Bovine milk microbiome: a more complex issue than expected

The aim of this study was to analyze bacterial profiles of bovine mastitic milk samples and samples from healthy quarters using Next Generation Sequencing of amplicons from 16S rRNA genes and to compare results with microbiological results by PCR assays of the same samples. A total of 49 samples were collected from one single dairy herd during the same day. The samples were divided in two sample sets, which were used in this study. The DNA extraction as well as the library preparation and sequencing of these two sets were performed separately, and results of the two datasets were then compared. The vast majority of genera detected appeared with low read numbers and/or in only a few samples. Results of PCR and microbiome analyses of samples infected with major pathogens Staphylococcus aureus or Streptococcus uberis were consistent as these genera also covered the majority of reads detected in the microbiome analysis. Analysis of alpha diversity revealed a much higher species richness in set 1 than in set 2. The dominating bacterial genera with the highest read numbers clearly differed between datasets, especially in PCR negative samples and samples positive for minor pathogens. In addition to this, linear discriminant analysis (LDA) was conducted between the two sets to identify significantly different genera/family level microbes. The genus Methylobacterium was much more common in set 2 compared to set 1, and genus Streptococcus more common in set 1. Our results indicate amplification of contaminating bacteria in excess in samples with no or minor amounts of pathogen DNA in dataset 2. There is a need for critical assessment of results of milk microbiome analyses.