Dataset DOI: 10.5061/dryad.ttdz08m8q

Description of the data

These data and scripts were used in analyses of of morphological evolution in the adaptive radiation of Malagasy vangas.

Data files

RAW DATA FOR MORPHOMETRIC ANALYSES IN R

raw_coordinates.csv - a csv file containing the raw (compiled) landmark data for the cranium and upper bill

• Column 1 is the unique specimen identifier (e.g. Artamella_viridis_100859) 
• The other 408 columns are the x,y,z coordinates for each of the 136 landmarks in the cranium and upper bill. These are ordered the same as in landmark_names.csv

raw_coordinates_lower.csv - a csv file containing the raw (compiled) landmark data for the cranium and upper bill

• Column 1 is the unique specimen identifier (e.g. Artamella_viridis_100859) 
• The other 234 columns are the x,y,z coordinates for each of the 78 landmarks in the lower bill. These are ordered the same as in landmark_names_lower.csv

foot_data_raw.csv - a csv file containing the raw linear measurements of pedal bones

• Specimen ID- An extended specimen id, e.g. calicalicus_madagascariensis_feet_fluid_352874. Indicates species, body part, preservation type (fluid or [round] skin), and museum accession number. See table S1 for detailed specimen info, including the museum each specimen is from.
• Species- The currently recognized species scientific name
• Genus- The genus that species belongs to
• Vanga- Binary identifier of whether or not the species is a Malagasy vanga; 1 = Malagasy, 0 = non-Malagasy.
• • tmt-d4p4 (12 columns)-* Measurements of each pedal bone. See Table S2 for descriptions of each measurement.
• Mass- Body mass in grams
• Foraging- Classification of foraging behavior (gleaning, probing, or sallying)

foot_data_logshape.csv - a csv file containing the size-corrected foot measurements (log-shape variables)

• Column 1- Species name
• tmt-d4p4 (12 columns)- the mean size-corrected measurement for each bone for that species. See Table S2 for descriptions of each measurement.

landmark_names.csv - a csv file containing the names of each landmark. Used in data_subsetting_clean.R to define sliding landmarks and modules.

• number- the landmark number (1-136)
• name- the name of the landmark. Landmarks 1-10 are type 1 landmarks with unique names describing their location. All others are semilandmark curves. Each curve has a unique name (e.g. maxilla_midline), which is followed by a number for that curve (.1 through .6) and then each landmark in that curve is numbered (e.g. .01 - .22). See Table S2 for a definition of each landmark and semilandmark curve.

landmark_names_lower.csv - a csv file containing the names of each landmark in the lower bill. Used in data_subsetting_clean.R to define sliding landmarks and modules.

• number- the landmark number (1-78)
• name- the name of the landmark. Landmarks 1-3 are type 1 landmarks with unique names describing their location. All others are semilandmark curves. Each curve has a unique name (e.g. lower_tomium), which is followed by a number for that curve (.1 through .3) and then each landmark in that curve is numbered (e.g. .01 - .25). See Table S2 for a definition of each landmark and semilandmark curve.

PHYLOGENIES

vanga_timetree_10_8_24 - a tree file in nexus format, containing the time-calibrated phylogeny used in all analyses

foot_timetree_10_8_24 - a tree file in nexus format, the same phylogeny but pruned to the taxa represented in the foot dataset

BAYESTRAITS & DTT

vangas_foot.txt; vangas_mass.txt; vangas_pc9_bill.txt; vangas_pc13_all.txt; vangas_posthoc.txt - text files containing the PC scores for axes summarizing 95% of variance for each dataset. Used as input for Bayestraits & dtt analyses.

bayestraits_commfiles.zip - a zipped folder of files used to run Bayestraits analyses. Contains:

17 individual command files, setting up the bayestraits parameters for each model. These are the versions for the full skull dataset; the only differences for the other datasets are the name of the Log file and in some cases the number of iterations.

commands_list_all - the line of code used to actually run each model for the full skull dataset; this is set up for submitting as a batch to a supercomputer, but each can be run individually very easily on a laptop.

output_files.zip - a zipped folder containing all output files from every replicate for every model and every dataset.

Organized as a separate subfolder for each dataset: out_all, out_bill, out_feet, out_mass, and out_posthoc.See main text for descriptions of these datasets.

each individual analysis produces 3 (BM) to 5 (rj) output files, resulting in 345 files total per subfolder. These are: the Log file (.Log.txt), Schedule file (.Schedule.txt), results of the stepping stone sampler (.Stones.txt), .Output.trees (for all models but BM) the rescaled trees in nexus format, and .VarRates.txt (for rj models) which summarizes the modeling results.

reduced_output.zip - a zipped folder containing files from just one replicate of the best-supported models for the skull and foot datasets. Sufficient to replicate the figures and get a sense of how the full analyses work. Contains 10 files, one of each of the above for the reversible-jump delta transform for the feet, and the reversible-jump branch transform for the full skull landmark dataset.

Code

Most analyses were performed in R (see session_info.txt file for all version details). 

Bayestraits analyses were performed using Bayestraits V4.0. Stationarity and convergence (from the .Log files) were assessed with Tracer.

Code files

data_subsetting_clean.R - an R file that organizes the landmark data, does mirroring, procrustes superimposition, and creates data and tree subsets. MUST BE RUN prior to other morphometrics analyses; saves an Rdata file at the end. Not required to run Bayestraits or foot analyses.

\--INPUT: raw_coordinates, raw_coordinates_lower, landmark_names, landmark_names_lower, and vanga_timetree

feet_code_clean.Rmd - an R Markdown file that formats the raw foot measurement data and runs most foot morphometric analyses: PCA, ANOVA and MANOVAs, morphological disparity, phylogenetic signal, net evolutionary rate, and integration (eigenvalue dispersion). Written primarily by Euan Lim with contributions from Anya.

\--INPUT foot_data_raw or foot_data_logshape and foot_timetree
\--OUTPUT:
	\--Figure 2, Figures S2 & S3
	\--Tables 1 & 3, Tables S3, S4, S5, S6

morphometric_analyses_clean.R - an R file that performs many of the morphometric analyses on the skull landmark dataset. Requires Rdata file produced by data_subsetting_clean.R as input. Generates PCAs and runs analyses of phylogenetic signal, morphological disparity, MANOVAs, and net evolutionary rate.

\--OUTPUT:
	\--Figure 2, Figures S1, & S2
	\--Tables 1 & 3, Tables S4, S5, and S12

hypervolumes_clean.R - an R script that calculates and plots the morphospace hypervolumes and associated statistics for both skull and foot datasets. Requires Rdata file produced by data_subsetting_clean.R as input.

\--OUTPUT:
	\--Figure 3, Figure S4
	\--Table 1

integration_clean.R - an R script that runs all analyses of integration (eigenvalue dispersion and PLS). Requires Rdata file produced by data_subsetting_clean.R as input. Calculates and compares eigenvalue dispersion, phylogenetic 2-block partial least squares, and PLS of foot and skull. Makes PLS plots.

\--OUTPUT:
	\--Figure 4
	\--Tables S6, S7, S8

modularity_clean.R - calculates and compares skull modularity (covariance ratio) and plots null covariance ratio distributions. Requires Rdata file produced by data_subsetting_clean.R as input.

\--OUTPUT:
	\--Figure S5
	\--Table 2

bayestraits_dtt_clean.R - creates the morphometric data files for these analyses and analyzes output. For Bayestraits: calculates Bayes factors, summarizes output files, and plots the phylogenies. For dtt: performs dtt analysis, plots results.

\--INPUT:
	\--Bayestraits: output_files (or reduced_output) and both phylogenies as input
	\--dtt: PC score .txt files and phylogenies as input
\--OUTPUT:
	\--Figure 5, Figures S6 & S7 
	\--Tables S9, S10, S11