Sensitive, specific polymorphism discovery in bacteria using massively parallel sequencing

Abstract

This variant ascertainment algorithm, or VAAL, uses short sequence reads of haploid bacterial genomes to first locally assemble the reads and then compare these assemblies to the reference genome. This allows VAAL to detect all types of variants ranging from single-nucleotide polymorphisms to large insertions or deletions. Our variant ascertainment algorithm, VAAL, uses massively parallel DNA sequence data to identify differences between bacterial genomes with high sensitivity and specificity. VAAL detected ∼98% of differences (including large insertion-deletions) between pairs of strains from three species while calling no false positives. VAAL also pinpointed a single mutation between Vibrio cholerae genomes, identifying an antibiotic's site of action by identifying sequence differences between drug-sensitive strains and drug-resistant derivatives.