P683 - PLANT&IA : A NEW SOFTWARE TO OPTIMIZE PLANT-BASED PROTEINS FOR ELDERLY PEOPLE

P683

PLANT&IA : A NEW SOFTWARE TO OPTIMIZE PLANT-BASED PROTEINS FOR ELDERLY PEOPLE

B. Araujo^1,2,*, C. Breuillard¹, G. Cougoulat², C. Moinard¹

¹Biology, Laboratoire de Bioénergétique Fondamentale et Appliquée (LBFA), ²Computer Science, Grenoble Alpes Recherche - Infrastructure de Calcul Intensif et de Données (GRICAD), Grenoble, France

Rationale: Current recommendations promote a plant-based diet, but plant proteins present nutritional limitations, including low essential amino acid (AA) content and reduced digestibility, which are particularly concerning for elderly with increased protein requirements. This project aims to optimize plant protein blends using programming to overcome these challenges.

Methods: A database of 506 plant-based products was developed from scientific literature, including their AA composition, ileal or fecal digestibility (in humans, pigs, and rats), and classification by format and botanical category. In our model, whey protein was selected as reference. An algorithm was developed to:

1.   Calculate the "real protein contribution" by multiplying AA composition by digestibility.

2.   Generate and filter blends of plant products that meet two constraints: a maximum weight of 90g and an AA composition at least equal to that of 30g of whey protein. Optimal blends were identified by calculating the root mean squared error (RMSE): the lower the RMSE, the closer the AA compositions of the reference and the blend.

Results: During the tests, an optimal mixture of 1 cereal and 1 legume was found with a deviation (RMSE) of 3.42 from the reference. By using two legumes, the RMSE was reduced to 2.48, and another mix with 10 products achieved an even lower RMSE of 2.06. 

Conclusion: This project confirms that plant proteins, when strategically combined, can achieve an AA bioavailibility close to that of animal references. The use of our software addresses the increased nutritional requirements of specific populations, such as older adults. In the future, the software will evolve to include other nutritional, environmental and economic factors in order to obtain plant protein mixtures that are better suited to today's challenges and populations.

Disclosure of Interest: None declared