Crime-fighting algorithm to take up the battle against illegal drugs?
Researchers from the University of British Columbia (UBC), Canada, have trained a computer to predict designer drugs based on specific common molecules, even before the drugs hit the market.
Clandestine chemists are constantly manufacturing new and dangerous psychoactive drugs that law enforcement agencies struggle to keep up with. Many of these designer drugs can lead to irreparable mental damage and/or even death.
“The vast majority of these designer drugs have never been tested in humans and are completely unregulated,” says author Dr Michael Skinnider. “They are a major public health concern to emergency departments across the world.”
The algorithm behind drug forensicsThe algorithm used by the computer, called deep neural network, generated 8.9 million potential designer drugs that could be identified from a unique molecular make-up if they popped up in society.
The researchers then compared this data set to newly emerging designer drugs and found that 90% of the 196 new drugs were in the predicted data set.
“The fact that we can predict what designer drugs are likely to emerge on the market before they actually appear is a bit like the 2002 sci-fi movie, Minority Report, where foreknowledge about criminal activities about to take place helped significantly reduce crime in a future world,” explains senior author Dr David Wishart from the University of Alberta, Canada.
“Essentially, our software gives law enforcement agencies and public health programs a head start on the clandestine chemists, and lets them know what to be on the lookout for.”
With this level of prediction, forensic scanning of drugs can be cut from months to days.
The algorithm also learned which molecules were more and less likely to appear.
“We wondered whether 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 mass spectrometry,” says UBC’s Dr Leonard Foster, an internationally recognised expert on mass spectrometry.
Using only mass, the algorithm was able to correctly identify the molecular structure of an unknown drug in a single guess around 50% of the time, but the accuracy increased to 86% as more measurements were considered.
“It was shocking to us that the model performed this well, because elucidating entire chemical structures from just an accurate mass measurement is generally thought to be an unsolvable problem,” says Skinnider. “And narrowing down a list of billions of structures to a set of 10 candidates could massively accelerate the pace at which new designer drugs can be identified by chemists.”
The researchers say this AI could also help identify other new molecules, such as in sports doping or novel molecules in the blood and urine.
“There is an entire world of chemical ‘dark matter’ just beyond our fingertips right now,” says Skinnider. “I think there is a huge opportunity for the right AI tools to shine a light on this unknown chemical world.”
Image credits: Getty Images