarring after heart attacks or other injuries. This new approach has the potential to advance the development of targeted treatments for various diseases by predicting and explaining the effects of drugs.
