adamkucharski/covid-import-model

Modelling variant importations

covid-import-model

Preliminary modelling analysis of variant importations and onwards community transmission. Currently implementation uses adaptive Metropolis Hastings MCMC to estimate paramaters using a deterministic approximation of a continuous time branching process model.

Note: this is working repository, so code and data are likely to change over time

Quick start guide

1. Get required data file from COG-UK and place in a local folder: Download cog_metadata.csv

2. Define file paths:

   • Set GitHub directory, e.g. setwd("~/Documents/GitHub/covid-import-model")

   • Specify local data path with above files stored, e.g. data_path <- "~/Documents/COVID_data/B_617_2/"

3. Main model script in scripts/main_run.r, including dependencies and data loading. This calls the following R files:

   R/load_data.R - Load data from covidregionaldata, COG-UK and India sequences (stored in folder defined above).

   R/model_functions.R - Main model simulation functions.

   R/model_mcmc.R - MCMC inference functions.

   R/load_posteriors.R - Load and format MCMC posteriors.

   R/plot_outputs.R - Plot data, parameter estimates and model simulations.

The model also calls outbreakinfo_mutation_report_data.tsv, downloaded on 21st May from the lineage prevalence visualisation here at outbreak.info.

Archived code

An MLE framework used to implement an earlier version of this analysis is archived in V1_code/.

Citation

Reference for initial version of the analysis: Kucharski et al, CMMID Working Group Report, 12 May 2021.

If you plan to build on or cite this preliminary analysis for an academic publication, please ensure that you credit the underlying data sources above appropriately.