/timtam

Code to carry out the computational experiments described in the manuscript "A computationally tractable birth-death model that combines phylogenetic and epidemiological data"

Primary LanguageHaskellBSD 3-Clause "New" or "Revised" LicenseBSD-3-Clause

BDSCOD

*SEE TIMTAM2 FOR THE BEAST2 IMPLEMENTATION*

This repository contains an implementation of the approximate likelihood of a birth-death process with various observation processes described here. For a description of the code that generated the figures in that manuscript see here. An overview of the organisation of the code in this repository is given below.

Architecture

The diagram below shows the main source and documentation files in this repository. The computation is split up into several examples, for more details on the examples see the README. The main directories are

  • src which contains the library code which implements the likelihood calculation, ie the code you might want to look at if you wanted to apply the method to your own data,
  • apps which contains the application code for each example, ie the code specific to the examples presented in the manuscript linked above,
    • The mcmc application can be used to run an analysis from the command line with data and configuration read from a single JSON file.
  • and examples which contains the documentation and scripts needed to run each example, ie the scripts you will need to run on your machine to replicate the results.
    • The ape-simulation example is particularly important, it includes the script ape-sim.R which provides a wrapper around some R packages to make it easier to simulate suitable datasets.

.repo-diagram.svg

Acknowledgements

  • This project was done in collaboration with my co-authors: Louis du Plessis, Kris V. Parag, and Oliver G. Pybus.
  • Thank you to Yunjun Zhang from Peking University for reading through and helping to improve documentation.