Carnegie Mellon University’s Olexandr (Oles) Isayev is part of a new breed of researchers blurring the boundaries between computational, theoretical and experimental chemistry.

Isayev joined Carnegie Mellon’s Department of Chemistry as an assistant professor in 2020. Using machine learning and neural networks, Isayev is developing technology that has rapidly accelerated the pace at which new molecules are being discovered through a combination of artificial intelligence (AI), informatics and high-throughput quantum chemistry.

“While chemistry as a field traces its roots back to the earliest societies, computational chemistry is less than 100 years old,” he said. “People don’t realize it’s very important in the practical applications of materials and drugs. A lot of discoveries start from computations.”

During his brief time at Carnegie Mellon, Isayev’s research program has generated impactful papers and earned several grants that led him to being promoted to associate professor more rapidly than usual.

Prior to joining Carnegie Mellon, Isayev was a research assistant professor at University of North Carolina at Chapel Hill in the Eshelman School of Pharmacy.

“It was an interesting place to learn more of the patient-facing world related to drug discovery,” Isayev said. “I admired them and was grateful for everyone who helped me. But I felt a little bit that my home department was chemistry, so when the opportunity to come to Carnegie Mellon arrived, I had no second thoughts.”

Among the draws to Carnegie Mellon is access to computational power and advanced technology.

As a student at Dnepropetrovsk National University in Ukraine, Isayev was drawn to both computer science and chemistry.

“I mixed liquids, I grew up a nerdy kid. I liked to program and I participated in different team events based on chemistry and computers,” he said. He elected to earn his master’s degree in chemistry at Dnepropetrovsk National University in Ukraine and his doctorate in theoretical chemistry at Jackson State University. “Being a computational chemist combines both of these passions.”

Carnegie Mellon’s reputation for technological advances goes beyond access to computing resources. The idea of building an academic cloud lab was in the works when Isayev interviewed to join Carnegie Mellon’s Department of Chemistry.

“I spoke with Linda Peteanu, then head of Chemistry, and Rebecca Doerge, and I could see their vision of AI and automation. It was a bold vision of the technology,” Isayev said. “It clicked with my passion and hopes of where chemical sciences were going.”

Emerald Cloud Lab, a company founded by Carnegie Mellon alumni Brian Frezza and DJ Kleinbaum, runs a remotely operated research facility that can handle many aspects of daily lab work, from experiment design to data acquisition and analysis. The company is partnering with Carnegie Mellon to build the first academic cloud lab, which will be online later in 2023.

Isayev and his students have been at the forefront of Carnegie Mellon’s efforts to adopt Emerald Cloud Lab’s internal platforms, which they started using nearly two years ago.

“Since my lab is computational, we developed algorithms to better connect their platforms and instruments,” he said. The work creates a feedback loop where a model can execute experiments and teach machines to make adjustments based on if the certain criteria are met. Part of his funding supports work using the Emerald Cloud Lab facilities. Once Carnegie Mellon’s Academic Cloud lab opens later this year, that research will transfer to Pittsburgh.

The ability to automate some of the more repetitive tasks required to conduct experiments allows researchers, in particular graduate students, to be freed from tedious tasks that have long been considered rites of passage.

“It’s liberating. I know how hard graduate students work in the lab and have had long hours and long nights. If a culture needs to be fed, there are certain things you have to do, but optimization technology can take some of those tasks then all those new researchers can unleash their creativity in new ways,” Isayev said.

Automation also puts computational chemists in the driver’s seat in the quest for scientific discovery. In one example using AI, active learning and automation, Isayev’s team and researchers at the UNC Chapel Hill found the ideal candidates for an MRI imaging agent by testing less than 400 polymers in about two weeks of lab time. In a traditional lab setting, scientists would have had a nearly impossible task of creating and testing 50,000 monomer compositions.