CRISPRbuilder-TB, reconstructing the CRISPR-Cas Locus of Mycobacterium tuberculosis complex using short reads archives (SRA) and tracking genome evolution

International audienceCRISPR-cas locus in M. tuberculosis has long been considered as a mere bar code for approximate classification. Now that this locus seems to interfere with physiology as shown by its functionality in a heterologous system and potential implication in resistance acquisition and interaction with host cells (Wei 2019a, Wei 2019b, Mo, 2021), it becomes important to track its precise evolutionary patterns. CRISPR loci are difficult to reconstruct using Short Read Archives especially with short reads, as evidenced by the low reliability of a large proportion of CRISPR-Cas loci from assembled genomes deposited in Public databases. Until the advent of reliable long-read sequencing technologies, and to ease the analysis of any CRISPR-Cas structure, we propose a new tool, CRISPRbuilder-TB to reconstruct the CRISPR-Cas Locus of Mycobacterium tuberculosis complex using Short Reads Archives (SRA). This software, using a de Bruijn graph approach, is computationally optimized, allowing to treat one dataset in less than 5 min on standard laptops. CRISPRbuilder-TB is freely available on GitHub with adequate tutorials. Using this new software, we show that CRISPR loci are prone to SNPs, duplications, deletions including deletions related to IS6110 recombinations. Using a home made tool soon to be releases, TBtools-viewer, we characterize known and new convergence events in CRISPR locus