Pneumonia or lower respiratory tract infection (LRTI) remains the leading cause of child death outside of the neonatal period, worldwide. The microbiota of the human nasopharynx is considered important in the pathogenesis of LRTI. The potential bacterial pathogens of the respiratory tract, which include Streptococcus pneumoniae, Haemophilus influenzae and Moraxella catarrhalis have been well-characterized, however, relatively little attention has been focused on the other members of the nasopharyngeal microbiome.
Whole metagenome (or shotgun) sequencing (WMGS) is a powerful method for determining the microbial composition of a niche at the species or strain level. Few studies have used this approach to characterize the NP microbiome and none, to our knowledge in children with lower respiratory tract infections. We used WMGS to study the species and strain-level composition of the NP microbiome in children in a South African birth cohort (Drakenstein Child Health Study) at the time of acute LRTI.
Samples from 150 children presenting with LRTI were processed for WMGS. Despite high host contamination and low biomass of the nasopharyngeal microbiome, we obtained high-quality data with an average of 520,437 bacterial reads per sample. Nasopharyngeal samples were frequently dominated by several species of the genus Moraxella including M. catarrhalis, M. nonliquefaciens and M. lincolnii. The role of M. nonliquefaciens and M. lincolnii in the nasopharynx is unknown. Haemophilus influenzae was the second most abundant species. Human adenovirus C was the most abundant DNA virus identified. Corynebacterium and Dolosigranulum were frequently identified, however, their mean relative abundance across all samples was less than 6%. While genus Dolosigranulum consisted of only one species, multiple understudied species of the genus Corynebacterium were identified in the samples. A high abundance of M. catarrhalis, H. influenzae, or M. nonliquefaciens, may be associated with hospitalization. Targeted PCR for five LRTI causing pathogens done in parallel showed good correlation with WMGS data. A total of 142 high quality draft genomes were constructed and many virulence factor genes and antibiotic resistance genes were identified in the Haemophilus and Moraxella genomes.