AI identifies promising plant proteins for sustainable food ingredients

New research by scientists at the University of Leeds is using artificial intelligence (AI) and statistical physics to identify plant proteins that could potentially replace animal-derived emulsifiers, according to the University of Leeds website. The research has identified nearly 800 plant proteins with potential emulsifying properties, many of which had not previously been considered for this purpose.

Researchers from the University of Leeds’ School of Food Science and Nutrition have developed a computational approach that can rapidly screen plant proteins for their potential to act as emulsifiers. These functional ingredients are widely used in food, cosmetics, pharmaceuticals and other products.

The work could help address a major challenge in developing more sustainable food ingredients: identifying promising proteins from the millions of possibilities available without having to test each one individually in the laboratory.

Why new emulsifiers are needed

Emulsifiers help oil and water mix and remain stable. They are widely used in everyday products such as sauces, ice cream and mayonnaise, as well as in cosmetics, pharmaceuticals and industrial applications.

There is growing interest in natural and more sustainable alternatives to existing emulsifiers, including animal-derived proteins such as caseins and whey proteins.

However, millions of plant proteins could potentially possess useful functional properties. Identifying suitable candidates through conventional laboratory testing can be expensive and time-consuming, often involving extensive trial and error.

The Leeds research team therefore set out to develop a faster and more reliable way of predicting which plant proteins could function effectively as emulsifiers.

Combining AI with statistical physics

The research was led by Dr Simha Sridharan and supervised by Professor Anwesha Sarkar, both from the University of Leeds’ School of Food Science and Nutrition and the Sarkar Lab. The team also collaborated with Dr Rik Sarkar, a machine-learning expert at the University of Edinburgh.

According to the University of Leeds, the researchers first used a simulation model based on statistical physics to understand how proteins interact with oil-water interfaces. This interaction is critical because proteins must attach to the interface between oil and water and help stabilise the resulting mixture to function as emulsifiers.

The team then applied machine learning to identify specific sections and characteristics of proteins associated with this behaviour.

By combining statistical physics with machine learning, the researchers were able to predict which plant proteins were most likely to demonstrate emulsification properties similar to those of animal proteins.

The computational approach could allow researchers to narrow the enormous pool of potential plant proteins to the most promising candidates for laboratory testing, potentially reducing the time and resources required for conventional trial-and-error approaches.

Nearly 800 plant proteins identified

The model identified nearly 800 plant proteins with potential to act as emulsifiers. Importantly, many had not previously been considered for this application.

The researchers subsequently tested several commercially available proteins to determine whether experimental results matched the model’s predictions. Pea and potato proteins demonstrated effective emulsification properties, supporting the predictions generated by the AI-driven approach.

Professor Anwesha Sarkar, NAPIC Co-Director at the University of Leeds, said the model had identified nearly 800 plant proteins that could potentially act as emulsifiers, including many that had never previously been considered for the purpose.

The findings demonstrate how AI-based approaches could help researchers identify promising ingredients considerably faster than traditional experimental methods, the University of Leeds said.

Potential applications for the food industry

The approach could be particularly useful for companies developing plant-based and sustainable food products by providing a way to search a much larger pool of potential protein sources for functional ingredients.

Rather than relying exclusively on laboratory testing of proteins one at a time, computational screening could help researchers identify the strongest candidates before experimental validation.

This could accelerate the discovery and development of new ingredients while reducing the time and resources required during early-stage research.

The research also highlights the potential of combining food science, protein chemistry, statistical physics and artificial intelligence to address challenges associated with the development of more sustainable food systems.

For NAPIC, the research represents an example of interdisciplinary work aimed at developing the next generation of alternative protein technologies and ingredients.

The research paper, ‘Data-driven pipeline enables discovery of plant protein surfactants’, has been published in Communications Chemistry.