Integrated meta-omics reveals new ruminal microbial features associated with feed efficiency in dairy cattle

Abstract

Background: As the global population continues to grow, competition for resources between humans and livestock has been intensifying. Increasing milk protein production and improving feed efficiency are becoming increasingly important to meet the demand for high-quality dairy protein. In a previous study, we found that milk protein yield in dairy cows was associated with the rumen microbiome. The objective of this study was to elucidate the potential microbial features that underpins feed efficiency in dairy cows using metagenomics, metatranscriptomics, and metabolomics.

Results: Comparison of metagenomic and metatranscriptomic data revealed that the latter was a better approach to uncover the associations between rumen microbial functions and host performance. Co-occurrence network analysis of the rumen microbiome revealed differential microbial interaction patterns between the animals with different feed efficiency, with high-efficiency animals having more and stronger associations than low-efficiency animals. In the rumen of high-efficiency animals, Selenomonas and members of the Succinivibrionaceae family positively interacted with each other, functioning as keystone members due to their essential ecological functions and active carbohydrate metabolic functions. At the metabolic level, analysis using random forest machine learning suggested that six ruminal metabolites (all derived from carbohydrates) could be used as metabolic markers that can potentially differentiate efficient and inefficient microbiomes, with an accuracy of prediction of 95.06%.

Conclusions: The results of the current study provided new insights into the new ruminal microbial features associated with feed efficiency in dairy cows, which may improve the ability to select animals for better performance in the dairy industry. The fundamental knowledge will also inform future interventions to improve feed efficiency in dairy cows. Video Abstract.

Keywords: Dairy cattle; Feed efficiency; Metabolomics; Metagenomics; Metatranscriptomics; Rumen microbiome.

Fig. 1

Workflow of the integrated rumen metagenomes, metatranscriptomes, and metabolomes

Fig. 2

Comparison of phenotypic data and rumen bacterial taxa identified in the metagenomes between cows with different feed efficiencies. Feed conversion rate (FCR), milk yield, nitrogen (N) efficiency, and dry matter intake (DMI) were compared using a t test (A). The 10 most abundant bacterial phyla (B), 10 most abundant bacterial genera (C), and 50 most abundant bacterial species (D). The Wilcoxon rank-sum test was used for mean comparison. *P < 0.05

Fig. 3

Fold changes of metabolic pathways identified in the metagenomes and metatranscriptomes of the cows with high and low feed efficiencies. A Pathways identified in the metagenomes and B pathways identified in the metatranscriptomes. The Wilcoxon rank-sum test was used for mean comparison. *P < 0.05

Fig. 4

Co-occurrence networks of bacterial taxa. A The co-occurrence among rumen bacteria in the dairy cows with high and low feed efficiencies. B Relationships between rumen microbial taxa and feed efficiency-associated microbial functions. Only significant (P < 0.05) relationships are shown. Blue edges indicate positive relationships, and red edges indicate negative relationships. The node size is proportional to the mean abundance

Fig. 5

Prediction of host feed efficiency using rumen metabolites and microbe-metabolite interactions. Receiver operating characteristic (ROC) curve and the confusion matrix of the performance of the random forest model using the six selected metabolites (shown in red) whose mean decrease accuracy (MDA) was > 4 (A). Biplot drawn from the microbe-metabolite vectors (mmvec) conditional probabilities estimated for the dataset of high-efficiency (B) and low-efficiency (C) cows. Axes: principal components from the singular value decomposition of the microbe-metabolite conditional probabilities estimated using mmvec. Arrows: microbes, dots: metabolites, and colors of dots represent associations with host feed efficiency (blue: negative, red: positive). Heatmaps display the inferred conditional probabilities for various efficiency-associated metabolites given the presence of specific microbial taxa in the rumen of cows with high (B) and low (C) efficiencies

Fig. 6

A working model to illustrate the microbial taxa, active carbohydrate metabolism, and metabolites that might be associated with feed efficiency in dairy cows