Use of Metagenomic Next-Generation Sequencing in Diagnosis of Respiratory Viruses in a Cohort of Patients Admitted to Intensive Care Unit
Refereed conference paper presented and published in conference proceedings

Other information
Current laboratory diagnosis of respiratory viruses relies on target-based molecular assays such as polymerase chain reactions that require prior knowledge of suspected pathogens. In contrast, emerging metagenomic next-generation sequencing (mNGS) technology has no such constraint. Here, we aimed at evaluating the clinical utility of mNGS in diagnosing respiratory viruses in a cohort of patients hospitalized for severe respiratory illnesses.

This was a retrospective laboratory study. Leftover diagnostic respiratory samples collected from patients hospitalized for severe respiratory illness of unknown etiology and admitted to intensive care unit of our teaching hospital were randomly selected for further investigations. These samples had tested negative for common respiratory viruses (influenza A and B viruses, respiratory syncytial A and B viruses, and parainfluenza virus types 1 to 4) in routine diagnosis. After removal of cell-free genetic materials, total DNA and RNA were extracted and purified by MagMAX Viral RNA Isolation kit, followed by cDNA synthesis and random amplification. Sequencing library preparation was performed using Nextera XT DNA Library Prep Kit. Paired-end 2x75 bp mNGS was performed on a MiSeq System. Sequencing reads were trimmed by Trimmomatic to remove adapter sequences and poor quality region. Taxonomic classification of sequencing reads was performed using the web-based Taxonomer.

A total of 39 respiratory (24 nasopharyngeal aspirate [NPA] and 15 tracheal aspirate [TA]) samples were metagenomically sequenced. The patient median age was 35 years (interquartile range: 5-70 years). On average 570,000 high-quality reads were generated per sample. We identified four (10%) samples that had more than 1% of reads classified as viral and one predominant (>95% of viral reads) virus was identified in each of these samples: human coronavirus 229E (from NPA of a 50-year-old woman; 6.6% of total reads), TTV-like mini virus (TA of a 30-year-old woman; 5.7%), human metapneumovirus (NPA of a 3-year-old boy; 4.4%), and Torque teno virus (NPA of a 1-year-old boy; 1.4%). In addition, two NPA samples had more than 1% of reads mapped to fungal species. Contaminating bacterial reads known to originate from molecular reagents were abundant.

Our findings illustrated the clinical utility of mNGS in supplementing routine diagnosis of respiratory viruses. Technical challenges remain, especially in the interpretation of results on bacteria due to prevalent contaminating reads.

This study was supported in part by the Commissioned Health and Medical Research Fund (Phase 3) of the Food and Health Bureau of the Hong Kong SAR Government (to MCWC; CU-17-B6).
Acceptance Date11/01/2018
All Author(s) ListMartin Chi-Wai Chan, Lin-Yao Zhang, Kirsty Kwok, Wai-Tat Wong, Grace C.Y. Lui, Kam Lun Hon, David S.C. Hui, Paul K.S. Chan
Name of Conference2nd International Meeting on Respiratory Pathogens
Start Date of Conference07/03/2018
End Date of Conference09/03/2018
Place of ConferenceSingapore
Country/Region of ConferenceSingapore
LanguagesEnglish-United States
Keywordsinfluenza, intensive care unit, metagenomic NGS, respiratory virus

Last updated on 2018-31-05 at 15:37