Research emerging from the University of Toronto found that using machine learning models was able to guide the design of long-acting drug formulations.