A sweeter future? AI helps scientists discover new sugar substitutes

Researchers from the Hebrew University of Jerusalem have created a computational framework that identified 92 promising compounds for future use as sugar substitutes from over 50 million chemical compounds. The study, published in the journal Food Chemistry, explores how artificial intelligence and molecular modeling could expedite the discovery of new sweeteners, addressing the challenge of unwanted bitter aftertastes in existing options.

As demand for low-calorie alternatives rises, scientists seek sweeteners with pleasant taste, safety, and stability profiles. Excessive sugar intake is linked to health issues such as obesity, type 2 diabetes, fatty liver disease, and cardiovascular disorders, and the FDA has set daily limits for artificial sweeteners like aspartame, saccharin, and sucralose.

However, these sweeteners have limitations, including undesirable tastes. To tackle this, the researchers employed a machine-learning algorithm called Iterative Stochastic Elimination (ISE) to screen and score millions of compounds virtually, significantly reducing the number of compounds for testing. By building models based on molecular properties, the ISE method predicted whether a compound would be sweet or bitter, enabling researchers to distinguish desirable from undesirable taste characteristics.

The study combined machine learning with molecular docking to examine how candidate compounds interact with the human sweet-taste receptor, composed of T1R2 and T1R3 protein subunits. The detailed structural knowledge of this receptor, published recently, enhanced the docking process. After applying taste prediction, molecular docking, and chemical structure filters, the researchers narrowed the initial pool to 556 potential sweeteners, narrowing it further to 92 candidates with favorable predicted properties.

Many of these molecules have unique chemical structures, suggesting opportunities for entirely new classes of sweet-tasting compounds. The research also identified some previously patented sweet-receptor modulators, supporting the effectiveness of the computational approach. While the identified compounds are still predictions requiring laboratory validation, the study establishes a promising framework for discovering the next generation of sugar alternatives.