AUTHOR=Si Jiyeon , Choi Yongbin , Raes Jeroen , Ko Gwangpyo , You Hyun Ju 

TITLE=Sputum Bacterial Metacommunities in Distinguishing Heterogeneity in Respiratory Health and Disease

JOURNAL=Frontiers in Microbiology

VOLUME=Volume 13 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/microbiology/articles/10.3389/fmicb.2022.719541

DOI=10.3389/fmicb.2022.719541

ISSN=1664-302X

ABSTRACT=Background and Objective Cluster-based analysis, or community typing, has been attempted as a method for studying the human microbiome in various body niches with the aim of reducing variations in bacterial composition and linking the defined communities to host health and disease. Here, we present the bacterial sub-communities in healthy and diseased population cohorts and assess whether these sub-communities can distinguish different host health conditions. 
Methods We performed community typing analysis on the sputum microbiome dataset obtained from a healthy Korean twin-family cohort (n = 202) and an external chronic obstructive pulmonary disease (COPD) cohort (n = 324) and implemented networks analysis to investigate the associations of bacterial metacommunities with host health parameters and microbial interactions in disease.
Results Analysis of the sputum microbiome of a healthy Korean cohort revealed high levels of inter-individual variation, which was driven by two dominant bacteria: Neisseria and Prevotella. Community typing of the cohort samples identified three metacommunities, namely, Neisseria 1 (N1), Neisseria 2 (N2), and Prevotella (P), each of which showed different functional potential and links to host traits (e.g., triglyceride levels, waist circumference, and high-sensitivity C-reactive protein). In particular, the Prevotella-dominant metacommunity showed low community diversity, which implies an adverse health association. Network analysis of the healthy twin cohort illustrated co-occurrence of Prevotella with pathogenic anaerobic bacteria; this bacterial cluster was negatively associated with high-density lipoproteins but positively correlated with waist circumference, blood pressure, and pack-years. Community typing of the external COPD cohort identified three sub-metacommunities: one exclusively including healthy subjects (HS) and the other two (CS1 and CS2) comprising patients. The two COPD metacommunities, CS1 and CS2, showed different abundances of specific pathogens such as Serratia and Moraxella, as well as differing functional potential and community diversity. Network analysis of the COPD cohort showed enhanced bacterial co-exclusions in the CS metacommunities when compared with HS metacommunity. 
Conclusions Overall, our findings point to a potential association between pulmonary Prevotella and host health and disease, making it possible to implement community typing for diagnosis of heterogenic respiratory disease.