As part of the HealthFerm project, which aims to explore microbial diversity for improving plant-based food fermentations, a total of 102 high-quality microbial genomes were successfully sequenced, annotated, and submitted to the European Nucleotide Archive (ENA) of the European Bioinformatics Institute (EBI; Hinxton, United Kingdom). These microbial genomes include 81 bacterial and 21 yeast strains isolated from diverse fermented foods and countries in Europe. Of the bacterial genomes, 67 are fully circularised, indicating complete assemblies that include chromosomes and putative plasmids, while 14 remain non-circular but are of high quality. The yeast genomes were assembled at the contig level, in line with the complexity of their eukaryotic genome structures.

All strains sequenced were selected based on their (assumed) functional relevance to key activities such as the production of short-chain fatty acids (SCFA), exopolysaccharides (EPS), and vitamin B12, as well as the degradation of fibres, proteins and antinutritional factors. This genomic resource forms a robust foundation for understanding the genetic potential of microorganisms relevant to food fermentation at the genetic level. Furthermore, these genomes will support comparative genomics, genome mining, and genome-to-phenotype association studies in subsequent project stages, integrating these predictive genomic insights with laboratory validation to develop healthier and more sustainable fermented foods.

The overall objective of work package 2 of the HealthFerm project is to understand the impact of fermentation on grain-based raw materials and foods for developing novel foods optimized for health benefits without compromising flavour, sustainability and safety. An important requirement for achieving this objective is the availability of a broadly applicable analytical toolbox that allows in-depth characterization of changes in the chemical structure of the many constituents that occur in grain-based raw materials and foods.

The aim of Task 2.1 within this work package is to establish such an analytical toolbox, which all involved partners can make use of and which can facilitate interactions and collaboration within the consortium. Under the supervision of KU Leuven, the lead partner for this task and associated deliverable, several steps were taken within the first project year (1 September 2022 – 31 August 2023) leading to the successful development of the mentioned analytical toolbox. First, inventories were made on the required raw materials as well as the target (model) food products for all partners within worl package 2 and the analytical capabilities of all partners.

Next, the scope of the analytical toolbox was defined and discussed in several meetings involving all work package 2 partners. A decision was made to establish the analytical toolbox in the form of a publishable ‘Book of Methods’, focusing on chemical characterisation of constituents present in the raw materials used in HealthFerm and changes therein as the result of fermentation processes. Then, the different partners of work package 2 were made responsible for specific chapters of the Book of Methods, and discussions among partners with compatible analytical capabilities were set up, after which draft texts for the different chapters were prepared. This has led to the compilation of a first draft of a complete ‘Book of Methods’ for work package 2 of the HealthFerm project, attached to this document and available to all consortium members, representing the successful completion of Deliverable 2.1. In follow-up steps, internal peer review and final formatting will be undertaken, with the final goal of achieving an open access publication of the ‘Book of Methods’.

This deliverable presents predictive models that link fermentation conditions to health-related, sensorial, and techno-functional properties of cereals and pulses and derived food products. Faba bean and oat were selected as representative plant-based ingredients due to their nutritional potential, but also their technological challenges, including beany off-flavours, low solubility, and the presence of anti-nutritional factors. To address these limitations, a series of liquid, semi-solid, and solid food systems were designed, including cereal-pulse beverages, oat-based emulsions, oat-faba gurts, tempeh-like products, extruded meat alternatives, and wheat breads enriched with sourdough from faba bean flour.

Across these models, fermentation conditions and processing parameters (e.g. microbial composition, inoculum ratio, time, temperature, pre-treatment, drying method) were varied. The resulting products were characterised for compositional, nutritional, and sensory properties, and the data were integrated using advanced statistical approaches such as principal component analysis (PCA), partial least squares (PLS) regression, principal component regression (PCR), and response surface methodology (RSM). These analyses provided insights into how fermentation parameters shape food quality across different matrices.

The models consistently demonstrated that fermentation strongly modifies product attributes. In faba bean ingredients, metabolomic profiling revealed links between specific compounds and sensory drivers of bitterness, astringency, and sweetness. In beverages, microbial consortia were key determinants of sensory clustering and consumer liking. In semi-solid gurts, RSM and PLS showed how inoculum ratio, time, and temperature drove acidity and viscosity development, which in turn influenced sensory perception. Tempeh-like products displayed enhanced proteolysis and nutrient accessibility during fermentation, while PLS modelling linked free amino acids, volatile compounds, and texture attributes to desirable sensory profiles. Extruded meat analogues highlighted the combined effect of fermentation and drying on flavour reduction (beany, bitter) and enhancement (umami, roasted). In breads, PCA revealed the contribution of sourdough fermentation to changes in protein and starch digestibility, with effects dependent on microbial composition. Finally, a joint PLS model integrating multiple food systems demonstrated that protein hydrolysis during digestion was most strongly associated with consortia-based fermentations, underlining the possible role of multi-strain systems in enhancing nutritional quality.

Together, these models establish clear relationships between fermentation conditions and the resulting nutritional, sensory, and techno-functional outcomes. The findings provide a framework for the rational design of plant-based foods, enabling more targeted use of fermentation to deliver products that are not only sustainable but also nutritious and appealing to consumers.

Short-chain fatty acids (SCFA) are microbial-produced metabolites that potentially mediate various metabolic health benefits. They are predominantly produced in the human colon through fibre fermentation and absorbed by colonocytes. In contrast, fermented foods contain SCFA, which are absorbed in the small intestine. Unlike colonocytes, small intestinal epithelial cells do not utilise SCFA as an energy source. Therefore, we hypothesised that a greater proportion of SCFA would reach the systemic circulation following small intestinal SCFA administration. In contrast, SCFA stimulate enteroendocrine L-cells in the gut, triggering the release of glucagon-like peptide 1 (GLP-1) and peptide YY (PYY), which regulate appetite and glucose homeostasis. As L-cells are more abundant in the colon, we hypothesised that gut hormone release would be more pronounced following colonic SCFA administration. These hypotheses were evaluated in two studies with healthy volunteers using targeted delivery capsules that administered SCFA specifically in the small intestine and colon. Systemic availability was evaluated in the first study using capsules filled with 13C-labeled SCFA. Gut hormone release was investigated in the second study using capsules containing unlabeled SCFA, with an additional placebo-controlled study visit to account for the hormone release induced by the meals consumed during the visit. Systemic availability was higher for propionate after small intestinal administration compared to colonic administration. However, no significant differences were observed for acetate and butyrate. GLP-1 and PYY release increased with SCFA administration in both the small intestine and the colon. However, only PYY release was significantly higher with colonic compared to small intestinal delivery. While the gut hormone release did not affect the glucose homeostasis, SCFA administration did impact appetite. Small intestinal SCFA administration reduced hunger, whereas colonic SCFA administration increased satiety and reduced desire-to-eat compared to placebo. In addition, small intestinal delivery increased fullness and satiety and reduced desire-to-eat compared to colonic SCFA delivery. Systemic SCFA concentrations were higher following small intestinal administration compared to colonic delivery. Although this data could not be used to calculate the systemic availability, this observation suggests that the hypothesis of the systemic availability may still be valid, also for acetate and butyrate.