Discovering Bacterial-Bacterial Interactions from Nasopharyngeal Microbiome Profiles

Background and aim: A healthy and stable bacterial community in the human upper respiratory tract (URT) during infancy has a significant impact on future respiratory health. We have previously described predictable succession of the URT microbiota over the first year of life, largely independent of external exposures. Interactions between bacterial species in the URT are likely to be key determinants of this succession. In this study, we aimed to further decipher the role of bacterial-bacterial interactions in temporal patterns from longitudinal NP microbiome profiles in silico.

Methods: We used V4 16S amplicon sequencing to study the nasopharyngeal (NP) bacteriome of 103 South African infants, sampled monthly over the first year of life. We inferred correlations from NP microbiome compositional data using Sparse Correlations for Compositional data (SparCC) networks. The rich pairwise networks connecting hundreds of interacting species across individual time points within the first year of life were identified from 16S rRNA gene profiles.

Results and conclusions: The bacterial-bacterial interactions were identified in 1,358 NP specimens from 103 healthy infants. A total of 145 pairwise interactions (correlation coefficient, r>0.3) between bacterial specimens were observed. We determined a total of 16 negative pairwise interactions and 129 positive pairwise interactions from our healthy cohort. The strongest positive interaction was found between Dolosigranulum and Corynebacterium genera, whilst the strongest negative interaction was observed between two different Streptococcus amplicon sequence variants.