The plastic genome of Bordetella pertussis

  • Jonathan Simon Abrahams

Student thesis: Doctoral ThesisPhD

Abstract

The paradigm that single nucleotide polymorphisms (SNPs) are the primary metric to judge bacterial diversity is outdated. This is particularly true for the main causative agent of whooping cough, Bordetella pertussis- a species with limited nucleotide variation. Examined in a holistic way, however, B. pertussis has high potential for genetic diversity with over 250 copies of the same insertion sequence- perfect genetic material for structural variations to arise via homologous recombination. Indeed, many deletions and inversions have been described which is in contrast to the third type of structural variation: CNVs (copy number variations), which have been only infrequently described.In this thesis, I systematically investigated the prevalence and dynamics of CNVs (and other structural variants) in B. pertussis using both long and short read sequence data.I developed a reliable pipeline to predict CNVs in >2000 isolates by analysing the read-depth of Illumina sequencing samples to find regions of increased coverage. A low rate of false positives and negatives was achieved by normalising inter-sample noise. The majority of these mutations were predicted to be >50kb long and clustered at 11 hotspot loci (rather than evenly distributed throughout the genome), a phenomenon described and analysed using network graphs. One CNV was verified by qPCR and by capturing entire tandem arrays of CNVs in single ultra-long reads generated on the Oxford Nanopore sequencing platform. Further investigation demonstrated the plasticity of the B. pertussis genome and it was found that multiple putative structural variants were being generated genome-wide within a single culture. Finally, preliminarily work established the compatibility of the B. pertussis with the Genome Wide Association (GWA) framework. I investigated how to represent CNVs and how homoplasic deletions were, given the highly clonal nature of the species and its low mutation rate. . It was found that deletions were more homoplasic than previously thought but that there are still considerable hurdles to using CNVs in GWAS.
Date of Award22 Oct 2020
Original languageEnglish
Awarding Institution
  • University of Bath
SponsorsPublic Health England
SupervisorAndrew Preston (Supervisor) & Ruth Massey (Supervisor)

Cite this

'