Multiple diversification and fossil sampling models can make total-evidence phylogenies more imbalanced but only a few make them look like empirical trees

This repository includes R code for replicating the analyses and figures presented in this paper. It also includes several datasets summarising the outputs of these analyses.

Description of the data and file structures

R Code

1. functions.R A series of helper functions including tree building functions for each of the diversification models and sampling functions for each of the preservation regimes
2. simultating_trees_e100.R Script for simulating trees under each combination of diversification model and fossil sampling model
3. analysis_simulation.R Script for analysing the trees simulated using simultating_trees_e100.R
4. ClaDS_simulation.R Script for fitting ClaDS models to empirical total-evidence phylogenies
5. CLaDS_empirical_simulation.R Script for simulating trees from the parameter values estimated using ClaDS_simulation.R
6. lolipop.R Script for reproducing Figure 5.

Supplementary Data Summarising the Analyses

1. CI_simulation_output_ne100.csv

Summary of the output of the trees simulated under each combination of diversification models and fossil sampling models produced using simultating_trees_e100.R and analysis_simulation.R. The dataset includes the number of fossil lineages on the simulated tree (Fossil_no), the number of fossil preserved by the fossil sampling models (Preservation), the length of the simulated tree (Tree_length), the starting speciation rate (lambda_init), the starting extinction rate (mu_init), the lower and upper bounds of the distributions parameterising the shifts in speciation and extinction rates (min_shift and max_shift), the number of extant species on the tree (n_extant), the diversification model under which the tree was simulated (model, see manuscript for a description of each model), the point in time in which the mass extinction event occurs (extinction_date, only relevant for models which include a mass extinction), the proportion of lineages which survive the extinction event (survivorship), the length of the root edge branch (root.edge, used to correct calculations of Tree_length for some of the diversification models), the point in time at which a universal shift in speciation rate occurs (shift_date, only relevant to the universal diversification rate shift model), the scaler for the universal speciation rate shift (shift_scaler), the limit of the size of the clade that arises from the key innovation (n_sub, only applicable to the density dependent with key innovation model), the point in time at which the key innovation arises (tinn), the fossil sampling model used to sample the extinct lineages (preservation), the Colless Index of imbalance for the simulated phylogeny (Index).

2. ClaDS_summary_all_trees.csv

Summary of the output of trees simulated using the parameters derived from ClaDS_simulation.R and ClaDS_empirical_simulation.R. The dataset includes the Colless-Index of imbalance for the complete tree and the tree sampled under each of the fossil sampling modes (CI_all, CI_diversified, CI_random, CI_trait_dependent, CI_time_dependent), the number of extinct lineages on the simulated tree (Fossils), Pybus and Harvey's Gamma statistic for each of the fossil sampling treatments (Gamma_*), the starting speciation rate (lambda_init), the starting extinction rate (mu_init), the number of extinct lineages sampled by the fossil sampling model (Preservation), the proportion of extant species sampled to bring the tree size into accordance with the empirical tree (sampling_fraction), , the length of the simulated tree (Tree_length), the diversification model under which the tree was simulated (model), the point in time in which the mass extinction event occurs (extinction_date, only relevant for models which include a mass extinction), the proportion of lineages which survive the extinction event (survivorship), the point in time at which a universal shift in speciation rate occurs (shift_date, only relevant to the universal diversification rate shift model), the scaler for the universal speciation rate shift (shift_scaler), the published total-evidence tree on which the simulation was based (tree, see manuscript for detailed description of each empirical phylogeny). preservation regime.

Sharing/Access Information

The study uses a number of previously published total-evidence of which all but one are available online. At the time of writing these phylogenies can be accessed via:

• Dornburg et al.’s (2015) total-evidence phylogeny of Holocentridae is available in the Dryad Digital Repository, at https://doi.org/10.5061/dryad.b708j

• Gustafson et al.’s (2017) total-evidence phylogeny of whirligig were provided by Dr Grey Gustafson by permission. Data will be shared on request to the corresponding author with permission of Dr Grey Gustafson.

• Slater et al.’s (2017) total-evidence phylogeny of Mysticeti is available in the Dryad Digital Repository, at https://doi.org/10.5061/dryad.b68g0

• Brennan’s (2019) total-evidence phylogeny of Macropodidae is available on the authors GitHub, at https://github.com/IanGBrennan/FossilUncertainty

• Koch and Thompson’s (2021) total-evidence phylogeny of Echinoidea is available in the Dryad Digital Repository, at https://datadryad.org/stash/dataset/doi:10.5061/dryad.7h44j0zr3

The best_model_sim.txt includes the model summary for the glmm fit to the dataset generated from the general simulation procedure and the [study]_best_model.txt files are model sumaries for each of the glmm models fit to the dataset generated from the informed simulation procedure.