Species Identification and Antibiotic Resistance Prediction by Analysis of Whole-Genome Sequence Data by Use of ARESdb: an Analysis of Isolates from the Unyvero Lower Respiratory Tract Infection Trial.

Ines Ferreira,Matthias Klein,Arndt Von Haeseler,Andreas E Posch,Thomas Weinmaier,Lukas Lueftinger,Johannes Bacher,Robin Patel,Stephan Beisken,Alexander Mellmann

doi:10.1128/jcm.00273-20

Abstract

Whole-genome sequencing (WGS) is now routinely performed in clinical microbiology laboratories to assess isolate relatedness. With appropriately developed analytics, the same data can be used for prediction of antimicrobial susceptibility. We assessed WGS data for identification using open-source tools and antibiotic susceptibility testing (AST) prediction using ARESdb compared to matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) identification and broth microdilution phenotypic susceptibility testing on clinical isolates from a multicenter clinical trial of the FDA-cleared Unyvero lower respiratory tract infection (LRTI) application (Curetis). For the trial, more than 2,000 patient samples were collected from intensive care units across nine hospitals and tested for LRTI. The isolate subset used in this study included 620 clinical isolates originating from 455 LRTI culture-positive patient samples. Isolates were sequenced using the Illumina Nextera XT protocol and FASTQ files with raw reads uploaded to the ARESdb cloud platform (ares-genetics.cloud; released for research use in 2020). The platform combines Ares Genetics' proprietary database ARESdb with state-of-the-art bioinformatics tools and curated public data. For identification, WGS showed 99 and 93% concordance with MALDI-TOF MS at the genus and species levels, respectively. WGS-predicted susceptibility showed 89% categorical agreement with phenotypic susceptibility across a total of 129 species-compound pairs analyzed, with categorical agreement exceeding 90% in 78 species-compound pairs and reaching 100% in 32. Results of this study add to the growing body of literature showing that, with improvement of analytics, WGS data could be used to predict antimicrobial susceptibility.

Highlights

Whole-genome sequencing (WGS) is routinely performed in clinical microbiology laboratories to assess isolate relatedness
We compared WGS-based bacterial identification and antibacterial susceptibility testing against matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) identification and broth microdilution susceptibility testing using a set of clinical isolates from the clinical trial of the Unyvero lower respiratory tract infection (LRTI) application, a multiplex-PCR based sample-to-answer solution manufactured by Curetis
Isolates were subjected to identification via matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF MS) and phenotypic susceptibility testing by broth microdilution in parallel to WGS and subsequent WGS-based identification using open-source tools and antibiotic susceptibility testing (AST) prediction using ARESdb (Fig. 1)

Summary

Introduction

Whole-genome sequencing (WGS) is routinely performed in clinical microbiology laboratories to assess isolate relatedness. We assessed WGS data for identification using open-source tools and antibiotic susceptibility testing (AST) prediction using ARESdb compared to matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) identification and broth microdilution phenotypic susceptibility testing on clinical isolates from a multicenter clinical trial of the FDA-cleared Unyvero lower respiratory tract infection (LRTI) application (Curetis). Bacterial cultures and antimicrobial susceptibility testing (AST) are established tools for detecting, identifying, and defining the antimicrobial susceptibility of bacteria in clinical practice Their advantages and disadvantages have been reviewed extensively [8, 9]. We compared WGS-based bacterial identification and antibacterial susceptibility testing against matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) identification and broth microdilution susceptibility testing using a set of clinical isolates from the clinical trial of the Unyvero lower respiratory tract infection (LRTI) application, a multiplex-PCR based sample-to-answer solution manufactured by Curetis. The incidence of hospital-acquired LRTI in Europe and the United States is estimated at 1 to 12 cases per 1,000 individuals [10,11,12]; 10 to 39% of patients receive inappropriate empirical treatment [13, 14], in part because of the spread of AMR

Methods

Results

Discussion

Conclusion