Skip to Main content Skip to Navigation

Pea protein based foods for the future : From chemical composition to statistical models to understand the mechanisms behind perceptions of beany, bitterness and astringency

Abstract : With population growth and the coming climate crisis, how can we offer a diet with higher levels of plant proteins? The use of pulses such as peas in foods would contribute to the evolution of this offer. However, their bitterness, their "beany" aromatic notes and their persistence are obstacles to their use. The objective of this PhD was to understand the role of their composition (volatile and non-volatile compounds) on perceptions. The role of food formulation was also studied on their physicochemical properties and perceptions during the oral process. For this, an original and multidisciplinary strategy combining sensory analysis, physico-chemistry and statistics was set up. A mixing plan was built to create a range of 23 solutions formulated from 6 fractions derived from commercial pea protein isolates. A profile was used to characterize the perceptions of these solutions In parallel, their compositions in volatile compounds, peptides, phytochemical compounds (phenolic acids, flavonoids and terpenoids), carbohydrates and minerals were characterized. The links between sensory and chemical profiles were analysed using Pearson correlations, linear (Partial-Least-Square) and non-linear (Artificial Neural Network) models. From these pea solutions, 79 volatile compounds were identified, including 34 correlated to beany notes; 3500 peptides including 14 correlated to bitterness; and 54 phytochemical compounds, including 29 correlated to bitterness and astringency. Then, multi-block PLS models were carried out to identify the relative role of these blocks of chemical compounds on beany, bitter and astringent perceptions. The second part of the work focused on understanding how the formulation of a pea-based beverages can modulate perceptions. A wide range of plant beverages was then developed by varying the type of proteins and the fat, texturizing and salt contents. The temporal sensory properties were evaluated when consuming a spoon, but also a whole portion of food; rheological properties and volatile compound profiles were determined. The results showed that the beany and bitterness notes depend mainly on the type of protein and gellan content. Astringency, which is highly persistent, can be limited by the formulation. In addition, rheological properties allow a reliable prediction of astringency perceptions. Finally, volatile molecule profiles, strongly affected by the presence of saliva, have shown the key role of the oral process on these perceptions. Beyond the original methodological approaches and new knowledge on products, this work opens up avenues to revisit pea fractionation processes and help the formulation of plant protein in foods, in order to reduce their off-notes.
Document type :
Complete list of metadata
Contributor : ABES STAR :  Contact
Submitted on : Tuesday, January 25, 2022 - 5:38:31 PM
Last modification on : Friday, August 5, 2022 - 2:41:13 PM
Long-term archiving on: : Tuesday, April 26, 2022 - 7:17:51 PM


Version validated by the jury (STAR)


  • HAL Id : tel-03543099, version 1


Audrey Cosson. Pea protein based foods for the future : From chemical composition to statistical models to understand the mechanisms behind perceptions of beany, bitterness and astringency. Food and Nutrition. Université Paris-Saclay, 2021. English. ⟨NNT : 2021UPASB012⟩. ⟨tel-03543099⟩



Record views


Files downloads