FarmR

Estimating farm-to-farm disease transmission using pathogen genetic and animal movement data

Concept

We integrate transmission tree analysis (https://github.com/xavierdidelot/TransPhylo) with social network analysis to uncover the livestock farm-level disease transmission link. Series of pathogen genetic data collected from multiple farm outbreaks along with corresponding animal movement data are recommended to be an input of this analysis.

Prerequisites

Time resolved phylogenetic tree must be created from the well-aligned pathogen genetic sequences via BEAST software (https://beast.community/).

Input data; example file

1. Time resolved phylogenetic tree (NEXUS format from BEAST MCC tree) of pathogen's genetic sequences; beastmcc.tree
2. Metadata of the sequence samples including at least ID of isolates, sampling date, and farm premises; metadata.csv
3. Animal movement data comprising origin (tail), destination (head), and date of shipment; movement.csv

Optional data for ERGMs analysis

1. Metadata spreadsheet may include farm production type, herd size, or farm's geographical coordinates
2. All other farms' coordinates in the area where the samples were collected may be used for farm's spatial point density calculation; locationall.csv

Getting started

1. Open "Restimation-ergm.R"
2. Install all required packages (first 10 lines)
3. At line 12 of the script: change "x" to "path to your working directory" that keeps the input files
4. Run them all!

Primary outputs

1. Transphylo's transmission tree: MCMC diagnostics (trace & ESS) and colored transmission tree
2. Infection chain length (ICL; animal level) vs Movement path-length (MPL; farm-level): Normality test, Pearson's correlation, and linear regression plot
3. Farm-level reproduction number (R) estimation: Boxplot of ICL ~ MPL with the cut-off value at MPL = 1 and statistic summary of the farm-level R

Farm-to-farm transmission network (Additional outputs for ERGMs analysis)

1. Dyads: dyadic relationship between all samples, including transmission link (1,0), movement path-length (steps), sampling interval (days), and distance between farms (km)
2. Node attributes: same as the metadata file plus sampling season, farm's spatial point density, and in-degree and out-degree of the 6-month period (according to sampling date) movement network

What's ERGMs? How to use it?

https://cran.r-project.org/web/packages/ergm/ergm.pdf