Image created by Dr. Michael J. Miller
A multidisciplinary research team co-led by investigators at the University of North Carolina at Chapel Hill and Tulane University has developed a CRISPR-based diagnostic assay that rapidly detects and differentiates 15 nontuberculous mycobacteria (NTM) species and subspecies directly from patient specimens.
Developed by UNC and Tulane researchers, the microfluidic assay demonstrated over 97% sensitivity directly from clinical specimens, cutting turnaround time from weeks to hours.
Published in Science Translational Medicine, the study outlines the CRISPR-assisted nanodroplet differential identification (CANDI) platform, which integrates population genomics, clinical microbiology, and microfluidic CRISPR engineering. The investigators have also secured a five-year R01 grant to support further clinical validation and assay expansion.
NTM represent an environmentally widespread and biologically diverse group of bacteria, dozens of which cause human disease. Because pathogenicity, therapeutic susceptibility, and clinical management differ substantially across species, rapid and accurate identification is essential for guiding targeted antimicrobial therapy.
To establish the diagnostic targets, the research team analyzed seven years of diagnostic records across the UNC Health system to identify the organisms most frequently encountered in clinical practice.
“Species-level identification is critical in NTM because different organisms can require very different treatment approaches,” says Melissa B Miller, PhD, director of the UNC clinical microbiology laboratory, in a release. “Being able to obtain that information much earlier could help clinicians make more informed treatment decisions.”
Population Genomics Target Selection
To ensure diagnostic reliability across circulating bacterial strains, the researchers evaluated candidate target sequences using data from more than 100,000 mycobacterial genomes. The team systematically screened candidate regions across the genetic breadth of each target species, identifying sequences that were strictly conserved within a species while remaining distinct from other mycobacteria.
“A diagnostic target cannot simply work in one reference genome—it needs to work across the diversity we actually see in patients,” says Qingyun Liu, assistant professor in the department of genetics at UNC, in a release. “By combining population genomics, clinical microbiology and CRISPR engineering, we were able to design the assay around real-world bacterial diversity from the beginning.”
Microfluidic Multiplexing and Clinical Performance
To overcome the multiplexing constraints common to CRISPR-based diagnostics, the Tulane engineering team created a keyhole-shaped microfluidic chip that organizes thousands of miniature CRISPR reactions in parallel. Each reaction is tagged with a distinct fluorescence barcode and combined with DNA from a patient sample, enabling simultaneous detection of multiple targets while keeping individual reactions physically partitioned.
In retrospective evaluations of archived clinical isolates, the CANDI platform delivered 98.20% sensitivity and 99.65% specificity in culture-derived samples.
The investigators then evaluated the assay on residual specimens collected during routine clinical care across a 125-participant cohort, which included 61 culture-positive and 64 culture-negative individuals. CANDI demonstrated 96.0% concordance with traditional culture results, correctly categorized all culture-negative samples, and demonstrated an overall sensitivity of 97.08% and specificity of 99.73%.
The platform yielded species-level results within approximately 3.5 hours, contrasting with the days to weeks typically required by standard culture-based diagnostic pathways.
Panel Expansion and Longitudinal Monitoring
Under the new five-year R01 grant, the research group plans to broaden the CANDI panel to cover approximately 30 clinically relevant NTM species and subspecies by incorporating additional population genomic targets into the microfluidic system.
The team is also initiating multicenter clinical validation using retrospective and prospective cohorts, including the seven-center HALT cohort funded by the Cystic Fibrosis Foundation, to facilitate integration into clinical laboratory workflows. Furthermore, researchers will investigate whether quantitative changes in NTM DNA can serve as a biomarker for treatment response, evaluating the approach in mouse infection models of Mycobacterium avium and Mycobacterium abscessus before transitioning to longitudinal patient cohorts.
Reference
Hongquan Gou et al., A rapid CRISPR-based nanodroplet assay enables direct clinical identification of mycobacteria species. Sci. Transl. Med.18, eaef2648(2026).DOI:10.1126/scitranslmed.aef2648.
Abstract
The global incidence and mortality of nontuberculous mycobacterial infections have risen sharply with population aging. In some regions, they are now surpassing Mycobacterium tuberculosis complex infections, imposing a substantial clinical and economic burden. Because nontuberous mycobacteria exhibit species-level heterogeneity and require prolonged culture for identification, their diagnosis remains slow and is frequently inaccurate. Here, we describe a multiplexed clustered regularly interspaced short palindromic repeats (CRISPR)–assisted nanodroplet differential identification (CANDI) diagnostic platform that integrates species-agnostic target amplification with species-specific CRISPR-associated protein 12a (Cas12a) detection in fluorescence-barcoded nanodroplets. By spatially compartmentalizing CRISPR reactions into color-encoded nanodroplets, CANDI overcomes the multiplexing limitations of conventional CRISPR diagnostics and enables simultaneous interrogation of multiple mycobacterial targets in a single assay. We designed a 16-plex panel that distinguishes 15 clinically relevant Mycobacterium species and subspecies. CANDI achieved high analytical sensitivity and accurate discrimination in samples containing coinfections with multiple species or subspecies. When applied to 230 clinical specimens, including sputum, tracheal aspirates, and other respiratory fluids, CANDI delivered subspecies-level results within 3.5 hours, achieving 97.08% sensitivity and 99.7% specificity relative to culture-based identification. By combining multiplexed, high-specificity CRISPR detection with scalable droplet-based engineering, CANDI has the potential to overcome the culture dependency of current diagnostics and enable species- and subspecies-level identification across the genetically complex Mycobacterium genus, offering a clinically adaptable framework for rapid, precision diagnosis of mycobacterial infections.