UBC researchers train computers to predict upcoming designer drugs

Science, health and technology

Global law enforcement is already using the new method

UBC researchers have trained computers to predict the next synthetic drugs before they even hit the market, a technology that could save lives.

Law enforcement is in a race to identify and regulate new versions of dangerous psychoactive drugs such as bath salts and synthetic opioids, even as underground chemists scramble to synthesize and distribute new molecules that have the same psychoactive effects as conventional drugs of abuse.

Dr Michael Skinnider

Identifying these so-called “legal plans” in the pills or powders seized can take months, during which time thousands of people may have already used a new designer drug.

But new research is already helping law enforcement around the world reduce the identification time from months to days, crucial in the race to identify and regulate new versions of dangerous psychoactive drugs.

“The vast majority of these synthetic drugs have never been tested in humans and are completely unregulated. They are a major public health concern for emergency services around the world, ”says Dr. Michael Skinnider, a medical student at UBC, who completed the research as a doctoral student at UBC. University of British Columbia. Michael Smith Laboratories.

A minority report for new synthetic drugs

Dr Skinnider and his colleagues used a database of known psychoactive substances provided by forensic laboratories around the world to train an artificial intelligence algorithm on the structures of these drugs. The algorithm they used, known as the Deep Neural Network, is inspired by the structure and function of the human brain.

Based on this training, the model then generated around 8.9 million potential designer drugs.

These molecules were then tested against 196 new synthetic drugs which appeared on the illicit market after the formation of the model. The researchers found that more than 90 percent were present in the set generated.

In other words, the model was able to predict almost all new drugs discovered since its formation.

“The fact that we can predict which designer drugs are likely to emerge on the market before they actually appear is kind of like the science fiction movie from 2002, Minority report, where prior knowledge of criminal activity about to take place has significantly reduced crime in a future world, ”says lead author Dr David Wishart (he him), professor of computer science at the University of Alberta.

“Essentially, our software gives law enforcement agencies and public health programs a head start on underground chemists and lets them know what to look for. “

Identification in days instead of months

That still left the problem of how to easily identify completely unknown substances.

Dr Leonard Foster

Dr Leonard Foster

The researchers found that the model also learned which molecules were most likely to appear on the market and which were less likely.

“We wondered if we could use this probability to determine what an unknown drug is, based solely on its mass, which is easy for a chemist to measure for any pill or powder using the mass spectrometry ”, explains Dr Leonard Foster (he him), professor in the Department of Biochemistry at UBC and internationally recognized expert in mass spectrometry.

The researchers tested this hypothesis using each of the 196 new designer drugs. Using only mass, the researchers found that their model ranked the correct chemical structure of an unidentified synthetic drug among the top 10 candidates 72 percent of the time. The integration of tandem mass spectrometry data, another easily obtainable measurement, improved this figure to 86%. When it was a single guess, the model could predict the correct structure 51% of the time.

“It shocked us that the model works so well, because elucidating entire chemical structures from a simple, precise mass measurement is generally considered an intractable problem. And reducing a list of billions of structures to a set of 10 candidates could dramatically speed up the rate at which new synthetic drugs can be identified by chemists, ”notes Dr. Skinnider.

The same type of model could be used to discover all kinds of new molecules, adds Dr Skinnider, from identifying new performance-enhancing drugs for sports doping to identifying previously unknown molecules in blood and urine. humans.

“There’s a whole world of chemical ‘dark matter’ just beyond our fingers right now. I think there is a huge opportunity for the right AI tools to bring this unknown chemical world to light, ”said Dr Skinnider.

Securely distributed by the Novel Psychoactive Substance Data Hub, the UBC model is used by the United States Drug Enforcement Agency, United Nations Office on Drugs and Crime, European Monitoring Center for Drugs and Drug Addiction and the Federal Criminal Police Office of Germany.

The study ‘Deep generative model enables automated elucidation of the structure of new psychoactive substances‘was published today in Nature Machine Intelligence.

Interview language (s): English (Skinnider, Wishart, Foster)

Comments are closed.