Rapid and reliable identification of bacterial pathogens directly from patient samples is required for optimizing antimicrobial therapy. Although Sanger sequencing of the 16S ribosomal RNA (rRNA) gene is used as a molecular method, species identification and discrimination is not always achievable for bacteria as their 16S rRNA genes have sometimes high sequence homology. Recently, next generation sequencing (NGS) of the 16S–23S rRNA encoding region has been proposed for reliable identification of pathogens directly from patient samples. However, data analysis is laborious and time-consuming and a database for the complete 16S–23S rRNA encoding region is not available. Therefore, a better, faster, and stronger approach is needed for NGS data analysis of the 16S–23S rRNA encoding region. We compared speed and diagnostic accuracy of different data analysis approaches: de novo assembly followed by Basic Local Alignment Search Tool (BLAST), operational taxonomic unit (OTU) clustering, or mapping using an in-house developed 16S–23S rRNA encoding region database for the identification of bacterial species. De novo assembly followed by BLAST using the in-house database was superior to the other methods, resulting in the shortest turnaround time (2 h and 5 min), approximately 2 h less than OTU clustering and 4.5 h less than mapping, and a sensitivity of 80%. Mapping was the slowest and most laborious data analysis approach with a sensitivity of 60%, whereas OTU clustering was the least laborious approach with 70% sensitivity. Although the in-house database requires more sequence entries to improve the sensitivity, the combination of de novo assembly and BLAST currently appears to be the optimal approach for data analysis.

Rapid and sensitive diagnostic strategies are necessary for patient care and public health. Most of the current conventional microbiological assays detect only a restricted panel of pathogens at a time or require a microbe to be successfully cultured from a sample. Clinical metagenomics next-generation sequencing (mNGS) has the potential to unbiasedly detect all pathogens in a sample, increasing the sensitivity for detection and enabling the discovery of unknown infectious agents. High expectations have been built around mNGS; however, this technique is far from widely available. This review highlights the advances and currently available options in terms of costs, turnaround time, sensitivity, specificity, validation, and reproducibility of mNGS as a diagnostic tool in clinical microbiology laboratories. The need for a novel diagnostic tool to increase the sensitivity of microbial diagnostics is clear. mNGS has the potential to revolutionize clinical microbiology. However, its role as a diagnostic tool has yet to be widely established, which is crucial for successfully implementing the technique. A clear definition of diagnostic algorithms that include mNGS is vital to show clinical utility. Similarly to real-time PCR, mNGS will one day become a vital tool in any testing algorithm.

Whole-genome sequencing (WGS) of Mycobacterium tuberculosis (MTB) isolates can be used to get an accurate diagnosis, to guide clinical decision making, to control tuberculosis (TB) and for outbreak investigations. We evaluated the performance of long-read (LR) and/or short-read (SR) sequencing for anti-TB drug-resistance prediction using the TBProfiler and Mykrobe tools, the fraction of genome recovery, assembly accuracies and the robustness of two typing approaches based on core-genome SNP (cgSNP) typing and core-genome multi-locus sequence typing (cgMLST). Most of the discrepancies between phenotypic drug-susceptibility testing (DST) and drug-resistance prediction were observed for the first-line drugs rifampicin, isoniazid, pyrazinamide and ethambutol, mainly with LR sequence data. Resistance prediction to second-line drugs made by both TBProfiler and Mykrobe tools with SR- and LR-sequence data were in complete agreement with phenotypic DST except for one isolate. The SR assemblies were more accurate than the LR assemblies, having significantly (P<0.05) fewer indels and mismatches per 100 kbp. However, the hybrid and LR assemblies had slightly higher genome fractions. For LR assemblies, Canu followed by Racon, and Medaka polishing was the most accurate approach. The cgSNP approach, based on either reads or assemblies, was more robust than the cgMLST approach, especially for LR sequence data. In conclusion, anti-TB drug-resistance prediction, particularly with only LR sequence data, remains challenging, especially for first-line drugs. In addition, SR assemblies appear more accurate than LR ones, and reproducible phylogeny can be achieved using cgSNP approaches.

Outbreak management of extended spectrum β-lactamase (ESBL)-producing pathogens requires rapid and accurate diagnosis. However, conventional screening is slow and labor-intensive. The vast majority of the screened samples are negative and detection of non-outbreak-related resistant micro-organisms often complicates outbreak management. In a CTX-M-15-producing Escherichia coli outbreak, 149 fecal samples and rectal eSwabs were collected by a cross-sectional survey in a Dutch nursing home. Samples were processed by routine diagnostic methods. Retrospectively, ESBL-producing bacteria and resistance genes were detected directly from eSwab medium by an accelerated workflow without prior enrichment cultures by an amplicon-based next-generation sequencing (NGS) method, and culture. A total of 27 (18.1%) samples were positive in either test. Sensitivity for CTX-M detection was 96.3% for the phenotypic method and 85.2% for the NGS method, and the specificity was 100% for both methods, as confirmed by micro-array. This resulted in a positive predictive value (PPV) of 100% for both methods, and a negative predictive value (NPV) of 99.2% and 96.8% for the phenotypic method and the NGS method, respectively. Time to result was four days and 14 h for the phenotypic method and the NGS method, respectively. In conclusion, the sensitivity without enrichment shows promising results for further use of amplicon-based NGS for screening during outbreaks.

Background: Machine learning (ML) applied to lung and gut microbiota offers substantial yet underexplored potential for improving tuberculosis (TB) diagnosis and prognosis. Progress has been limited by insufficiently large and standardised datasets. Here, we applied ML models to globally curated microbiota datasets to classify TB cases and healthy controls.Methods: Datasets are from our recently published systematic review of 42 studies (2010–2024) reporting microbiota changes associated with TB. The harmonised lung and gut datasets include n=1,617 participants. Six ML model types were optimised across 36 parameters for diagnostic (TB status) and prognostic (treatment status) classification. Models were trained, cross-validated, and independently validated using 70%, 20%, and 10% of each dataset, respectively.Findings: The dataset included 1,181 lung and 550 gut microbiota samples, of which 1,264 (73%) were microbiologically confirmed TB cases; 89% of samples originated from Africa and Asia. In the global lung dataset, Lasso II and Ridge II models achieved an area under the receiver operating characteristic curve (AUROC) of 0.93 and 0.92 during cross-validation and correctly classified 64 of 75 TB cases. External independent validation (n=94) showed comparable performance, correctly classifying 63 of 75 cases. Although Africa- and Asia-specific models performed similarly to global models during cross-validation, their performance did not generalise to external validation. Random Forest models performed best on treatment-phase lung samples and on gut microbiota datasets, achieving cross-validation metrics comparable to those from the lung analyses. Genera such as Rothia, Gemella, Prevotella, and Streptococcus consistently ranked among the strongest predictors of pulmonary TB, suggesting potential biomarker value.Interpretation: ML models leveraging lung and gut microbiota show promise as complementary tools for TB diagnosis, particularly for extrapulmonary or hard-to-diagnose cases. This work provides a foundation for clinical metagenomics in TB diagnostics and prognostics. To support reproducibility and future innovation, we release MLTBdx, an open-source pipeline and dataset.