README: The impact of within-host coinfection interactions on between-host parasite transmission dynamics varies with spatial scale ? ?️

Overview ?

These data were collected in 2012 as part of a broader piece of work exploring the parasite communities of the wood mouse (Apodemus sylvaticus). This code base takes input data from trapping grids and calculates neighbourhood prevalence relative to each individual on the grid. It does this by iteratively identifying each individual in the population as the ‘focal’ individual, and each other as a potential neighbour, with a distance to the focal individual calculated. Then neighbourhood prevalence and mean infection burden of the potentially interacting parasite (the nematode Heligmosomoides polygyrus) are calculated for each user-defined neighbourhood size around the focal individual. These are then able to be used in statistical models as an explanatory factor for the specified infection state (Eimeria presence/absence or intensity) of all focal individuals.

Input Data Structure ?

The core data, included in raw data.csv includes:

• Animal identifier

• Temporal data in DD/MM/YYYY format

• X & Y grid coordinates

• Animal demographic data

• Parasitological data

And more specifically:

Output Data Structure - neighbourhood data generation

These files are titled:

DATA - Eapp_17.csv,DATA - Eapp_34.csv,DATA - Eapp_51.csv, DATA - Eapp_all.csv, DATA - Ehung_17.csv, DATA - Ehung_34.csv, DATA - Ehung_51.csv, DATA - Ehung_all.csv

Each file will have all of the following fields:

And files with "Ehung" in the file name will have:

While files with "Eapp" in the file name will have:

All files will have the following fields:

Output Data Structure - statistical model output files

These files are titled:

STATS - Eapp INF NHMeanburden 17.csv, STATS - Eapp INF NHMeanburden 34.csv, STATS - Eapp INF NHMeanburden 51.csv, STATS - Eapp INF NHMeanburden all.csv, STATS - Eapp INF NHprev 17.csv, STATS - Eapp INF NHprev 34.csv, STATS - Eapp INF NHprev 51.csv, STATS - Eapp INF NHprev all.csv, STATS - Eapp Int NHMeanburden 17.csv, STATS - Eapp Int NHMeanburden 34.csv, STATS - Eapp Int NHMeanburden 51.csv, STATS - Eapp Int NHMeanburden all.csv, STATS - Eapp Int NHprev 17.csv, STATS - Eapp Int NHprev 34.csv, STATS - Eapp Int NHprev 51.csv, STATS - Eapp Int NHprev all.csv, STATS - Ehung INF NHMeanburden 17.csv, STATS - Ehung INF NHMeanburden 34.csv, STATS - Ehung INF NHMeanburden 51.csv, STATS - Ehung INF NHMeanburden all.csv, STATS - Ehung INF NHprev 17.csv, STATS - Ehung INF NHprev 34.csv, STATS - Ehung INF NHprev 51.csv, STATS - Ehung INF NHprev all.csv, STATS - Ehung Int NHMeanburden 17.csv, STATS - Ehung Int NHMeanburden 34.csv, STATS - Ehung Int NHMeanburden 51.csv, STATS - Ehung Int NHMeanburden all.csv, STATS - Ehung Int NHprev 17.csv, STATS - Ehung Int NHprev 34.csv, STATS - Ehung Int NHprev 51.csv, STATS - Ehung Int NHprev all.csv

Which contain the statistical outputs of the models run with the file run.statistical.analysis.R. These files contain the following columns.

*Capture date is a second-order polynomial variable which is why there are two estimates for this.

Instructions for use ?‍?

There are three ways to recreate the figures in the paper:

1. from scratch, generating the neighbourhood data files (step #1 below), followed by running the statistical analysis code (step #2) and the plotting code to generate the figures (step #3);
2. use the previously-generated neighbourhood data files (“DATA -…csv”) and just run the statistical analysis code (step #2) and the plotting code to generate the figures (step #3);
3. use the previously-generated statistical output files (“STATS - …csv”) and just run the plotting code to generate the figures (step #3).

Step 1: Open file execute.parasite.function.R. This file takes raw data (described below) and generates neighbourhood data files that includes information on number of neighbours, and their infection statuses, for each focal capture in the raw data set. These neighbourhood data files ("DATA - …csv") are then used for the statistical analyses in step #2 below.

Lines 6 - 22 clears the environment and loads the required data file ("raw data.csv"), helper function files (exppoints.R, make.dist.vector.R, prop.on.grid.R) and package (gtools).

Lines 26 - 37 sets up distance matrices and calculates the range of distances to be used as neighbourhood sizes.

Lines 41 - 61 defines the neighbour and focal variables to be used and then prepares the dataframes accordingly.

Lines 65 - 94 defines column names for the output dataframes.

Lines 98 - 106 runs the parasite function across the selected temporal ranges.

Lines 110 - 119 write output files for these runs ("DATA - …csv").

Step 2: Open file run.statistical.analysis.R. This file runs Bayesian Generalised Linear Mixed Effects Models on the data generated by the file above to determine the effect of neighbour parasites on focal parasites at increasing spatial scales and specified time windows. Outputs are stored in csv files ("STATS - …csv"`) used to generate the plots in step #3 below.

Line 8 clears the environment.

Line 12 - 22 loads the data output from the neighbourhood analysis generated in the execute.parasite.function.R file ("DATA - …csv").

Line 26 - 28 loads the required helper function file (model.run.R) and statistical package (brms).

Line 33 sets the labels to be used which is a list of the distances which have been selected.

Lines 42 - 60 selects response and explanatory parasites and variable names.

Lines 65 - 85 runs the appropriate suite of models.

Lines 90 - 136 saves the output files for the models that have been run ("STATS - …csv").

Step 3: Open file combined.plotting.code.R. This file generates multipanel plots showing H. polygyrus neighbourhood effects on both E. hungaryensis (left-hand column) and E. apionodes (right-hand column) focal infection status (y-axis) against neighbourhood size (x-axis) for the different time windows (panels). Uses statistical output files (“STATS - …csv”) generated from Step 2.

Line 8 clears the environment.

Lines 12 - 58 loads the appropriate statistical output files generated in the run.statistical.analysis.R file (“STATS - …csv”).

Lines 64 – 73 creates functions to colour code points, depending whether the 95% CIs cross zero.

Line 78 sets the labels to be used which is a list of the distances which have been selected.

Lines 83 – 84 generate the output pdf with an appropriate name.

Lines 91- 94 sets plotting parameters (layout, margins etc).

Lines 97 – 162 creates a function (plotfunc) for creating the paired E. hungaryensis and E. apionodes plots for each time window in turn.

Lines 166 – 170 calls plotfunc for each time window in turn.

Lines 174 - 189 adds a single x-axis label for the whole plot and species labels for the two columns.

Lines 193 – 198 adds an appropriate single y-axis label, depending on the neighbourhood and focal variables.

Line 200 closes the pdf.