Metagenomic Assessment of the Pathogenic Risk of Microorganisms in Sputum of Postoperative Patients With Pulmonary Infection

Abstract

Respiratory infections are complicated biological processes associated with an unbalanced microbial community and a wide range of pathogens. To date, robust approaches are still required for distinguishing the pathogenic microorganisms from the colonizing ones in the clinical specimens with complex infection. In this study, we retrospectively analyzed the data of conventional culture testing and metagenomic next-generation sequencing (mNGS) of the sputum samples collected from 50 pulmonary infected patients after cardiac surgery from December 2020 and June 2021 in Ruijin Hospital. Taxonomic classification of the sputum metagenomes showed that the numbers of species belonging to bacteria, fungi, and viruses were 682, 58, and 21, respectively. The full spectrum of microorganisms present in the sputum microbiome covered all the species identified by culture, including 12 bacterial species and two fungal species. Based on species-level microbiome profiling, a reference catalog of microbial abundance detection limits was constructed to assess the pathogenic risks of individual microorganisms in the specimens. The proposed screening procedure detected 64 bacterial pathogens, 10 fungal pathogens, and three viruses. In particular, certain opportunistic pathogenic strains can be distinguished from the colonizing ones in the individual specimens. Strain-level identification and phylogenetic analysis were further performed to decipher molecular epidemiological characteristics of four opportunistic etiologic agents, including Klebsiella pneumoniae, Corynebacterium striatum, Staphylococcus aureus, and Candida albicans. Our findings provide a novel metagenomic insight into precision diagnosis for clinically relevant microbes, especially for opportunistic pathogens in the clinical setting.

Keywords: limit of detection; metagenomic next-generation sequencing; opportunistic pathogen; pathogen risk; pulmonary infection; sputum microbiome; strain profiling.

Figure 1

Bacterial community structure and diversity of the sputum microbiota from the post-surgery patients with pulmonary infection. (A) An accumulation curve of the microorganisms in the sputum samples. The boxplot denotes the species richness according to the observed species numbers for a given number of samples collected from 100 random permutations of all the samples. (B) Percent stacked barplot for species-level taxonomic profiling of the sputum bacterial microbial communities. Labels stand for the top 12 abundant species. The samples are grouped by clinical outcomes Recover and Worse. (C) Chao1 index of the sputum microbiota from the two groups Recover and Worse. (D) The ANOSIM boxplot ranks between and within the Recover and Worse groups.

Figure 2

Maximum-likelihood phylogenetic trees of three bacterial pathogens. Using StrainPhlAn trees are built based on the sample-specific consensus sequences of marker genes from K. pneumoniae (A), C. striatum (B), and S. aureus (C), respectively. The strains detected in the metagenomic samples are labeled in red. The isolate genomes are labeled by the GenBank accessions, strain identifiers, isolation country, and year. Bootstrap support values greater than 80% are displayed.

Figure 3

Maximum-likelihood phylogenetic trees of the fungal pathogen C. albicans. Trees are built according to the StrainPhlAn analysis (A) and MLST analysis (B), respectively. The strain of C. albicans in the metagenome sample #13 is excluded from the MLST tree due to its low sequencing depth and failure to reconstruct the MLST loci-sequences.