The latest international research: machine learning methods can help identify man-made drugs

China News Service, Beijing, November 16 (Reporter Sun Zifa) Springer Nature's professional academic journal "Nature-Machine Intelligence" published a computational biology research paper saying that researchers have discovered an automated, generative machine The learning method can determine the chemical structure of unknown new psychotropic drugs (also known as artificial drugs) using mass spectrometry alone.
Understanding these structures can help forensic laboratories identify suspected man-made drugs more quickly
.

The paper pointed out that a large number of new psychotropic drugs appear on the illegal market every year.
These substances will cause similar psychoactive effects to known illegal drugs; however, because their synthesis method makes them chemically different, these drugs circumvent the existing The drug laws and regulations are even difficult to detect
.

In response to the above problems, the corresponding author of the paper, Michael Skinnider of the University of British Columbia, Canada, and colleagues used confidential data from forensic laboratories around the world to train a machine learning model to produce a structure and nature similar to recent artificial drugs.
Molecule
.

They tested the model with the newly collected data after the model training, and found that this method can identify unknown man-made drugs using mass spectrometry alone
.

The author of the paper concluded that similar generation methods trained with other data sets can also help identify the structure of unknown molecules in other specific fields, such as identifying new stimulants or environmental pollutants
.