Avian cranial evolution is influenced by shape interactions between hard and soft tissue traits
Data files
Nov 11, 2025 version files 6.50 GB
-
morphometrics_data.zip
6.42 GB
-
mvSLOUCH_data.zip
74.55 MB
-
README.md
17.13 KB
Abstract
Changes in the structure and relative size of the brain are thought to be key transformations in the evolution of birds, reflecting innovations and diversity of neurosensory and cognitive capabilities. These changes do not occur in isolation, being accompanied by many other derived morphological characteristics. In the avian head alone, these include the evolution of a toothless beak, increase in relative eye size, and reduction and restructuring of jaw muscles. Several developmental trade-offs have been proposed to explain the interrelationships among these traits, but how these developmental patterns translate into evolutionary correlations among cranial traits is poorly understood. Here, we use two-block partial least squares analyses and Ornstein-Uhlenbeck models of adaptive trait evolution to explore the phenotypic evolution of hard and soft cranial tissues and test hypotheses of correlated trait evolution. In pairwise analyses, we found that all traits are significantly correlated, and found support for a form of adaptive trait evolution across the whole head in which traits interact reciprocally via the neurocranium. Together, these results highlight the integrated nature of the avian head and reveal that the evolution of diverse phenotypes is a result of complex multiple interactions among hard and soft tissue traits.
https://doi.org/10.5061/dryad.sxksn03dp
Description of the data and file structure
This repository contains all data, scripts and supplementary figures from "Avian cranial evolution is influenced by shape interactions between hard and soft tissue traits."
Code/software
All code is written for R (v. 2024.12.1+563), with required packages labelled at the start of each script. Packages should be installed prior to running scripts.
The scripts for replicating this study are in two parts (Morphometrics and mvSLOUCH) and are available via Zenodo. The morphometrics scripts generate the data input into the mvSLOUCH analysis and should be run first. The required input data for each set of scripts is included in Dryad in the relevant section. Instructions for constructing the phylogeny used in these analyses are included in the Phylogeny Construction section, also available via Zenodo. This section is included for reference, and it is not required to run through this script because the final phylogeny is included in both the Morphometrics and mvSLOUCH sections as the tree.tre object.
Part 1: Morphometrics
1.1 Contents
Morphometrics data (available via Dryad)
morphometrics_data.zip
┣ input_data
┃ ┣ classifier.csv
┃ ┣ endo_resampling_template.csv
┃ ┣ patchpoints.pts
┃ ┣ skull_resampling_template.csv
┃ ┣ slidedlms.endocast.rda
┃ ┣ slidedlms.skull.rda
┃ ┗ tree.tre
┣ output_data
┃ ┣ ES.rda
┃ ┣ PCA.3D.endo.rda
┃ ┣ PCA.3D.muscle.rda
┃ ┣ PCA.3D.neur.rda
┃ ┣ PCA.3D.rostrum.rda
┃ ┣ PCA.3D.skull.rda
┃ ┣ reflected_patched_endocasts.rda
┃ ┣ Y.gpa.endo.rda
┃ ┣ Y.gpa.muscle.rda
┃ ┣ Y.gpa.neur.rda
┃ ┣ Y.gpa.rostrum.rda
┃ ┗ Y.gpa.skull.rda
┣ ply_ASCII_endocast
┣ ply_ASCII_skull
┣ pts_endocast
┗ pts_skull
Morphometrics scripts (available via Zenodo)
morphometrics_scripts
┣ scripts
┗ ┣ 1.skull_resampling_sliding.R
┣ 2.endocast_resampling_sliding.R
┣ 3.skull_alignment.R
┣ 4.endocast_patching_alignment.R
┣ 5.ES_script.R
┣ 6.two_block_pls.R
┣ 7.data_export.R
┣ 8.plotting_code.R
┗ source_code
┗functions.R
1.2 Description of files and variables
1.2.1 morphometrics_data
Folder: input_data (available via Dryad)
File: classifier.csv
Description: Contains information for all 322 species included in the study. This includes clade, specimen number, and source, body mass in grams, and brain volume in mm3.
File: patchpoints.pts
Description: Template for endocast surface semilandmark patch, used in 2.endocast_resampling_sliding.R script (see details below). The template is created on the surface mesh of Anhinga anhinga, which is contained within the ply_ASCII_endocast.zip file.
File: tree.tre
Description: Full time-scaled phylogeny in .tre format for 322 species included in study. Instructions for recreating the phylogeny are contained in phylogeny_construction.zip.
File: skull_resampling_template.csv
Description: Table containing template for standardised resampling of skull semilandmark curves, used in 1.skull_resampling_sliding.R script. Information includes start point, end point, and the number of curves required for each curve. Only lm1,lm2, and ptswanted columns are necessary for the curve resampling script.
Variables
- Curves: Curve identifier number
- lm1: Start point of curve (landmark number)
- lm2: End point of curve (landmark number)
- ptswanted: Number of points needed for resampling
- start: Landmark number of start of curve (after resampling, excluding lm1)
- end: Landmark number of end of curve (after resampling, excluding lm2)
File: endo_resampling_template.csv
Description: Table containing template for standardised resampling of endocast semilandmark curves, used in 2.endocast_resampling_sliding.R script. Information includes start point, end point, and the number of curves required for each curve. Only lm1,lm2, and ptswanted columns are necessary for the curve resampling script.
Variables
- Curves: Curve identifier number
- lm1: Start point of curve (landmark number)
- lm2: End point of curve (landmark number)
- ptswanted: Number of points needed for resampling
- start: Landmark number of start of curve (after resampling, excluding lm1)
- end: Landmark number of end of curve (after resampling, excluding lm2)
File: slidedlms.endocast.rda
Description: R data file containing endocast landmarks for all specimens after resampling and sliding procedure in 2.endocast_resampling_sliding.R. This information is used in the 4.endocast_patching_alignment.R script.
File: slidedlms.skull.rda
Description: R data file containing endocast landmarks for all specimens after resampling and sliding procedure in 1.skull_resampling_sliding.R. This information is used in the 3.skull_alignment.R script.
Folder: pts_endocast
Description: Individual endocast landmark files for all 322 species included in the study and used in 2.endocast_resampling_sliding.R. Landmark files are in pts format and each contains 14 anatomical (Bookstein Type I and II) landmarks and 15 semilandmark curves. Each row represents a single landmark and contains three columns of X, Y, and Z spatial coordinates. Refer to the Supplementary File for details of each landmark.
Folder: pts_skull
Description: Individual skull landmarks for all 322 species included in the study and used in 1.skull_resampling_sliding.R. Landmark files are in pts format and each contains 20 anatomical (Bookstein Type I and II) landmarks and 21 semilandmark curves. Each row represents a single landmark and contains three columns of X, Y, and Z spatial coordinates. Refer to the Supplementary File for details of each landmark.
Folder: ply_ASCII_endocast
Description: Contains three-dimensional (3D) endocast surface meshes for all 322 species included in the study, in ASCII .ply format and used in 2.endocast_resampling_sliding.R script. Meshes are scaled in mm and can be opened and viewed in any 3D viewing software that supports .ply format. The source of all scans is listed in the classifier.csv document.
Folder: ply_ASCII_skull
Description: Contains three-dimensional (3D) skull surface meshes for all 322 species included in the study, in ASCII .ply format and used in 1.skull_resampling_sliding.R script. Meshes are scaled in mm and can be opened and viewed in any 3D viewing software that supports .ply format. The source of all scans is listed in the classifier.csv document.
1.2.2 output_data (available via Dryad)
This folder contains output files generated in the morphometrics section.
File: ES.rda
Description: R data file containing adjusted orbit diameter values for all species.
File: PCA.3D.endo.rda
Description: R file containing PCA eigenvalues for patched endocast landmarks.
File: PCA.3D.muscle.rda
Description: R file containing PCA eigenvalues for locally-aligned jaw muscle landmarks.
File: PCA.3D.neur.rda
Description: R file containing PCA eigenvalues for locally-aligned neurocranium landmarks.
File: PCA.3D.rostrum.rda
Description: R file containing PCA eigenvalues for locally-aligned rostrum landmarks.
File: PCA.3D.skull.rda
Description: R file containing PCA eigenvalues for globally-aligned skull landmarks.
File: reflected_patched_endocasts.rda
Description: R file containing landmark data for all specimens after application of surface patch semilandmarks.
File: Y.gpa.endo.rda
Description: R file containing Procrustes-aligned endocast landmark coordinates for all specimens.
File: Y.gpa.muscle.rda
Description: R file containing Procrustes-aligned jaw muscle landmark coordinates for all specimens.
File: Y.gpa.neur.rda
Description: R file containing Procrustes-aligned neurocranium landmark coordinates for all specimens.
File: Y.gpa.rostrum.rda
Description: R file containing Procrustes-aligned rostrum landmark coordinates for all specimens.
File: Y.gpa.skull.rda
Description: R file containing Procrustes-aligned skull landmark coordinates for all specimens.
1.2 Instructions for running morphometric analyses.
1.2.1 Data (available via Dryad)
morphometrics/input_data/ contains input data needed to run morphometric analyses, including species data for all specimens (classifier.csv), landmark patching template for endocast (patchpoints.pts), semilandmark curve resampling templates (endo_resampling_template.csv and skull_resampling_template.csv) and the full, timescaled phylogeny for the dataset (tree.tre). It also contains the output of the resampling and sliding portion of the script, i.e. the input landmark data for the morphometric analyses (slidedlms.endocast.rda and slidedlms.skull.rda).
morphometrics/output_data/ contains the output from the morphometric analyses, including Procrustes alignments Y.gpa....rda and principal component analyses PCA.3D....rda for each landmark subset (endocast, neurocranium, jaw muscle and rostrum).
morphometrics/ply_ASCII_endocast/ and morphometrics/ply_ASCII_skull/ contain meshes for all endocasts and skulls used in the analysis, in ply ASCII format.
morphometrics/pts_endocast/ and morphometrics/pts_skull/ contain landmark data for all specimens, for endocasts and skulls, respectively.
1.2.2 Scripts (available via Zenodo)
Folder: morphometrics_scripts
morphometrics/scripts/ contains all scripts required to run morphometric analyses on the dataset, which should be run in order. All required packages are listed in the scripts and should be installed prior to running the scripts. The scripts 1.skull_resampling_sliding.R and 2.endocast_resampling_sliding.R import landmark data from morphometrics/pts_skull/ and morphometrics/pts_endocast/, respectively, and resample the semilandmark curves to standard numbers of landmarks based on the relevant ...resampling_template.csv file. The scripts include the process for estimating missing landmarks, sliding the resampled semilandmark curves to minimise bending energy, and exporting the resultant landmark files as a single R script for each landmark set (endocasts and skulls). 3.skull_alignment imports the skull landmarks from step 1, reflects the landmarks across the sagittal plane and performs a generalised Procrustes alignment, and a PCA on the subsequently aligned right-side landmarks for each subset of skull landmarks. 4.endocast_patching_alignment does the same, but first places surface semilandmarks on each specimen using the patchpoints template before the sagittal reflection step. 5.ES_script.R uses landmarks from the skull to calculate orbit diameter adjusted for neurocranium size. 6.two_block_pls.R performs two-block partial least squares on all pairwise combinations of traits. 7.data_export.R exports all results for mvSLOUCH analyses. 8.plotting_code.R contains code for plotting data.
The folder source_code contains all specialist functions required to perform morphometrics analyses.
Part 2: mvSLOUCH process
2.1 Contents
mvSLOUCH data (available via Dryad)
mvSLOUCH_data.zip
┣ indivResults
┃ ┣ bootstrap
┃ ┣ initialdiagUTASR
┃ ┣ model01
┃ ┣ model02
┃ ┣ model03
┃ ┣ model04
┃ ┣ model05
┃ ┣ model06
┃ ┣ model07
┃ ┣ model08
┃ ┣ model09
┃ ┣ model10
┃ ┣ model11
┃ ┣ model12
┃ ┗ modelBM
┣ RandomSeeds
┃ ┣ bootstrap
┃ ┣ initialdiagUTASR
┃ ┣ model01
┃ ┣ model02
┃ ┣ model03
┃ ┣ model04
┃ ┣ model05
┃ ┣ model06
┃ ┣ model07
┃ ┣ model08
┃ ┣ model09
┃ ┣ model10
┃ ┣ model11
┃ ┣ model12
┃ ┣ modelBM
┃ ┣ BestFoundModel.RData
┃ ┗ Optima_bestmodel.pdf
┣ raw_data
┃ ┣ classifier_mvSLOUCH.csv
┃ ┣ mvslouch_data.csv
┃ ┗ tree.tre
┣ Results
┗ ┣ bootstrap
┣ initialdiagUTASR
┣ model01
┣ model02
┣ model03
┣ model04
┣ model05
┣ model06
┣ model07
┣ model08
┣ model09
┣ model10
┣ model11
┣ model12
┣ modelBM
┣ BestFoundModel.RData
┣ Optima_bestmodel.pdf
┗ Optima_bestmodel_traits.pdf
mvSLOUCH scripts (available via Zenodo)
mvSLOUCH_scripts
┗ scripts
┣ 1_startingscript.R
┣ datasetup.R
┣ do_parametric_bootstrap.R
┣ model01.R
┣ model02.R
┣ model03.R
┣ model04.R
┣ model05.R
┣ model06.R
┣ model07.R
┣ model08.R
┣ model09.R
┣ model10.R
┣ model11.R
┣ model12.R
┣ modelBM.R
┣ network_plot.R
┣ regimes.R
┣ startingmodels.R
┣ tosource.R
┗ usercontrolledvariables.R
2.2 Description of files and variables
Folder: mvSLOUCH_data (available via Dryad)
Folder: indivResults
Description: Contains output folders into which model outputs are saved.
Folder: RandomSeeds
Description: Contains random seeds generated for each model runthrough.
Folder: raw_data
Desription: Contains input data for mvSLOUCH analyses.
File: classifier_mvSLOUCH.csv
Description: Contains taxon information for each species included in study.
File: mvslouch_data.csv
Description: Contains PC scores and adjusted orbit diameter data for mvSLOUCH analysis. This information is generated in the morphometric analysis section.
File: tree.tre
Description: Timescaled phylogenetic tree for all 322 species included in study.
Folder: scripts (available via Zenodo)
Description: Contains all scripts necessary for running mvSLOUCH analysis (see section 2.2: Instructions for running mvSLOUCH analysis, below). These scripts are available via Zenodo.
2.3 Instructions for running mvSLOUCH analysis
Input folder (available via Dryad)
mvSLOUCH/raw_data/ contains input data files for mvSLOUCH analyses. classifier.csv contains taxon data. mvslouch_data.csv contains relevant PC scores for each trait and adjusted orbit diameter for each specimen. tree.tre is the timescaled phylogeny for all taxa, created with scripts contained in the phylogeny.zip folder.
Output folders (available via Dryad)
The folders named mvSLOUCH/indivResults/, mvSLOUCH/RandomSeeds/ and mvSLOUCH/Results contain output folders for all models for each analysis. Results will be saved into these folders.
Folder: mvSLOUCH_scripts
Script folder (available via Zenodo)
mvSLOUCH_scripts/ contains all R scripts needed to perform mvSLOUCH analyses. All scripts in this folder are adapted from Bartoszek et al., 2024 (https://doi.org/10.1111/2041-210X.14376). The supplementary data for this paper contains the original scripts and should be referred to for any queries (original code available at Zenodo: https://doi.org/10.5281/zenodo.11286569). 1_starting_script.R is the shell script containing all separate scripts, which should be run through in order. usercontrolledvariables.R defines pre-set starting variables for the analyses, including file pathways. datasetup.R contains code for reading in and standardising data contained in mvslouch_data.csv file prior to analysis. tosource.R contains functions specific to the mvSLOUCH analysis. startingmodels.r is the script to initiate the starting conditions for each model, and should be run to completion before attempting to run any individual model. Scripts model01.R to model12.R contain the model parameters for each hypothesis outlined in the manuscript, and should be run through in order in the shell script.
Finally, do_parametric_bootstrap.R contains a script to sample outputs from all models and summarises the best supported model and its parameter. Trait optima plots (Fig. 2 in manuscript) can be made from the output of the parametric bootstrap with the make_optima_plot.R script.
3. Phylogeny construction
Folder: phylogeny_construction (available via Zenodo)
phylogeny_constructioncontains R code and instructions for constructing the phylogenetic tree used in the study, using the tree.merger function from the R package RRphylo. Refer to Marek&Felice_treemerging_README.md for detailed instructions. This information is included for reference, the final tree used in the analysis is contained in both morphometrics/input_data/tree.tre and mvSLOUCH/raw_data/tree.tre .
phylogeny_construction
┣ Marek&Felice_treemerger_part1.r
┣ Marek&Felice_treemerger_part2.r
┗ Marek&Felice_treemerging_README.md