We’ve all heard the commercials: a drug promises amazing results for treating a disease, and then the remainder of the commercial is filled with a mind-numbingly long list of potential side effects. Side effects plague prescription drugs, sometimes prompting drug approval agencies to reject the drug or making patients wonder if the cure can be worse than the disease. Now, researchers from Stanford University have leveraged the Summit supercomputer at Oak Ridge National Laboratory (ORNL) to work on redesigning those drugs without the side effects.

“What if we could redesign drugs to keep their benefits while eliminating their unwanted side effects?” said Ron Dror, the associate professor of computer science at Stanford University whose lab is leading the research, in an interview with Stanford University’s Tom Abate. Dror’s lab targeted drugs that work with G protein-coupled receptors (GPCRs), proteins that are found in all human cells and which serve as the attachment points for a wide range of drugs – everything from psychedelics like LSD to blood pressure medications.

Drugs that attach to GCPRs cause multiple simultaneous reactions in the protein, which is responsible for many side effects. ORNL is no stranger to GCPRs: just a few months ago, they highlighted research exploring applications of machine learning in drug design for GCPRs. Dror’s lab, on the other hand, set out to use supercomputing power to simulate a GCPR attached to a series of different molecules, aiming to pin down how each molecule changed the ordering of the GCPR’s atoms.

To run these detailed simulations, the researchers turned to ORNL’s Summit supercomputer, which is currently the most powerful publicly ranked supercomputing in the world according to the most recent Top500 list. Summit’s 4,608 nodes (each powered by two IBM Power9 CPUs and six Nvidia Volta GPUs) deliver 148 Linpack petaflops.

The results were promising, and based on what they found, the researchers designed a set of molecules that – in the simulations, at least – produced the desired atomic reordering without also inducing the atomic shifts that produce undesirable side effects. While there is a long road between this research and any drug that may be approved for human consumption, the results are a promising milestone in the path toward a new era of drug design.

“In addition to revealing how a drug molecule could cause a GPCR to trigger only beneficial effects, we’ve used these findings to design molecules with desired physiological properties, which is something that many labs have been trying to do for a long time,” Dror said. “Armed with our results, researchers can begin to imagine new and better ways to design drugs that retain their effectiveness while posing fewer dangers.”

The research discussed in this article was published as “Molecular Mechanism of Biased Signaling in a Prototypical G protein–Coupled Receptor” in the February 2020 issue of Science. The article, which is accessible here, was written by Carl-Mikael Suomivuori, Naomi R. Latorraca, Laura M. Wingler, Stephan Eismann, Matthew C. King, Alissa L. W. Kleinhenz, Meredith A. Skiba, Dean P. Staus, Andrew C. Kruse, Robert J. Lefkowitz and Ron O. Dror.