DOI: 10.5061/dryad.p8cz8w9xd

This dataset was created for a study of how the grain size of clustering of environmental heterogeneity could alter how like plastic generalism is to evolve using computer simulation models. All of the code for simulations was written in R and consists of two main experiments.

In one experiment (using the “clustering simulation script.R”), random landscapes were generated with variable levels of clustering to evaluate how the likelihood of plastic generalism arising on landscapes with a variety of clustering levels changed when populations were already located across the landscape vs. when they were expanding through the landscape. This experiment showed that in settled populations, the likelihood of plastic generalism evolving decreased with increasing grain size of clusters. However, in expanding populations coarse-grained landscapes led to plastic generalism being much more likely to evolve. This surprising result led to the second experiment.

In the second experiment (using “stripes simulation script.R”), landscapes with alternating stripes of different types of environments were generated to evaluate if the increase in evolution of plastic generalism in coarse-grained landscapes was because alternating blocks of the environment created impassable barriers that forced incremental evolutionary steps. Stripe widths varied from thin stripes that organisms could disperse across to wide stripes that they had to evolve to be able to use to cross. In one treatment, populations were initially placed so that expansion would happen parallel to stripes while in the other they were placed so expansion would happen perpendicular to stripes. These different treatments were designed so that the parallel one (“with the grain” or wg) showed how the population would evolve when alternating environments were not an explicit barrier to expansion and the perpendicular one (“against the grain” or atg) would show how the population would evolve when alternating environment were a barrier to expansion. This treatment showed that alternating blocks of environments being a barrier to range expansion was the component that increased the likelihood of plastic generalism evolving, with perpendicular treatments leading to many more replicates evolving plastic generalism.

This repository contains the code used to run simulations and generate results (all written in R), the code used to generate figures shown in the paper associated with this repository (all written in R), and the results from simulations (as text files and jpegs).

The code used for the simulations is all written in R, and can be opened in R or in any text editing window. It can be run in R.

For the first experiment, the code used to generate the simulations is in the file titled “clustering_simulation_script.R”. This file uses functions located in the file “functions_for_clustering_sims.R” by calling the file as source code on the 79th line. The pathname for the source code will need to be changed to reflect the files location on each individual users computer to access it. At the beginning of the script (line 2) there is also a line to set the working directory to direct the data files that are output to the appropriate location on a users computer. Lines 15-36 allow setting parameters that did not vary in the results presented in the accompanying paper, line 39 allows you to set the number of replicates. Line 43 allows users to alter whether a set of replicates will have the range expansion (“re”) or no range expansion (“nr”) treatment, and line 49 allows users to set the grain size of landscape for a treatment (with eligible options shown in lines 121-130). Once these parameters are set, the entire script can be run at once and will output the number of replicates set in the treatment that has been set.

Results from the first experiment can be found in the zipped files that start are named with the format “theta_#_treatment.zip”, where # can be 0.5, 2, 6, 24, 43, 90, 175, or 370 and treatment can be re or nr. Each zipped file is a folder that contains several text files for each replicate containing genotype information and a jpeg of the end state of the landscape in that replicate. These data files were processed using “Analysis_of_data.R” script to create summary files that indicate whether plastic generalism arose in each replicate and when it arose. These summary files can be found in the zipped file “clusters_gen_emerge.zip” which is a folder of text files. These summary files were used to create figure 2 in the accompanying paper using the script “Figure_2.R” to generate images in R. Figure 3 was also generated using the data from this experiment in R, using the script “Figure_3.R”. This used summaries of the number of steps to generalism that can be found in each treatment’s individual zipped folder and randomly selected replicates that evolved generalism to plot trajectories of, drawing from the data in individual treatments folders. Supplemental figure 2 was a similar figure to figure 3, and used the same pool of data and a similar method located in the “Supplemental_figure_2.R” script.

For the second experiment, the code used to generate the simulations is in the file titled “stripes_simulation_script.R”. This file is an R script that also uses the same source file as the first experiment, and has similar locations of changeable parameters. This script generated data located in zipped files titled “treatment_#.zip”, where treatment can be “wg” or “atg”, and # can be 3, 5, 7, 10, 11, 12, 13, 14, 15, 20, 30, or 40. Each of these zipped files is a folder containing several text files for each treatment and a jpeg for the state of the population at the end of the replicate. These results were also processed using the “Analysis_of_data.R” script, and the summary text files can be found in the zipped folder “stripes_gen_emerge.zip”. These data were used to generate figure 4 in the accompanying paper, using the script “Figure_4.R”.

In addition to the figures derived directly from results, there is a script called “Figure_1.R” which is a script that generates a conceptual figure that was in the methods. There is also a script called “SF1_making_blank_landscapes.R” that generated the first supplemental figure, which is a series of blank landscape subsections that allows readers to get a sense of what different levels of clustering actually look like.