README.md

In pratikunterwegs/patho-move-evol: Analysis code for the _pathomove_ model.

Source Code and Supplementary Material for Novel pathogen introduction triggers rapid evolution in animal social movement strategies

This is the source code for the analyses and figures for a manuscript that reports on a model of the evolution of social movement strategies, following the introduction of a novel pathogen to an animal population. This work was developed in the Modelling Adaptive Response Mechanisms Group (Weissing Lab) at the Groningen Institute for Evolutionary Life Science, at the University of Groningen.

Contact and Attribution

Please contact Pratik Gupte for questions on the model or the associated project.

Name: Pratik Rajan Gupte
Email: [e-mail]
ORCID: https://orcid.org/0000-0001-5294-7819

You can cite all versions of this repository, archived on Zenodo, by using the DOI 10.5281/zenodo.6334026. This DOI represents all versions, and will always resolve to the latest one.

Simulation Source Code

The simulation source code is provided in a different repository, pathomove. The pathomove simulation is archived on Zenodo. You can cite cite all versions of Pathomove by using the DOI 10.5281/zenodo.6331815.

Simulation Data

The simulation data are originally generated on the University of Groningen's Peregrine high-performance computing cluster (using the R code in the scripts/ folder: scripts/00_use_cluster.Rmd).

The data used for this the manuscript are available on Zenodo, and you can cite all versions by using the DOI 10.5281/zenodo.6331756. This DOI represents all versions, and will always resolve to the latest one.

Workflow: Running the Simulation

Warning: This is a relatively advanced computational study. Replicating it requires many interacting components. Please do not expect it to work out of the box.

A brief description of this workflow is:

1. Download and install the Pathomove simulation Rcpp package from https://github.com/pratikunterwegs/pathomove locally (on your computer).

2. Download this repository locally.

3. Run the simulation with the required parameter combinations. Parameter combinations used in this study are found in data/parameters/.

4. Collect the simulation output in data/output/, and analyse it using the scripts in scripts/. Analysis scripts are written to find the output in data/output/.

5. Prepare figures using the scripts in figure-scripts/.

Workflow: Replicating this Study

Note: In order to have completely reproducible simulations, it is necessary to run the simulation in single-threaded mode. Multi-threaded simualtion runs are not reproducible.

1. Download and install the Pathomove simulation Rcpp package from https://github.com/pratikunterwegs/pathomove using:

   Install directly from an R terminal without downloading the simulation package source code separately.

   ```r

   in R

   devtools::install_github("pratikunterwegs/pathomove") ```

   You should be prompted to also install the simulation dependencies, Rcpp, RcppParallel, and BH (Boost Headers).

2. Download this repository, using:

   ```sh

   in the shell

   git clone [e-mail]:pratikunterwegs/patho-move-evol.git ```

3. In this repository, prepare an R script that runs a single replicate of the simulation.

   This script should essentially run the commands

   ```r

   in R

   command to run simulation and save output

   data = pathomove::run_pathomove(...)

   save data as Rds object

   saveRDS(data, file = "data/output/output_N.Rds") ```

   This script should be able to take command line arguments, including at least: (1) the name of a file containing parameter combinations, and (2) the row number in the parameter file, indicating which combination should be run.

   An example of such a script (which is also used in our analyses) is scripts/do_sim_pathomove.R. This script should specify where to save the simulation output, and the file type. We recommend saving .Rds (R data) files to data/output.

At this stage, there are two options:

• Running simulation replicates locally (on your computer), or,
• Running simulation replicates on a high-performance computing (HPC) cluster.

Replicating Analyses: Running Simulation Replicates Locally

Continuing from (3.):

1. Prepare a file with parameter combinations, and a shell script that instructs R to run the script from (3.), taking as arguments the parameter file, and a row number.

   • The same procedure applies for multiple replicates of the same parameter combination.

   • An example of a script which automates these processes, for Windows systems, is scripts/00_make_batch_script.R.

   • This script generates .bat batch files. Run the batch file.

   • Simulation output in the form of Rds files should be saved to data/output, or some other path that you have chosen.

Replicating Analyses: Running Simulation Replicates on an HPC Cluster

Continuing from (3.):

1. Prepare a file with parameter combinations.

   • The same procedure applies for multiple replicates of the same parameter combination.

2. Prepare a directory structure to store the output. A template directory structure can be found at https://github.com/pratikunterwegs/patho-move-evol (this repository).

   There should be at least the following paths:

   ```sh

   from the shell, upon running tree (or equivalent)

   yourFolder ├───bash ├───data │ ├───output │ ├───parameters └───scripts ```

3. Prepare a template job. An example is found in bash/main_job_maker.sh. This script is written for an Ubuntu-based, SLURM-scheduler HPC cluster.

4. Run the following commands locally from R. The specific commands, and parameter combinations used in this study are found in scripts/00_use_cluster.Rmd.

   ```r

   this should be your R terminal

   be careful about working directories etc.

   load the package locally

   library(pathomove)

   make a parameter file with all the combinations required

   or with multiple replicates

   pathomove::make_parameter_file( ..., replicates = N, which_file = "" )

   above, ... indicates the simulation parameters

   use the use_cluster function to send in jobs

   try sending in a job

   pathomove::use_cluster( ssh_con = "[e-mail]", password = password, script = "scripts/do_sim_pathomove.R", # e.g. scripts/do_sim_pathomove.R tag = scenario_tag, folder = "patho-move-evol", # folder for the output, must contain data/ template_job = "bash/main_job_maker_default.sh", # the shell script from (5) parameter_file = param_file_default # the parameter data ) ```

5. Simulation output should be returned as Rds files into the data/output folder specified above on the cluster, or your custom equivalent. Move these Rds files to your local system for further analysis.

Workflow: Analysing Output

1. Gather simulation output data in the form of Rds files in the data/output/ folder.

2. Process the output by running, in sequence the analysis source code described below.

Analysis Source Code

The source code for the analyses reported here can be found in the directory scripts/, and are explained briefly here:

• scripts/01_process_eco_evo_data.Rmd: Process the output, in the form of Rds objects, that result from running pathomove replicates or parameter combinations.

• scripts/02_process_networks.Rmd: Process the pairwise individual associations logged during the simulation into social networks.

  • These two steps create the directory structure:

  ```sh

  from the shell, upon running tree

  . ├── data │   ├── 00_data.txt │   ├── output │   │   └── 00_output_exists.txt │   ├── parameters │   │   ├── 00_parameters.txt │   └── results │   ├── 00_data.txt │   ├── gen_data │   ├── morph_data │   ├── networks │   └── si_imp_data ```

  • gen_data/ holds data on the individuals in each generation stored from the simulation.
  • morph_data/ holds data on the proportions of the social movement strategies in each stored generation.
  • si_imp_data/ holds data on the importance of social information to movement strategies in each stored generation.

  • move_assoc_data/ holds data on the number of associations per movement distance in each stored generation.

    Across these data, there is one .csv file per simulation replicate.

  • networks/ holds .Rds files which store the social networks emerging over the simulation; each file is for one simulation replicate, and holds a list object whose components are tidygraph network objects.

• scripts/03_sir_models.Rmd: Run SIR models on the emergent social networks acquired from simulation runs.

Helper scripts

Used in previous steps.

• scripts/00_use_cluster.Rmd: Passes the simulation run commands to the University of Groningen's Peregrine high-performance computing cluster. May also work with HPC cluster running Ubuntu, with required libraries installed --- this includes Boost and potentially Intel's TBB --- and with a SLURM-scheduler. Use with caution.

• scripts/00_make_batch_script.R: An alternative to using an HPC cluster, written for Windows systems. Makes a batch script and parameter set to run simulations in sequence. Use with caution. Can be easily adapted to working with Linux systems.

Analysis Functions

Note that analysis function provided in this repo in earlier versions have now been moved to the simulation package repository; https://github.com/pratikunterwegs/pathomove.

Figure Source Code

The source code for the figures in this manuscript is in the directory figure_scripts/. These scripts are well commented, and are not explained further.

Manuscript Text

The main text of the manuscript is written in LaTeX and is stored in the (private) submodule, ms-pathomove. Using the shell scripts provided in bash/, the LaTeX files are converted into date-stamped PDFs. These are not uploaded here, but the docs/ folder indicates their storage location.

Supplementary Material

The supplementary material provided with this manuscript is generated from the supplement.Rmd file in the supplement/ directory.

Other files in this directory are helper files required to format the supplementary material.

Other Directories

• bash/ Some useful shell scripts for output rendering.

pratikunterwegs/patho-move-evol documentation built on April 15, 2023, 5:54 p.m.