P019

1:San Raffael Scientific Institute; 2:Statens Serum Institut

Targeted Next-Generation Sequencing (tNGS) of Mycobacterium tuberculosis (MTB) is approved by the World Health Organization for direct drug-resistant tuberculosis (DR-TB) testing of respiratory specimens. The new Deeplex Myc-TB XL assay (Genoscreen) targets several DR-associated loci, including novel key drugs as bedaquiline and pretomanid. Direct whole genome sequencing (dWGS) emerged as an innovative approach for full DR profiling and epidemiological surveillance without prior target selection. The new QIAseq xHYB MTB panel (Qiagen) enables dWGS through hybridization-based MTB genome capture. This study evaluates the detection limits (LOD) and performance for DR and genotyping predictions of the two assays.

We analyzed 89 MTB-positive sputum sediments from a TB vaccine trial with prevention of recurrence design. DNA was extracted and host DNA depleted using the Ultra-Deep Microbiome Prep (Molzym), then quantified by real-time PCR. Deeplex and QIAseq libraries were sequenced on Illumina systems. FASTQ files were processed with adapted MTBseq pipeline and benchmarked against WGS from corresponding cultures.

MTB genome copies (gc) quantification from extracted samples ranged from <10 to >1000 gc/uL. Deeplex yielded complete DR profiles (coverage depth >200x) in 97% of specimens, with LOD of 10 total gc. QIAseq produced high-quality genome coverage (depth >30x; breadth >97%) in 60% of samples, enabling transmission analysis consistent with that of culture-based WGS (concordant clustering for all but one sample). DR profiling was successful in 74% of samples; LOD was >100 total gc. Both platforms achieved >98% sensitivity/specificity for DR mutations identification.

Bacillary load-guided selection of tNGS or dWGS would improve culture-free DR-TB sequencing efficiency.