In addition to providing nutritional and bioactive factors necessary for infant development, human breast milk contains bacteria that contribute to the establishment of commensal microbiota in the infant. However, the composition of this bacterial community differs considerably between studies. We hypothesised that bacterial DnA extraction methodology from breast milk samples are a substantial contributor to these inter-study differences. We tested this hypothesis by applying five widely employed methodologies to a mock breast milk sample and four individual human breast milk samples. Significant differences in DNA yield and purity were observed between methods (P < 0.05). Microbiota composition, assessed by 16S rRNA gene amplicon sequencing, also differed significantly with extraction methodology (P < 0.05), including in the contribution of contaminant signal. Concerningly, many of the bacterial taxa identified here as contaminants have been reported as components of the breast milk microbiome in other studies. These findings highlight the importance of using stringent, well-validated, DNA extraction methodologies for analysis of the breast milk microbiome, and exercising caution interpreting microbiota data from low-biomass contexts. In addition to nutrients and bioactive components, human breast milk contains bacteria 1 , commonly referred to as the breast milk microbiome. These bacterial populations are thought to contribute to the establishment of commensal bacterial communities in the infant, as well as supporting the maturation of the gut, and development of the immune system 1,2. The potential importance of microbes introduced with breast milk on early-life development has led to increasing efforts to accurately define them. Early culture-and PCR-based methods, predominantly reported the presence of members of the Staphylococcus, Streptococcus, Lactobacillus and Bifidobacterium genera 3-12. More recently, 16S rRNA gene amplicon sequencing has become an increasingly popular approach to characterising the bacterial content of breast milk samples 13-33. By enabling the detection of species that are refractory to standard culture methodologies, as well as providing insight into the relative abundance of constitutive bacteria, 16S rRNA gene sequencing offers substantial analytical advantages. However, the composition of the breast milk microbiome reported in sequencing-based studies varies substantially. For example, while two studies have described a "core" human breast milk microbiome (bacteria that are present in all samples analysed) from healthy mothers, consisting of 7 and 9 taxa respectively, only three of these taxa were present in both (Propionibacterium, Staphylococcus, and Streptococcus) 27,34. Notably, many of the other bacterial taxa often reported as part of the breast milk microbiome, including