Exact mapping of Illumina blind spots in the Mycobacterium tuberculosis genome reveals platform-wide and workflow-specific biases

Abstract

Whole genome sequencing (WGS) is fundamental to M. tuberculosis basic research and many clinical applications. Coverage across Illumina-sequenced M. tuberculosis genomes is known to vary with sequence context, but this bias is poorly characterized. Here, through a novel application of phylogenomics that distinguishes genuine coverage bias from deletions, we discern Illumina "blind spots" in the M. tuberculosis reference genome for seven sequencing workflows. We find blind spots to be widespread, affecting 529 genes, and provide their exact coordinates, enabling salvage of unaffected regions. Fifty-seven PE/PPE genes (the primary families assumed to exhibit Illumina bias) lack blind spots entirely, while remaining PE/PPE genes account for 55.1% of blind spots. Surprisingly, we find coverage bias persists in homopolymers as short as 6 bp, shorter tracts than previously reported. While GC-rich regions challenge all Illumina sequencing workflows, a modified Nextera library preparation that amplifies DNA with a high-fidelity polymerase markedly attenuates coverage bias in GC-rich and homopolymeric sequences, expanding the "Illumina-sequencable" genome. Through these findings, and by defining workflow-specific exclusion criteria, we spotlight effective strategies for handling bias in M. tuberculosis Illumina WGS. This empirical analysis framework may be used to systematically evaluate coverage bias in other species using existing sequencing data.