GitHub - emvolz/PhyDyn-simulations (original) (raw)

This repository is for simulation experiments to evaluate performance of the PhyDyn BEAST2 package. We simulate a SIRS infectious disease epidemic such that transmission rates vary over the course of infection and between risk groups. Simulations are individual-based, stochastic, and continuos time using a Gillespie exact algorithm implemented in Python. A genealogy of the transmission history is reconstructed by subsampling lineages over time. PhyDyn is then used to re-estimate parameters. For a description of the model including equations, see the manuscript: https://www.biorxiv.org/content/early/2018/04/10/268052

Note the following

• This simulation test is based on genealogies, not on simulated sequences
• BEAST is not used to simulate data, only to re-estimated parameters. We simulate epidemics using a python script and use an R script to sample genealogies.

Simulation sets are included:

1. 2000 initial susceptible individuals and sampling 250 lineages over a long period of time.
2. 5000 initial susceptible, sampling 500 over a short period of time.

In both cases, sampling is heterochronous following epidemic peak.