Rapid host switching in generalist Campylobacter strains erodes the signal for tracing human infections – Characterising zoonosis – Modelling zoonosis

View Github Repository Open presentation in a new window

bdearlove

See all presentation from bdearlove

Rapid host switching in generalist Campylobacter strains erodes the signal for tracing human infections – Characterising zoonosis – Modelling zoonosis

0 0

notMolEpi2015

On Github bdearlove / notMolEpi2015

Rapid host switching in generalist Campylobacter strains erodes the signal for tracing human infections

Bethany Dearlove

University of Cambridge

 @bethanydearlove

Campylobacter

  • Major cause of bacterial gastroenteritis.
    • Causes severe diarrhoea, vomiting and abdominal pain.
    • Symptoms last 7-10 days.
    • Can lead to more serious diseases such as Guillain-Barré syndrome and reactive arthritis.

Increasing incidence

Data source: HPE

Transmission

  • Common transmission routes:
    • Undercooked meat and unpasteurised milk
    • Untreated water
  • Human-to human-transmission rare

Zoonosis

  • Range of wild and domestic hosts
  • Asymptomatic carriage in the gut microbiota.
  • Contamination can occur any time from farmyard to fork.
      by Mauricio Duque, snap2objects

Characterising zoonosis

  • MLST analyses have shown that:
    • Populations in different host species tend to be genetically isolated, despite shared geography.
    • However, some of the most common disease-causing strains are isolated from multiple host species.

Characterising zoonosis

Sheppard et al. (2013). PNAS.

Characterising zoonosis

  • Genuine generalists?
  • Just the limited resolution of MLST?

Aim:

To investigate the utility of whole-genome sequences for

  • estimating the rate of zoonosis
  • improving the accuracy of source attribution for clinical cases

in these clonal complexes.

Challenges

  • Zoonosis model
  • Recombination
  • Computation

Modelling zoonosis

  • Phylogeography model (Lemey et al., 2009)
  • If adapted in the past, would expect isolates from same host population to cluster together.

Phylogeography

  by Mauricio Duque, snap2objects

Transitions between hosts

Human = chicken or pig or cattle

Transitions between nucleotides

N = A or C or G or T

Accounting for ancestral recombination

  • Substitution model
    • Relaxed clock
    • Gamma site heterogeneity
  • Removal of homoplasies
    • Hedge and Wilson (2014)

ST-45 complex

ST-21 complex

ST-828 complex

Parameter estimates

Estimated one jump every...

  • 1.6 years in ST-21
  • 1.8 years in ST-45
  • 12 years in ST-828

Parameter estimates

Source of clinical cases

Summary

  • Isolates from different host species are often more closely related than those isolated from the same host species.

  • As expected from previous MLST studies, much ancestry is inferred to occur within chickens.

    • MRCA of all three complexes.

Summary

  • Cannot properly attribute at the individual level in these complexes due to weak host signal.

    • Most likely due to generalist nature of these complexes, and uncertainty reflects make up of host population.
    • Signal could be drowned out or in accessory genome.

  • Evidence of potential for whole genomes

    • BUT the extra information from WGS does not overcome the need for detailed sampling.

Acknowledgements

Daniel Wilson

Samuel Sheppard

Ben Pascoe

Guillaume Meric

Alison Cody