The research, led by Timothy J. Snelling and Liam A. Sinclair at Harper Adams University together with colleagues at the Alltech Bioscience Centre in Ireland, used a 4 × 4 Latin square design with four second-lactation Holstein–Friesian cows fitted with permanent rumen cannulae. Each cow weighed around 650 kilograms, was roughly 69 days in milk, and yielded about 40 kilograms of milk per day. The animals received a partial mixed ration containing a pelleted concentrate, grass silage, and maize silage at dry matter proportions of 0.297, 0.318, and 0.385 respectively. To mimic the effect of a cow selecting concentrates from a poorly mixed ration, an additional 4 kilograms of concentrate per cow per day was added either all at once at the morning feedout or split evenly across four deliveries at 0600, 1000, 1400, and 1700 hours.

The concentrates were prepared with or without live S. cerevisiae yeast, included at the manufacturer’s recommended rate of just 1 gram per cow per day, providing 1 × 10⁹ colony forming units per gram of product. This created four treatments: an uneven or even pattern of concentrate allocation, each with or without yeast. Each of the four experimental periods lasted 35 days, with 28 days for adaptation and sampling during the final 7 days. Rumen fluid was collected at six time points across the day, from before the morning feedout through to 2100 hours, allowing the researchers to track pH, volatile fatty acids, and ammonia nitrogen through the full diurnal cycle. The rumen microbial community was characterised using 16S rRNA gene amplicon sequencing, with libraries normalised to 47,000 reads per sample and classified against the SILVA 138 reference database.

Perhaps the most striking finding is what did not change. Neither the pattern of concentrate allocation nor the yeast supplement had any significant effect on dry matter intake, milk yield, milk fat or protein content, live weight, or apparent whole-tract digestibility of dry matter, organic matter, or neutral detergent fibre. Total dry matter intake ranged from 22.6 to 23.6 kilograms per day across treatments, and milk yield from 38.8 to 40.0 kilograms per day, with no statistically significant differences. Blood metabolites, including plasma glucose, β-hydroxybutyrate, and urea, were similarly unaffected, although plasma ammonia tended to be lower in yeast-supplemented cows.

Rumen pH told a more nuanced story. The pattern of concentrate allocation did not change mean rumen pH, but it did alter the diurnal fluctuation. There was a significant interaction between concentrate pattern and sampling time, with lower pH in cows on the uneven pattern during samples taken three to six hours after feedout. This reflects the rapid burst of rumen fermentation that follows the ingestion of a large single meal of rapidly fermentable concentrates, exactly the scenario a cow creates when it sorts concentrates from a poorly mixed ration. However, the mean minimum pH never fell below 5.6 at any sampling time, and the authors note that the ration was deliberately not formulated to induce acidotic conditions for ethical reasons. In susceptible animals or with higher concentrate proportions, this pattern could push ruminal pH below the 5.8 threshold for more than five to six hours per day, which is the diagnostic criterion for sub-acute ruminal acidosis.

The yeast supplement, by contrast, produced measurable shifts in rumen fermentation. Yeast supplementation significantly increased total volatile fatty acid concentration (p = 0.033) and propionate concentration (p = 0.016), with acetate and valerate tending to be higher (p = 0.076 and p = 0.091 respectively). Rumen pH also tended to increase with yeast (p = 0.088), and rumen ammonia nitrogen tended higher post-feeding (p = 0.100). Because both acetate and propionate rose proportionately, the acetate-to-propionate ratio remained unchanged. Rumen lactate concentrations were undetectable below 1 millimolar across all treatments, suggesting that the increased propionate reflected heightened microbial activity rather than a redirection away from lactate accumulation.

The microbiome data revealed the deepest effects. The phylum composition was consistent with a typical rumen community, dominated by Bacteroidota, Bacillota, and Pseudomonadota along with methanogenic Archaea. Alpha diversity, measured by Faith’s phylogenetic diversity and evenness, showed no differences between treatments. However, Bray–Curtis dissimilarity analysis detected significant microbiome clustering by yeast supplementation (p = 0.002) and by digesta phase (p < 0.001), with only a tendency for clustering by concentrate pattern (p = 0.061). Concentrate allocation had little overall effect on beta diversity, with a single Prevotellaceae feature reduced in relative abundance under the uneven pattern.

Linear discriminant analysis effect size, at FDR-corrected p < 0.05 and LDA scores above 2.0, identified three features associated with yeast supplementation. Two features, classified as unclassified Gammaproteobacteria and unclassified Prevotellaceae, decreased in abundance with yeast, while a third feature belonging to the Christensenellaceae R-7 group increased. The Christensenellaceae feature was distantly related to the type strain Gracilibacter thermotolerans at only 87 percent identity, meaning it represents a novel species yet to be isolated from the rumen. Members of Christensenellaceae play an important role in dietary fibre metabolism and have previously been shown to benefit from yeast-supplemented diets, consistent with the oxygen-scavenging mode of action proposed for S. cerevisiae, in which yeast cells consume oxygen in the rumen and create more favourable conditions for fibre-degrading bacteria.

The authors used two complementary statistical tools, ANCOM-BC and LEfSe, which differ in their statistical philosophy and provided different results, highlighting the difficulty of drawing firm conclusions from microbiome data. Many of the type strains identified by BLASTn searches showed low percentage identity to the feature sequences, limiting classification to approximately the genus or family level. This underscores how little is known about the species contained in the rumen microbiome, despite major efforts such as the Hungate1000 collection and the assembly of 913 microbial genomes from cow rumen metagenomes. The increased propionate concentration with yeast supplementation may have promoted the activity of Acidaminococcaceae species, which are abundant in the rumen and capable of metabolising succinate to propionate when succinate accumulates following ingestion of large quantities of rapidly fermentable concentrates.

The study has limitations that the authors acknowledge candidly. With only four animals, the statistical power was constrained, and the a priori power calculation set a β error probability of 0.60, lower than the preferred 0.80, increasing the probability of type II errors in the production and rumen metabolism data. The simulation of diet selection also differs from commercial conditions, where cows are fed as a group and some individuals consume substantially more concentrates than others. Nevertheless, the findings suggest that uneven concentrate delivery primarily shifts the diurnal pattern of rumen pH rather than its daily mean, while live yeast supplementation promotes microbial activity, increases propionate production, and reshapes the rumen microbiome in ways that favour fibre degradation. For dairy farmers, the message is clear: mixing the ration properly remains the first line of defence, but a modest daily dose of live yeast may help the rumen cope when the mixer wagon falls short.

Subject of Research: Microbiomes

Article Title: Modulation of the rumen microbiome and metabolism in dairy cows by altering the concentrate feeding pattern and the inclusion of Saccharomyces cerevisiae yeast

Article References: Snelling, T. J., Johnson, C. A., Warren, H. E., Taylor-Pickard, J., Huntington, J. A., & Sinclair, L. A. (2026). Modulation of the rumen microbiome and metabolism in dairy cows by altering the concentrate feeding pattern and the inclusion of Saccharomyces cerevisiae yeast. Frontiers in Microbiomes, 5, Article 1884444. https://doi.org/10.3389/frmbi.2026.1884444

Image Credits: AI Generated

DOI: 10.3389/frmbi.2026.1884444

Keywords: Modulation, rumen, microbiome, metabolism, dairy, cows, altering, concentrate, feeding, pattern, inclusion, Saccharomyces