Dataset Information

Milk microbiome in dairy cattle and the challenges of low microbial biomass and exogenous contamination

ABSTRACT:

Background

The blanket usage of antimicrobials at the end of lactation (or “drying off”) in dairy cattle is under increasing scrutiny due to concerns about antimicrobial resistance. To lower antimicrobial usage in dairy farming, farmers are now encouraged to use “selective dry cow therapy” whereby only cows viewed as at high risk of mastitis are administered antimicrobial agents. It is important to gain a better understanding of how this practice affects the udder-associated microbiota and the potential knock-on effects on antimicrobial-resistant bacterial populations circulating on the farm. However, there are challenges associated with studying low biomass environments such as milk, due to known contamination effects on microbiome datasets. Here, we obtained milk samples from cattle at drying off and at calving to measure potential shifts in bacterial load and microbiota composition, with a critical assessment of contamination effects.

Results

Several samples had no detectable 16S rRNA gene copies and crucially, exogenous contamination was detected in the initial microbiome dataset. The affected samples were removed from the final microbiome analysis, which compromised the experimental design and statistical analysis. There was no significant difference in bacterial load between treatments (P > 0.05), but load was lower at calving than at drying off (P = 0.039). Escherichia coli counts by both sequence and culture data increased significantly in the presence of reduced bacterial load and a decreasing trend of microbiome richness and diversity. The milk samples revealed diverse microbiomes not reflecting a typical infection profile and were largely comprised of gut- and skin-associated taxa, with the former decreasing somewhat after prolonged sealing of the teats.

Conclusions

The drying off period had a key influence on microbiota composition and bacterial load, which appeared to be independent of antimicrobial usage. The interactions between drying off treatment protocol and milk microbiome dynamics are clearly complex, and our evaluations of these interactions were restricted by low biomass samples and contamination effects. Therefore, our analysis will inform the design of future studies to establish whether different selection protocols could be implemented to further minimise antimicrobial usage.

Supplementary Information

The online version contains supplementary material available at 10.1186/s42523-021-00144-x.

SUBMITTER: Pollock J

PROVIDER: S-EPMC8600933 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:BackgroundMilkability, primarily evaluated by measurements of milking speed and time, has an economic impact in milk production of dairy cattle. Recently the Italian Brown Swiss Breeders Association has included milking speed in genetic evaluations. The main objective of this study was to investigate the possibility of implementing genomic selection for milk flow traits in the Italian Brown Swiss population and thereby evaluate the potential of genomic selection for novel traits in medium-sized populations. Predicted breeding values and reliabilities based on genomic information were compared with those obtained from traditional breeding values.MethodsMilk flow measures for total milking time, ascending time, time of plateau, descending time, average milk flow and maximum milk flow were collected on 37 213 Italian Brown Swiss cows. Breeding values for genotyped sires (n?=?1351) were obtained from standard BLUP and genome-enhanced breeding value techniques utilizing two-stage and single-step methods. Reliabilities from a validation dataset were estimated and used to compare accuracies obtained from parental averages with genome-enhanced predictions.ResultsGenome-enhanced breeding values evaluated using two-stage methods had similar reliabilities with values ranging from 0.34 to 0.49 for the different traits. Across two-stage methods, the average increase in reliability from parental average was approximately 0.17 for all traits, with the exception of descending time, for which reliability increased to 0.11. Combining genomic and pedigree information in a single-step produced the largest increases in reliability over parent averages: 0.20, 0.24, 0.21, 0.14, 0.20 and 0.21 for total milking time, ascending time, time of plateau, descending time, average milk flow and maximum milk flow, respectively.ConclusionsUsing genomic models increased the accuracy of prediction compared to traditional BLUP methods. Our results show that, among the methods used to predict genome-enhanced breeding values, the single-step method was the most successful at increasing the reliability for most traits. The single-step method takes advantage of all the data available, including phenotypes from non-genotyped animals, and can easily be incorporated into current breeding evaluations.