Scent plays a central role in the lives of living things, from locating food and mates to detecting and avoiding danger. Insects use different types of scents such as sex pheromones, trace pheromones, alarm pheromones, aggregation pheromones, and plant odor to locate host plants. Because insect populations are rapidly declining, negatively impacting ecosystem function, it is important to know what types of molecules interact with insect olfactory receptors. In this way, the newly developed chemicals do not interfere with the insect's chemical communication signals and cues.
To date, much is still unknown about what properties of scent chemicals cause their interaction with insect olfactory receptors. Researchers at MMD TechHub therefore aim to create the first large-scale database of insect odors known to bind to insect receptors and use them in machine learning models that can identify patterns in these odors.
The project is a collaboration between evolutionary biologist Astrid Groot, mathematician Joe Ellis Monaghan, data scientist Patrick Foret, chemist Sael Samanipour, and postdoctoral researcher Thanet Pitakbat.
binding to receptors
Identifying insect scent patterns – artistic impressions. Image credit: Alius Noreika / AI
The researchers currently have a database of about 25,000 data points. The data includes different types of scents, such as pheromones and communication chemicals.
For each scent molecule, the database includes both the chemical structure and the receptor protein to which it binds. However, the exact binding location on the receptor is largely unknown. Scientists aim to link chemicals to specific regions of receptor proteins and identify binding patterns.
Comprehensive database development
There is much scattered and unpublished data regarding insect odor chemicals and olfactory receptors. To build the most comprehensive database, the researchers established an international consortium of insect olfactory scientists willing to share all available data.
This consortium therefore fosters collaboration and knowledge sharing. The consortium is currently establishing standard protocols for data collection, a key aspect of the project. Pitakbut: “Ultimately, people benefit by sharing their data. This way we can use AI technologies together and move forward together.”
insect experiment
Although the MMD project is currently focused on data collection and machine learning development, the project's ultimate goal is to develop a pipeline for the safe design of scent molecules. If a machine learning model can find binding patterns, it can be used to predict whether newly designed chemicals are likely to bind to insect receptors.
The accuracy of these predictions can be tested by making so-called electroantennary recordings. When a chemical is sprayed onto an insect's antennae (the insect's nose), this chemical binds to one or more olfactory receptors present on the antenna, causing a change in electrical potential within the antenna. These recordings are relatively simple and provide researchers with a unique opportunity to biologically test the model's predictions.
Source: University of Amsterdam
