Comparison among four different bacterial DNA extraction protocols for analysing milk metagenomics

Bovine udder is colonized by a huge number of microorganisms that constitute the intramammary ecosystem, with a specific role in modulating not only the udder homeostasis and mastitis susceptibility but also the quality of the dairy products. Therefore, information on milk microbiota composition will facilitate the dairy industry in the production of safe and high-quality products. However, generating high-quality bacterial DNA could be critical. In the present study, bacterial DNA from healthy milk samples was isolated by four different protocols to evaluate the effect of the extraction procedures on milk microbiota composition. For the characterization of the milk microbiota by 16S deep sequencing, 500 ml of bulk tank milk samples were aseptically collected from three different farms and bacterial DNA was extracted by using an internal laboratory protocol and three commercial kits. Bacterial DNA was then amplified using the primers for the V3-V4 hypervariable regions and sequenced in one MiSeq (Illumina) run with 2×250-base paired-end reads. Data analysis was performed by using QIIME suite and SILVA 132 as reference database for taxonomy. The results showed that the four extraction kits performed very differently and showed a significant separation on both microbial richness (alpha-diversity) and composition (beta-diversity). In particular, the relative abundance of some genera (e.g.: Lactobacillus, Acinetobacter and Microbacterium) were consistently altered by the extraction method. Based on these data, then, particular attention must be kept in choosing the proper extraction method, for example, carefully evaluating eventual biases towards or against bacterial genera of interest. Moreover, we believe that, in order to define which kit best resembles the original bacterial community, an additional set of experiments with a mock community with known composition should be performed.