Rapid comparative genomic analysis for clinical microbiology: The Francisella tularensis paradigm

Abstract

It is critical to avoid delays in detecting strain manipulations, such as the addition/deletion of a gene or modification of genes for increased virulence or antibiotic resistance, using genome analysis during an epidemic outbreak or a bioterrorist attack. Our objective was to evaluate the efficiency of genome analysis in such an emergency context by using contigs produced by pyrosequencing without time-consuming finishing processes and comparing them to available genomes for the same species. For this purpose, we analyzed a clinical isolate of Francisella tularensis subspecies holarctica (strain URFT1), a potential biological weapon, and compared the data obtained with available genomic sequences of other strains. The technique provided 1,800,530 bp of assembled sequences, resulting in 480 contigs. We found by comparative analysis with other strains that all the gaps but one in the genome sequence were caused by repeats. No new genes were found, but a deletion was detected that included three putative genes and part of a fourth gene. The set of 35 candidate LVS virulence attenuation genes was identified, as well as a DNA gyrase mutation associated with quinolone resistance. Selection for variable sequences in URFT1 allowed the design of a strain-specific, highly effective typing system that was applied to 74 strains and six clinical specimens. The analysis presented herein may be completed within approximately 6 wk, a duration compatible with that required by an urgent context. In the bioterrorism context, it allows the rapid detection of strain manipulation, including intentionally added virulence genes and genes that support antibiotic resistance.