Development, Evaluation, and Implementation of a House-Made Targeted Next-Generation Sequencing Spoligotyping in a French Laboratory.

Epidemiological studies investigating transmission chains of tuberculosis are undertaken worldwide to tackle its spread. CRISPR locus diversity, called spoligotyping, is a widely used genotyping assay for Mycobacterium tuberculosis complex (MTBC) characterization. Herein, we developed a house-made targeted next-generation sequencing (tNGS) spoligotyping, and compared its outputs with those of membrane-based spoligotyping. A total of 144 clinical MTBC strains were retrospectively selected to be representative of the local epidemiology. Data analysis of a training set allowed for the setting of "presence"/"absence" thresholds for each spacer to maximize the sensibility and specificity related to the membrane-based spoligotyping. The thresholds above, in which the spacer was considered present, were 50 read per millions for spacers 10 and 14, 20,000 for spacers 20, 21, and 31, and 1000 for the other spacers. The confirmation of these thresholds was performed using a validation set. The overall agreement on the training and validation sets was 97.5% and 93.8%, respectively. The discrepancies concerned six strains: Two for spacer 14, two for spacer 31, and two for spacer 32. The tNGS spoligotyping, whose thresholds were finely-tuned during a careful bioinformatics pipeline development process, appears be a technique that is reliable, inexpensive, free of handling errors, and automatable through automatic transfer into the laboratory computer system.

Consistency of Mycobacterium tuberculosis Complex Spoligotyping between the Membrane-Based Method and In Silico Approach.

To tackle the spread of tuberculosis (TB), epidemiological studies are undertaken worldwide to investigate TB transmission chains. Clustered regulatory interspaced short palindromic repeats (CRISPR) locus diversity, also called spoligotyping, is a widely used genotyping assay for the characterization of Mycobacterium tuberculosis complex (MTBC). We compared herein the spoligotyping of MTBC clinical isolates using a membrane-based method (following an initial PCR step) and whole-genome sequencing (WGS)-based method (i.e., in silico spoligotyping). All MTBC strains isolated at the Lyon University Hospital, France, between November 2016 and December 2020 were included (n = 597). Spoligotyping profiles were also used for species identification among the MTBC. Outputs of both methods were analyzed, and discrepant results were investigated thanks to CRISPRbuilder-TB. The overall agreement was 85.7%. Spacer discrepancies observed between the methods were due to the insertion of IS6110 within the direct repeat (DR) sequence upstream or downstream of spacers, mutated DR sequences, or truncated spacers. Discrepancies did not impact species identification. Although spoligotyping-based species identification was inconclusive for 29 isolates, SNP-based phylogeny conducted after WGS allowed the identification of 23 M. tuberculosis (Mtb), 2 M. canettii, and 4 mixed MTBC infections. WGS yielded very few discrepancies compared to membrane-based spoligotyping. Overall agreement was significantly improved (92.4%) by the CRISPR locus reconstruction using CRISPRbuilder-TB for the MTBC isolates with the shared international type 53 in silico spoligotyping. A smooth transition from the membrane-based to the in silico-based genotyping of M. tuberculosis isolates is, therefore, possible for TB diagnosis and epidemiologic survey. IMPORTANCE Whole-genome sequencing (WGS) has profoundly transformed the perspectives of tuberculosis (TB) diagnosis, providing a better discriminatory power to determine relatedness between Mycobacterium tuberculosis complex (MTBC) isolates. Previous genotyping approaches, such as spoligotyping consisting of an initial PCR step followed by reverse dot hybridization, are currently being replaced by WGS. Several pipelines have been developed to extract a spoligotype from WGS data (in silico spoligotyping) allowing for the continuity of MTBC molecular surveys before and after WGS implementation. The present study found very good overall agreement between hybridization to membrane-based spoligotyping and in silico spoligotyping, indicating the possibility of a smooth transition from the traditional to the in silico-based genotyping of MTBC isolates for TB diagnosis and epidemiological survey.