Students from the University of North Carolina at Chapel Hill Eshelman School of Pharmacy has created artificial intelligence (AI) which designs drug molecules from scratch.
Olexandr Isayev, Mariya Popova and Alexander Tropsha, created a computer and algorithm programme that comprises two neural networks named Reinforcement Learning for Structural Evolution (ReLeaSE), which work similarly to a teacher and a student.
Over time, by working with the teacher, the “machine learning student” will become better at proposing molecules that are likely to be useful as new medicines.
“If we compare this process to learning a language, then after the student learns the molecular alphabet and the rules of the language, they can create new ‘words,’ or molecules,” explained Alexander Tropsha, one of the ReLeASE creators, to ScienceDaily.
“If the new molecule is realistic and has the desired effect, the teacher approves. If not, the teacher disapproves, forcing the student to avoid bad molecules and create good ones.”
ReLeASE is an innovation to virtual screening. It has the ability to create and evaluate new molecules used by the pharmaceutical industry, as well as to create and evaluate new molecules by accelerating the design of new drug candidates.
The university team are also using ReLeASE to generate molecules with properties such as desired bioactivity and safety profiles.
Written by Leah Alger