New AI Model Can Identify Designer, Research Drugs on the Fly

Identifying suspected drugs in the field has long-since been an issue for law enforcement to keep pace with. At a certain point, there are just far too many different psychoactive chemicals to keep track of and the number grows larger every day. However, a new AI language model just won an award for identifying unknown and untested compounds with surprising accuracy.
AI
Shutterstock

A prestigious science award has been given to a man who created a new generative AI language model capable of identifying the exact chemical structure of designer drugs, even drugs that have not been tested on humans yet. 

The winning entry for the 2023 NOMIS & Science Young Explorer competition was a new AI language model trained by Princeton biologist Michael Skinnider. This new AI model can reportedly identify the chemical structure of research chemicals or “legal highs,” terms which refer to any number of chemical compounds which produce psychoactive effects but have not yet been scheduled by the FDA. The speed at which these compounds can be synthesized has created a legislative loophole where chemists and labs can more or less sell and ship dangerous drugs to people with a label on it that says “not for human consumption” without fear of legal repercussions. 

As a result of this loophole, law enforcement professionals are often faced with a situation wherein they suspect someone is carrying dangerous drugs but can’t prove it, or someone is experiencing adverse effects from a drug that they are unable to identify and thus, unable to properly treat. Traditional field testing kits can’t identify the drug because traditional field testing kits only look for the most commonly used psychoactive compounds (heroin meth, cocaine, the usual suspects), whereas Skinnider’s AI model is reportedly capable of generating and identifying entirely new chemical structures on the fly. 

“Conventional drugs of abuse such as cocaine or methamphetamine still dominate the market, but enterprising chemists have realized that with slight chemical modifications to these drugs, they can create new derivatives that are completely legal,” Skinnider said in an essay on Science. “And because these synthetic drugs have never been tested on humans, they can have unpredictable and damaging side effects.”

Skinnider began his work by first training an AI language model on a simplified molecular input line entry system, otherwise known as SMILES, which is a fancy way of saying he taught the language model a new language which is used to represent different complex chemical structures in a simple text-based format. 

Skinnider taught the AI model a way of then identifying chemicals using a process called “mass spectrometry” which, god help me I’m probably butchering this, but according to Waters, mass spectrometry is a process of measuring the different ratios of electrical charges at the molecular level of whatever you want to test, drugs in this case, to determine the exact molecular weight of the particles in the sample. These molecular weights are used to identify and map the chemical compound. 

“As a MD/PhD student, I saw firsthand how patients could present with devastating symptoms of designer drug intoxications, but emergency physicians had few options to treat them. I wondered whether artificial intelligence could help,” Skinnider said. “Specifically, I asked whether AI could automatically elucidate the chemical structures of new designer drugs from mass spectrometry data. Scientifically, this was a tall order.”

Skinnider then used information from existing research about commonly used designer drugs to further educate the AI model, using 1,753 known examples of such. What he found was the program could then generate examples of entirely new chemical structures that might have similar effects. Not only that, but it could also be used to predict what undiscovered chemicals are most likely to become popular in the future based on what drug users have responded well to in the past. 

Skinnider intimated that these advancements in drug identification technology have very practical real world applications in identifying and responding to drug crises. He also said his technology has already been used to identify new and dangerous psychoactive compounds. 

“I have now applied this technology to tens of thousands of patient samples and used it to discover several new designer drugs, such as a new analog of fentanyl that emerged last year. Currently, I am working with the British Columbia Centre for Disease Control to implement this AI technology in routine clinical practice to automatically discover new drugs as soon as they are introduced into the population,” Skinnider said. “Ultimately, my dream is that first responders, emergency physicians, and public health officials will all be able to take advantage of generative AI to make more informed decisions when treating patients and managing outbreaks.

Total
0
Shares
1 comment
Leave a Reply

Your email address will not be published. Required fields are marked *

Related Posts
Abstrax
Read More

Brand Spotlight: Abstrax

Enter the cannaquantum realm. Abstrax reaches beyond the world of terpenes to discover new, exotic flavor compounds.
Total
0
Share