Targeted Sequencing Technique May Improve Bacterial Sepsis Diagnoses

Healthcare providers taking care of patients with bacterial sepsis have a very short window in which to diagnose and treat the infection. Not only does the bacterial cause of infection need to be pinpointed, but the antibiotic susceptibility profile of the isolate will determine the efficacy of any drug regimen. A new report from mBio describes a sequencing platform called BacCapSeq that may speed these processes.
mBio: BacCapSeq: a Platform for Diagnosis and Characterization of Bacterial Infections

The BacCapSeq method, described by first author Orchid Allicock and senior author W. Ian Lipkin, uses targeted sequencing data to provide insights on identity, pathogenicity markers, and antibiotic resistance characteristics, all simultaneously. “Our initial intent was to increase the sensitivity and coverage of high throughput sequencing while decreasing the cost and complexity of bioinformatics analysis,” said Lipkin.

To increase sensitivity, the research team designed a set of 4.2 million oligonucleotide probes to target open reading frames, including all known antibacterial resistance genes and virulence factors, found within the 307 bacterial species known to cause human disease. Probes incubated with a sample will hybridize to their target sequence and can be purified away from the rest of the sample, enriching a sequence which may be masked by other bacterial or human sequences. The overlapping nature of the probes captures nearly the complete protein coding sequence information, and results in higher coverage than unbiased high-throughput sequencing (see example, right). This translates to the ability to detect even 40 bacterial cells per ml of human blood, a hundred- or even thousand-fold increase in test sensitivity over unbiased methods.
While many proof-of-principle tests were run using spiked samples, the true test of the diagnostic capability of BacCapSeq lies in its ability to determine infectious etiologies of human samples. “Our focus is on diagnosis and improvements in clinical care,” said Lipkin. To this end, the scientific team tested 2 patient blood samples using BacCapSeq and compared it to UHTS. Both methods correctly identified the cause of infection in the samples, but BacCapSeq provided better genome coverage than unbiased methods.
From sample processing through bioinformatics analysis takes about 70 hours, or almost 3 days, which is comparable to the diagnostic identification timeline for some bacteria, but offers a faster diagnostic for slow-growing or difficult-to-cultivate species. But because BacCapSeq provides isolate identity, as well as resistance and virulence gene information, healthcare providers may more quickly know whether a patient is suffering from a drug-resistant or particularly virulent strain.
The technique may help better understand how bacteria respond to antibiotic stress, too. “As the sensitivity of BacCapSeq became apparent we realized that we could use it as a discovery tool for identifying biomarkers for antibiotic resistance that were expressed early after exposure to antibiotics in culture,” said Lipkin. The scientific team, based at Columbia University, is now working with bacteria exposed to antibiotics to develop probes to these antibiotic-responsive genes, which may act as bacterial biomarkers to predict responsiveness to drug exposure – and can provide valuable information about how bacteria respond to antibiotic stress. “We hope that others will find these data useful for basic and applied bacteriology,” said Lipkin.

Author: Julie Wolf

Julie Wolf
Dr. Julie Wolf is in science communications at Indie Bio, and was a former ASM employee. Follow Julie on Twitter for more ASM and microbiology highlights at @JulieMarieWolf.