Researchers at Oregon State University College of Engineering have harnessed the power of artificial intelligence to help protect bees from pesticides.
Bee populations are regularly destroyed by pesticides used in agriculture. Image credit: Anivesh Agrawal via WikimediaCC BY-SA 4.0
Cory Simonassistant professor of chemical engineering, and Xiaoli Fernassociate professor of computer science, leading the project, involves training a machine learning model to predict whether any of the proposed new herbicides, fungicides, or pesticides will be toxic. for honey bees based on the molecular structure of the compound.
The findings, featured on spread of the Journal of Physics and Chemistry In some special cases, “Chemical Design by Artificial Intelligence” is essential because many fruit, nut, vegetable and seed plants rely on bee pollination.
Without bees to transfer the pollen needed for reproduction, nearly 100 commercial crops in the United States would disappear. The annual global economic impact of bees is estimated to exceed $100 billion.
“Pesticides are widely used in agriculture, helping to increase crop yields and provide food security, but pesticides can harm untargeted species such as bees,” says Simon. “And because insects, weeds, etc eventually develop resistance, new insecticides have to be constantly developed, ones that don’t harm bees.”
Graduate students Ping Yang and Adrian Henle used honey bee toxicity data from pesticide exposure experiments, involving nearly 400 different pesticide molecules, to train an algorithm Mathematical calculations predict whether a new pesticide molecule will be toxic to honey bees.
“The model represents pesticide molecules by a set of random walks on their molecular graph,” says Yang.
Random walk is a mathematical concept that describes any meandering path, such as on the complex chemical structure of pesticides, where each step along the path is decided by chance. , as if by tossing a coin.
Imagine, explains Yang, that you are walking aimlessly along the chemical structure of a pesticide, going from atom to atom through the bonds that hold the compound together. . You move in random directions but keep track of your route, the sequence of atoms, and the bonds you visit. Then you come out to another molecule, compare the helix and switch with what you did before.
“The algorithm declares two molecules the same if they share multiple steps with the same sequence of atoms and bonds,” Yang said. “Our model serves as a proxy for an experiment in bee toxicity and can be used to rapidly screen proposed pesticide molecules for their toxicity.”
Source: Oregon State University
