ABSTRACT: The human breast milk (HBM) bacteriome is an important, continuous source of microbes to the neonate in early life, playing an important role in shaping the infant's intestinal bacteriome. Study of the composition of the HBM bacteriome is an emerging area of research, with little information available, particularly from low- and middle-income countries. The aim of this study was to characterize the diversity of bacterial communities in HBM samples collected between 6-10 weeks postpartum from lactating South African women and to study potential influencing factors of the bacteriome. Using 16S rRNA gene sequencing of samples from 554 women, we demonstrated that the HBM bacteriome was largely dominated by the phyla Firmicutes (mean relative abundance: 71.1%) and Actinobacteria (mean relative abundance: 16.4%). The most abundant genera identified from the HBM bacteriome were Streptococcus (mean relative abundance: 48.6%), Staphylococcus (mean relative abundance: 17.8%), Rothia (mean relative abundance: 5.8%), and Corynebacterium (mean relative abundance: 4.3%). "Core" bacterial genera including Corynebacterium, Streptococcus, Staphylococcus, Rothia, Veillonella, Gemella, Acinetobacter, Micrococcus and a genus belonging to the Enterobacteriaceae family were present in 80% of samples. HBM samples were classified, according to their bacteriome, into three major clusters, dominated by the genera Staphylococcus (cluster 1), a combination of Staphylococcus and Streptococcus (cluster 2), and Streptococcus (cluster 3). The cluster groups differed significantly for Shannon and chao1 richness indices. Bacterial interactions were studied using co-occurrence networks with positive associations observed between the abundances of Staphylococcus and Corynebacteria (members of the skin microflora) and between Streptococcus, Rothia, Veillonella, and Gemella (members of the oral microflora). HBM from older mothers had a higher Shannon diversity index. The study site was associated with differences in HBM bacteriome composition (permutational multivariate analysis of variance using distance matrices (PERMANOVA), p?
