Inferring spatiotemporal dynamics of the H1N1 influenza pandemic from sequence data


In 2009, a variant of influenza H1N1 emerged through reassortment of bird, swine and human flu viruses, with surface proteins HA and NA of swine origin. Owing to a lack of previous immunity to these proteins, this virus spread rapidly through the human population, causing a worldwide pandemic. Here, we will investigate the growth of the virus population and its global spread using publicly available sequence data. This tutorial walks through how to use BEAST and associated software to infer spatiotemporal dynamics from viral sequence data. Accompanying slides on the theory behind this analysis are available in theory/slides.pdf.

Required software

  • BEAST is used to infer evolutionary dynamics from sequence data.
  • BEAGLE is a helper library that allows faster and more advanced functions to be run in BEAST. For this practical, it is not necessary to install CUDA drivers (step 2 in the BEAGLE installation).
  • Tracer is used to analyze parameter estimates from BEAST.
  • FigTree is used to analyze phylogeny estimates from BEAST.
  • Google Earth is used to display phylogeographic reconstructions.


  1. Compile sequence data
  2. Prepare a skyline analysis
  3. Run the skyline analysis
  4. Examine the skyline output
  5. Examine the skyline tree
  6. Prepare a logistic growth analysis
  7. Run the logistic growth analysis
  8. Examine the logistic growth output
  9. Prepare a phylogeographic analysis
  10. Run the phylogeographic analysis
  11. Examine the phylogeographic output

Next section