Integrative microbiomes analyses identify opportunistic pathogens of patients with lower respiratory tract infections based on Metagenomic Next-Generation Sequencing

Abstract

Abstract Background Lower Respiratory tract infections (LRTIs) are one of the most widespread and harmful diseases. As an emerging technology, the rapid development of metagenomic next-generation sequencing (mNGS) has advantages for the unbiased etiological detection and greatly meet the needs for the clinical diagnosis. However, little is known about how to interpret the application of mNGS results, especially for the classification of opportunistic pathogens from colonization or infection in patients with LRTIs. Methods We performed a retrospective study of respiratory tract samples from 394 patients and 179 health based on the mNGS to assess pathogens and the airway microbiome. Results 576 discriminant species (442 bacteria and 134 fungi) were achieved from the patients with LRTIs and controls by discriminant analysis. Specifically, these differential species were clustered and charactered into cohort 1, 2, 3 including the colonizing microbiota, emerging opportunistic pathogens and rare opportunistic pathogens on the basis of their correlation profiles, detection frequency and relative abundance. In these 3 cohorts, pathogens from the cohort 2 obtained an average area under the curve (AUC) of 0.976 for the best predictive performance, followed by cohort 1 (0.961) and cohort 3 (0.887). In addition, 46 responsible pathogens (30 bacteria and 16 fungi) were further identified from the three cohorts and achieved good performance of predictive value in LRTIs diagnose (AUC = 0.988). Co-abundance analysis of the ecological network revealed patients with LRITs were more complex and appeared modularly in the focus of the opportunistic pathogens. Conclusion Our findings provided a profile of LRTIs-associated bacterial and fungal colonization or opportunistic pathogens in relatively large-scale statistics, which provides potential reference evaluation criterions that contribute to the mNGS report result interpretations including those caused by unknown pathogens in clinical practice.