afrenoy/2ordercoop

Code and data for Frénoy et al 2017 (Second-order cooperation: cooperative offspring as a living public good arising from second-order selection on non-cooperative individuals)

Code and Data for Frénoy et al 2017, submited to Evolution

code_model: Matlab implementation of our model

• s2.m is the main file, it is a Matlab function which runs one replicate of evolution using a given random seed. Provides 2 kinds of output: allele frequencies at each time point (as a returned value) and optionally full spatial snapshot of the population at each time point (written in specified binary file)
• repls2.m runs several replicates of evolution, calling s2 several times, and saves the average output (allele frequency, with standard deviation) into a .mat file
• s2phenotypic.m is a variant of the model in which pure cooperation is a phenotypic switch (while in s2.m it is a genetic switch). repls2phenotypic.m runs several replicates of evolution for this alternative model
• s2invasion.m is a variant of the model in which we measure how much an initially rare allele can invades a population not containing this allele, repls2invasion runs several replicates simulations of invasion
• s2async.m is a variant of the model in which simulations are asynchronous: a time step corresponds to the random death of an individual, and competition in the neighborhood for reproduction and replacement of the dying individual

data: Contains all the simulation data

Each subfolder corresponds to a set of simulations. The correspondence between the figures and the simulations is given below:

data subfolder	figures	description
exp1	fig 1,2,3,S3	Main simulation: evolution of cooperation with evolvable mutation rate, high mutation rate is 0.01
exp1_ctl	fig 2	exp1 when enforcing a low mutation rate
exp1_cth	fig 2	exp1 when enforcing a high mutation rate
exp1_sterile	fig S6	exp1 when pure cooperators are sterile
exp1_phenotypic	fig S3	exp1 when cooperative behavior is controled by a non-heritable switch
exp1_liquid	fig S1	exp1 when neighborhood are randomized
exp1_async	fig S7	exp1 when the modeled is simulated asynchronously
exp2	fig S4	Similar to exp1, but high mutation rate is 0.1
exp3	fig S5	Similar to exp1, but high mutation rate is 0.5
invasion	fig S2	Simulations of invasion of a clonal population by a rare mutant allele

code_analysis: Matlab functions to produce all figures in the paper from the simulation data

These functions are commented, and all the input parameters needed to produce exactly all the figures presented in the paper (including supplementary material) from the content of the data directory are provided in a Makefile. Runing make all will produce all the figures using the content of the data directory.

code_runs: Bash scripts used to run the simulations (on a computing cluster) producing all the data presented in this work

These scripts rely on a function matlabcmd (defined in common.sh) that runs the matlab command given as argument. Such command depends on the computer platform used. We give 2 examples:

• launchmatlabcommandrnd.sh, a script that submits the matlab command as a job on a Grid Engine computing platform (such as the CC-IN2P3 platform that we used for this study)
• launchmatlablocal.sh, a script that directly runs the command using a local version of Matlab and should be modified according to the Matlab installation

To locally rerun our simulations and produce a new version of any subdirectory in the data directory, one should:

• Remove the content of this subdirectory (eg rm -rf data/exp1/*)
• Edit launchmatlablocal.sh to match the Matlab installation
• Run the script matching the data subdirectory (eg cd code_runs && ./exp1.sh)