We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

q-bio.PE

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Quantitative Biology > Populations and Evolution

Title: A Bayesian model of acquisition and clearance of bacterial colonization

Abstract: Bacterial populations that colonize a host play important roles in host health, including serving as a reservoir that transmits to other hosts and from which invasive strains emerge, thus emphasizing the importance of understanding rates of acquisition and clearance of colonizing populations. Studies of colonization dynamics have been based on assessment of whether serial samples represent a single population or distinct colonization events. A common solution to estimate acquisition and clearance rates is to use a fixed genetic distance threshold. However, this approach is often inadequate to account for the diversity of the underlying within-host evolving population, the time intervals between consecutive measurements, and the uncertainty in the estimated acquisition and clearance rates. Here, we summarize recently submitted work \cite{jarvenpaa2018named} and present a Bayesian model that provides probabilities of whether two strains should be considered the same, allowing to determine bacterial clearance and acquisition from genomes sampled over time. We explicitly model the within-host variation using population genetic simulation, and the inference is done by combining information from multiple data sources by using a combination of Approximate Bayesian Computation (ABC) and Markov Chain Monte Carlo (MCMC). We use the method to analyse a collection of methicillin resistant Staphylococcus aureus (MRSA) isolates.
Comments: Machine Learning for Health (ML4H) Workshop at NeurIPS 2018 arXiv:1811.07216
Subjects: Populations and Evolution (q-bio.PE); Machine Learning (cs.LG); Applications (stat.AP)
Report number: ML4H/2018/87
Cite as: arXiv:1811.10958 [q-bio.PE]
  (or arXiv:1811.10958v1 [q-bio.PE] for this version)

Submission history

From: Marko Järvenpää [view email]
[v1] Tue, 27 Nov 2018 13:17:44 GMT (238kb,D)

Link back to: arXiv, form interface, contact.