Abstract
Introduction/Aim: To gain better understanding of the human lung virome associated with health and disease, we retrieved shotgun metagenomic data from public repositories and implemented a newly developed pipEline for Viral assEmbly and chaRactEriSaTion (EVEREST) (https://agudeloromero.github.io/EVEREST/).
Methods: Lung virome studies (n = 85) utilising shotgun metagenomics from lung-attributed samples were collated from public databases. Following manual curation (n = 43), bioproject data (n = 30) was entered and run through EVEREST resulting in viral discovery and taxonomic classification of known and unknown viruses. Complete and high-quality contigs were further assessed for genomic characterisation.
Results: From 30 bioprojects, 868 samples were run through EVEREST, 201 samples contained 1850 viral contigs, with 149 classified as complete/high-quality genomes. Prokaryotic viruses and bacteriophages comprised 84% of viral contigs found, while 16% were eukaryotic viruses. Majority of bacteriophages were classified as Siphoviridae and Myoviridae, while Siphoviridae and Phycodnaviridae accounted for the highest proportion of eukaryotic viruses. Healthy and diseased cohorts showed significant differences in diversity, with acute respiratory infection (p-val = 0.017) and sarcoidosis (p-val = 0.0015) showing decreased viral diversity when compared to healthy cohorts. Viral families that are representative of other human niches which have not been previously described in the lungs were also observed.
Conclusion: We uncover for the first time the human lung virome in health and disease, identifying known and novel viruses utilising publicly available lung metagenome studies. Knowledge gained in the pulmonary virome will facilitate additional understanding of the virus-host microbiome interactions in various disease states and age groups.
Keywords: Virome, viral genomics, lung
Grant Support: Pawsey Supercomputing Research Centre, and Australian Government Research Training Program Scholarship.