Detection of minority variants and mixed infections in Mycobacterium tuberculosis by whole genome sequencing using a specific-DNA capture strategy: A pilot study

Detection of mixed Mycobacterium tuberculosis infections is essential, particularly when resistance mutations are present in minority bacterial populations and may affect patient´s evolution and treatment. Whole genome sequencing has extended the amount of key information available for the diagnosis of Mycobacterium tuberculosis. The relevance of having genomic information at diagnosis for early intervention requires carrying out whole genome sequencing directly on the clinical samples. However, few studies have been successful with this approach due to the low representation of MTB-DNA in sputa. In this study, we evaluated the ability of a strategy based on specific MTB-DNA enrichment to detect minority variants and mixed infections by whole genome sequencing on controlled mixtures of MTB-DNAs in spiked sputa. A pilot study was carried out with 12 samples containing 98% of a DNA pool from sputa of non-tuberculosis patients and 2% of MTB-DNA mixtures at different proportions. Our strategy allowed us to obtain sequences with a quality equivalent to those obtained from culture: 62.5X depth coverage and 95% breadth coverage (for at least 20X reads). Assessment of minority variant detection was carried out by manually directed analysis and allowed to identify heterozygous positions up to 95:5 ratio. The strategy also distinguishes automatically mixed infections up to a proportion of 90:10. This strategy efficiently captures MTB-DNA in a non-specific genetic background, allows detecting minority variants and mixed infections, and has proven to be a promising tool for performing whole genome sequencing directly on clinical samples.