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

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.