Rxivist logo

Rapid statistical methods for inferring intra- and inter-hospital transmission of nosocomial pathogens from whole genome sequence data

By Marianne Aspbury, James Sciberras, Jukka Corander, Sion C Bayliss, Tjibbe Donker, Edward Feil, Richard James

Posted 19 Oct 2018
bioRxiv DOI: 10.1101/442319

Whole genome sequence (WGS) data for bacterial pathogens can provide evidence as to the source of nosocomial infection, and more specifically the ability to distinguish between intra- and inter-hospital transmission. This is currently achieved either through using SNP thresholds, which can lack statistical robustness, or by constructing phylogenetic trees, which can be computationally expensive and difficult to interpret. Here we compare two alternative statistical approaches using 1022 genomes of methicillin resistant Staphylococcus aureus (MRSA) clone ST22. In 71% of cases both methods predict the same hospital origin, which is also supported by the ML tree. Robust assignments are divided approximately equally between intra-hospital transmission and inter-hospital transmission. Our approaches are rapid and produce intuitive output that could inform on immediate infection control priorities, as well as providing long-term data on inter-hospital transmission networks. We discuss the strengths and weakness of our methods, and the generalisability of this approach.

Download data

  • Downloaded 438 times
  • Download rankings, all-time:
    • Site-wide: 87,836
    • In epidemiology: 3,835
  • Year to date:
    • Site-wide: 79,066
  • Since beginning of last month:
    • Site-wide: 149,848

Altmetric data

Downloads over time

Distribution of downloads per paper, site-wide