Microbiome networks and change-point analysis reveal key community changes associated with cystic fibrosis pulmonary exacerbations

Abstract

Over 90% of cystic fibrosis (CF) patients die due to chronic lung infections leading to respiratory failure. The decline in CF lung function is greatly accelerated by intermittent and progressively severe acute pulmonary exacerbations (PEs). Despite their clinical impact, surprisingly few microbiological signals associated with PEs have been identified. Here we introduce an unsupervised, systems-oriented approach to identify key members of the microbiota. We used two CF sputum microbiome data sets that were longitudinally collected through periods spanning baseline health and PEs. Key taxa were defined based on three strategies: overall relative abundance, prevalence, and co-occurrence network interconnectedness. We measured the association between changes in the abundance of the key taxa and changes in patient clinical status over time via change-point detection, and found that taxa with the highest level of network interconnectedness tracked changes in patient health significantly better than taxa with the highest abundance or prevalence. We also cross-sectionally stratified all samples into the clinical states and identified key taxa associated with each state. We found that network interconnectedness most strongly delineated the taxa among clinical states, and that anaerobic bacteria were over-represented during PEs. Many of these anaerobes are oropharyngeal bacteria that have been previously isolated from the respiratory tract, and/or have been studied for their role in CF. The observed shift in community structure, and the association of anaerobic taxa and PEs lends further support to the growing consensus that anoxic conditions and the subsequent growth of anaerobic microbes are important predictors of PEs.

Microbiology of Cystic Fibrosis Airway Disease

“…In addition, recent studies based on 16S rRNA sequencing have highlighted the potential significance of anaerobes, whereby the relative abundance of anaerobic taxa in respiratory tract specimens of individuals with CF was dominant during pulmonary exacerbations. 161,162 In summary, many studies of the CF microbiome have recently documented a diversity much more complex than that described by conventional culture alone, with changes in relative abundance and structure of microbial communities in response to age, disease progression, and acute clinical events. Further studies are needed to understand how these changes impact clinical outcomes and are affected by therapeutic interventions.…”

A network approach to elucidate and prioritize microbial dark matter in microbial communities

“…Hubs may have ecological relevance to the community as their removal affects the largest number of connections and paths, causing the highest impact on the connectivity of the network [19,21]. Although many advances have been made in microbial ecology using network-driven approaches [22][23][24][25][26][27][28][29][30][31][32][33][34][35][36][37] few, if any, revolve around Unknown taxa. Most network analyses focus on the role of known species, usually excluding any taxonomically unassigned or ambiguous taxa in early filtering steps.…”

NetCoMi: Network Construction and Comparison for Microbiome Data in R

“…For instance, users can input a user-defined association or dissimilarity matrix rather than a data matrix, and then proceed with NetCoMi's standardized network analysis modules. In summary, we believe that NetCoMi is a useful addition to the modern microbiome data analysis toolbox, enabling rapid and reproducible microbial network estimation and comparison and ideally leading to robust hypotheses about the role of microbes in health and disease [111].…”