Shield:

FBD+SSE

Repo holding simulation and analysis scripts for the simulation study testing SSE models on trees with fossil samples. Here we ran trait-dependent diversification simulations using the R package paleobuddy, generated phylogenetic trees with and without fossil samples, and then ran BiSSE analyses on RevBayes on both types of trees to assess the impact of including fossils on diversification rate estimates. See the manuscript for more details.

A detailed description of each file can be found in its respective directory. Here we describe how to reproduce the study. Note that each file is thoroughly commented to guide a user.

Data generation

1. Ensure the latest versions of the devtools, ape, and readr packages are installed. Install paleobuddy from github, making sure to install from the dev_traits branch. Also, ensure that file paths and Slurm configurations on simulation/pipeline and all analysis/ files are modified for your machine.
2. Run simulation/pipeline.R. It will generate 100 sets of 100-tip trees for each parameter combination and fossil sampling parameters. See the simulations/ directory for a summary of how the resulting replicates directory is organized. Note that as-written the pipeline will choose seeds at random. If interested in running using the same seeds as the original study, check the seeds.RData files for each parameter combination.
3. If using Slurm, run array.sh, which will then call bisse.sh to run the 76 parameter combinations, each for 100 reps corresponding to the trees generated in step 2. See the analysis/ directory for details on each analysis file.
4. Run plotting.R to parse the results, test the effective sample sizes, and plot summary statistics.

This work is licensed under a Creative Commons Attribution 4.0 International License.