Skip to main content
Premium Trial:

Request an Annual Quote

Familiar Tool, New Use: Researchers Turn AutoDock on its Head to Predict Side Effects


Taking a fresh look at the capabilities of AutoDock, which screens structural models of small-molecule compounds against the structures of receptor molecules, Elcock and MD/PhD student Bill Rockey have successfully used it to identify unintended drug interactions that might cause side effects.

The researchers simply reversed the way a typical AutoDock screening study is conducted: Rather than screen thousands of small molecules against a single receptor to find promising drugs, they started with a single small molecule, and screened thousands of receptor molecules against it. This identified which of those receptors the drug might also bind to, which would result in unintended side effects.

“It’s standard methodology that’s been out there for a while,” said Elcock. “People thought it was sort of too ambitious to apply it the way we have… People often feel that homology-based models are not accurate enough to do drug docking work.”

A paper on the work, which examined several kinase inhibitors, including Gleevec, was published in a recent issue of Proteins: Structure, Function and Genetics. The team compared its computational results with experimental data for the drugs, and was able to predict with a high degree of confidence which kinases would be inhibited by the drugs and which ones wouldn’t. “We just sort of flipped the thing on its head, really,” summed up Elcock.

Large protein families like kinases are ideal for this approach, Elcock said, because of the high degree of homology that exists within them. The method is only able to identify receptors that have a high degree of sequence similarity to the original receptor, so side effects caused by binding a totally unrelated structure cannot be detected, he noted.

Elcock said that it’s also crucial to begin with a high-quality x-ray crystal structure of a drug bound to a receptor protein. “If you’ve got a crystal structure of your protein bound to a drug, then you can make models of the other proteins with a high degree of confidence,” he said. “If you don’t have a good structure to start with — for example, if there’s no drug bound to it — then it’s probably not going to be a good model for building other receptors.”

This work grew out of Elcock’s primary research in simulating protein-protein interactions. “This was our first foray into the [drug docking] area,” he said. Recognizing that the work has obvious benefits in the drug arena, he added, “We’re hoping that maybe some of the drug companies will sit up and take note.”

Future plans include refining the approach, scaling it up, and adapting it to work with a lower quality of crystal structure than it currently requires. The researchers will also use AutoDock to study metabolite interactions with receptors in order to reconstruct metabolic pathways and, hopefully, identify key regulatory elements. “That’s a long shot,” Elcock noted. “I don’t know if the methods are going to be good enough for that, but we’ll find out.”

— BT

Filed under

The Scan

NFTs for Genome Sharing

Nature News writes that non-fungible tokens could be a way for people to profit from sharing genomic data.

Wastewater Warning System

Time magazine writes that cities and college campuses are monitoring sewage for SARS-CoV-2, an approach officials hope lasts beyond COVID-19.

Networks to Boost Surveillance

Scientific American writes that new organizations and networks aim to improve the ability of developing countries to conduct SARS-CoV-2 genomic surveillance.

Genome Biology Papers on Gastric Cancer Epimutations, BUTTERFLY, GUNC Tool

In Genome Biology this week: recurrent epigenetic mutations in gastric cancer, correction tool for unique molecular identifier-based assays, and more.