Environmental stress can have a profound effect on inbreeding depression. Quantifying this effect is of particular importance in threatened populations, which are often simultaneously subject to both inbreeding and environmental stress. But while the prevalence of inbreeding-stress interactions is well known, the importance and broader applicability of such interactions in conservation are not clearly understood. We used seed beetles, Callosobruchus maculatus, as a model system to quantify how environmental stressors (here host quality and temperature stress) interact with inbreeding as measured by changes in the magnitude of8inbreeding depression, δ, as well as the relative importance of inbreeding-stress interactions to overall fitness. We found that while both environmental stressors caused substantial inbreeding-stress interactions as measured by change in δ, the relative importance of these interactions to overall survival was modest. This suggests that assessing inbreeding-stress interactions within the framework of δ alone may give an inaccurate representation of the relevance of interactions to population persistence. Furthermore, we found that the effect of environmental stress on fitness, but not inbreeding depression, varied strongly among populations. These results suggest that the outcomes of inbreeding-stress interactions are not easily generalized, an important consideration in conservation settings.

Survival and mass data were collected from inbred and outbred seed beetles from different lines and stress treatments.

Below is a description of the files contained in this compressed directory, as well as a description of how the analyses based on these files were run.
The programs R, jags, and Rjags are needed for this analysis.

This folder contains eight files:

Abbreviations used in these files include:

• BF = the Burkina Faso seed beetle lineage
• CA = the California seed beetle lineage
• SI = the South India seed beetle lineage
  
• H = heat treatment (37 C)
• R = regular incubation temperature (27 C)
• M = mung bean host (native host for SI)
• C = cowpea host (native host for CA and BF)
• P = green pea host (stressful host for all lineages)

DATA FILES

The five data files include data_mass_BF.csv, data_mass_CA.csv, data_survival_BF.csv, data_survival_CA.csv, and data_survival_SI.csv
These files include the raw data from the inbreeding-stress seed beetle experiment.

MASS DATA FILES

Each row in the two 'mass' data files represents the data for a single beetle. Each of the 'mass' data files contains eight columns:

The two full-sibling families within each block treatment were assigned either a capital or lowercase letter.

For example, block A would contain two unrelated full-sibling families: 'A' and 'a'.
These families were used for each of the four crosses within each block: inbred (i.e. 'AA' or 'aa') or outbred ('Aa' or 'aA').

ALL of the beetles weighed for the mass files were female as we were using female mass as a proxy for fecundity.

SURVIVAL DATA FILES

The three 'survival' data files contain the same first seven columns as the mass data, but with 8 additional columns:

MODEL FILES

1. The file "model_mass.txt"; contains the linear mass model used to analyze the raw mass data using jags.
   NOTE: Because the CA lineage had no female beetles survive in the host+temp and inbreeding+host+temp treatment groups, the sample sizes for these treatments are both zero. Thus, to run the analysis for CA, the model must be altered to comment out lines 56 (del[4]), 66 (HLE[4]), 75 (deldif[4]) and 84 (HLEdif[4]) as these are undefined for CA
2. The file 'model_survival.txt" contains the linear survival model used to analyze the raw survival data using jags.
   NOTE: Again, because the CA lineage had no beetles survive in the inbreeding+host+temp treatment group, the percent survival for this treatment was zero. Thus, to run the analysis for CA without problems, the survival model must be altered to comment out line 56 (HLE[4]) as this is undefined for CA.

R SCRIPT

The file named 'script_filter_data_run_model' contains the R code to compile and run the statistical analyses for each of the raw data files. Additional details for how the data were formatted for analysis are found in this file.

Lines 1 through 300 in this script can be run as-is with no alterations needed.

Before running lines 301 through 325, both the mass model and survival model text files (see 'MODEL FILES' above) must be altered to remove undefined variables for the CA lineage.

After changing the model text files for the CA lineage, lines 301 through 325 can be run as-is.

The output of running this R-script (in two segments) includes the following:

• D.BF, D.CA, and D.SI: these three lists contain the raw survival data formatted for use in jags
• D.BF.M, and D.CA.M: these two lists contain the raw mass data formatted for use in jags
• BF.surv.out, CA.surv.out, and SI.surv.out: these are MCMC lists containing the thinned MCMC output from each of the three MCMC chains in the survival model
• BF.mass.out and CA.mass.out: These are MCMC lists containing the MCMC output from the mass model
• BF.surv.summary, CA.surv.summary, and SI.surv.summary: these three matrices contain the summarized MCMC output
  (95% equal-tail probabilities from the posterior distributions of the tracked variables in the linear model)
• BF.mass.summary and CA.surv.summary: these two matrices contain the summarized MCMC output
  (95% equal-tail probabilities from the posterior distributions of the tracked variables in the mass model)