ChemSweet: An AI-driven computational platform for next-gen sweetener discovery

Area of Science:

• Food Science
• Computational Chemistry
• Toxicology

Background:

• Growing concerns over artificial sweetener health risks necessitate safer alternatives.
• Current sweetener discovery methods are time-consuming and expensive.

Purpose of the Study:

• To introduce ChemSweet, an AI platform for rapid discovery of novel sweet molecules.
• To integrate safety, physicochemical properties, and sweetness prediction into the discovery process.

Main Methods:

• Developed machine learning models for physicochemical and toxicity properties.
• Integrated predictive models with the ChemSweet platform.
• Incorporated the SuperNatural database for rational screening.

Main Results:

• Identified 294 potential sweeteners meeting multiple criteria.
• ChemSweet evaluates biosafety, stability, and sweetness profiles.
• The platform ensures evaluation of six specific sweetener types.

Conclusions:

• ChemSweet offers a reduced-time, lower-cost approach to identifying safe and healthy sweeteners.
• The AI-based platform aids in the rational design and discovery of novel sweet compounds.