Exhaustive reconstruction of the CRISPR locus inMycobacterium tuberculosiscomplex using short reads

Abstract

AbstractSpoligotyping, a graphical partial display of the CRISPR locus that can be producedin vitroorin silico, is an important tool for analyzing the diversity of givenMycobacterium tuberculosiscomplex (MTC) isolates. As other CRISPR loci, this locus is made up of an alternation between direct repeats and spacers, and flanked bycasgenes. Unveiling the genetic mechanisms of its evolution requires to have a fairly large amount of fully reconstructed loci among all MTC lineages.In this article, we point out and resolve the problem of CRISPR reconstruction based on short read sequences. We first show that more than 1/3 of the currently assembled genomes available for this complex contain a CRISPR locus erroneously reconstructed, and errors can be very significant. Second, we present a new computational method allowing this locus to be reconstructed extensively and reliablyin silicousing short read sequencing runs. Third, using this method, we describe new structural characteristics of CRISPR locus by lineages. We show how both the classical experimentalin vitroapproach and the basicin silicospoligotyping provided by existing analytic tools miss a whole diversity of this locus in MTC, by not capturing duplications, spacer and direct repeats variants, and IS6110insertion locations. This description is extended in a second article that presents general rules for the evolution of the CRISPR locus in MTC.This work opens new perspectives for a larger exploration of CRISPR loci diversity and of mechanisms involved in its evolution and its functionality.