BacCapSeq: a Platform for Diagnosis and Characterization of Bacterial Infections

Abstract

We report a platform that increases the sensitivity of high-throughput sequencing for detection and characterization of bacteria, virulence determinants, and antimicrobial resistance (AMR) genes. The system uses a probe set comprised of 4.2 million oligonucleotides based on the Pathosystems Resource Integration Center (PATRIC) database, the Comprehensive Antibiotic Resistance Database (CARD), and the Virulence Factor Database (VFDB), representing 307 bacterial species that include all known human-pathogenic species, known antimicrobial resistance genes, and known virulence factors, respectively. The use of bacterial capture sequencing (BacCapSeq) resulted in an up to 1,000-fold increase in bacterial reads from blood samples and lowered the limit of detection by 1 to 2 orders of magnitude compared to conventional unbiased high-throughput sequencing, down to a level comparable to that of agent-specific real-time PCR with as few as 5 million total reads generated per sample. It detected not only the presence of AMR genes but also biomarkers for AMR that included both constitutive and differentially expressed transcripts.IMPORTANCE BacCapSeq is a method for differential diagnosis of bacterial infections and defining antimicrobial sensitivity profiles that has the potential to reduce morbidity and mortality, health care costs, and the inappropriate use of antibiotics that contributes to the development of antimicrobial resistance.

Keywords: antibiotic resistance; bacterial identification; diagnostics; sequencing.

FIG 1

BacCapSeq yields more reads and higher genome coverage than unbiased high-throughput sequencing. (A) Graphic representation of read depth obtained with BacCapSeq or UHTS across the K. pneumoniae genome. (B) Representative BacCapSeq results for the toxR virulence gene obtained from whole-blood nucleic acid spiked with 40,000 copies/ml of V. cholerae DNA and (C) the bla_KPC AMR gene obtained from whole blood spiked with 40,000 live K. pneumoniae cells/ml. Probes are shown in black. BacCapSeq reads are shown in blue. UHTS reads are shown in brown.

FIG 2

Mapped bacterial reads in blood spiked with bacterial cells. Mapped bacterial reads were normalized to 1 million quality- and host-filtered reads obtained by BacCapSeq (blue) or UHTS (brown). The data shown represent 40,000 cells/ml. No cutoff threshold was applied.

FIG 3

Identification of bacteria in two immunosuppressed patients with HIV/AIDS and unexplained sepsis. (A) Infection with Salmonella enterica; (B) coinfection with Streptococcus pneumoniae and Gardnerella vaginalis; (C) genomic coverage of Gardnerella vaginalis.

FIG 4

Levels of seven transcripts in Staphylococcus aureus sensitive (AMR−) or resistant (AMR+) to ampicillin after culture for 45, 90, and 270 min in the presence of ampicillin. Box plots represent the log of normalized transcript counts for each gene. Only results obtained with BacCapSeq are shown because no transcripts were detected in the presence of ampicillin with UHTS until later time points.