Data and code for: Evolutionarily distinct species in frugivore networks have fewer links and more specialized partners
Data files
Jul 08, 2025 version files 44.25 KB
-
birdsynonyms.csv
6.51 KB
-
README.md
4.81 KB
-
README.txt
1.26 KB
-
ZanneTreeTrim.tre
31.66 KB
Abstract
Ecological networks describe the complex set of interconnections among species and their environment, and network structure can inform the stability, resilience, and functioning of ecosystems. Increasing attention is being paid to the mechanisms that determine species interactions. Phylogeny has informed our understanding of connections within networks, mostly by describing the strength of phylogenetic conservation of interactions. Here, we examine how the phylogenetic position of a species relates to its functional position within a network, testing the relationship between phylogenetic and network topologies. Time period: Early 20th century to present. We used a large dataset of frugivore interactions to calculate partner degree and degree, and bird and plant phylogenies to calculate local evolutionary distinctiveness (ED), evolutionary distinctiveness calculated on a community-level phylogeny. We then fit binomial Bayesian models to estimate the effect of evolutionary distinctiveness on partner degree and degree, comparing the effects of known bird traits from AVONET. The partners of both high ED birds and plants are less likely to interact with other members of the community, and birds with high ED tend to interact with fewer plant species. In birds, the effect size of ED is comparable to or larger than measured bird traits. Our analysis illustrates how phylogenetic relationships affect present-day ecologies and underscores the unique ecological contribution of evolutionary outliers.
https://doi.org/10.5061/dryad.08kprr5dm
Description of the data and file structure
In this analysis, we model the relationship between local evolutionary distinctiveness (ED) and two metrics of species positions in networks: partner degree and degree. We additionally examine the relationship between local ED of partners between phylogenies. We find that species with higher local ED tend to have fewer connections and that their partners tend to be more specialized
Files and variables:
File: ZanneTreeTrim.tre
Description: A phylogeny of plants from Zanne et al. (2014), trimmed to only contain species represented in the interaction networks from Fricke and Svenning (2020).
File: birdsynonyms.csv
Description: A dictionary of bird binomial synonyms. The binomial Latin names of species can change when species relationships are revised; this file contains information to fix discrepancies between the species interaction networks and the bird phylogeny.
Variables
- Row: Row number
- original: Accepted bird species identity used in networks in Fricke and Svenning (2020).
- synonym: Synonym species identity represented in Ericson backbone phylogeny from Jetz et al. (2014).
File: README.txt
Description: Readme containing information about files, code, and instructions to reproduce analyses.
Code/software
Accompanying R code can be run in the open source language R. Analyses were conducted with R version 4.3.2.
Dependencies (version number when analyses were conducted)
- tidyverse 2.0.0
- cli 3.6.3
- modelr 0.1.11
- bipartite 2.20
- picante 1.8.2
- ape 5.8-1
- cmdstanr 0.7.1
- brms 2.20.4
- tidybayes 3.0.6
- ggplot2 3.5.1
- ggdist 3.3.1
- cowplot 1.1.3
- magick 2.8.2
- taxize 0.9.100
Setting up for analysis
Working directory should include folders named 'data', 'output', and 'plots.' 'output' should contain folders 'models', 'stats', and 'summaries.'
Run 1_file_downloads.R to download necessary files from various repositories. This includes downloading the large collection of bird phylogenies from birdtree.org and subsetting it to the six used in the analysis, producing 'data/EricsonBirdTreeTrim.tre.'
Run 2_synonymization.R to create 'data/birdsynonyms.csv', a species dictionary to match network species IDs to phylogeny species IDs. This step may be skipped as the file is included in the .zip file.
Run 3_analysis.R to perform all initial analyses and create publication plots.
Analyses may be performed on different subsets of the data by changing the 'subset_type' variable.
- subset_type="all" performs analyses on all networks
- subset_type="comp" repeats analyses on the subset of networks labeled as comprehensively sampled, as defined in the Fricke and Svenning data (as opposed to networks produced by focusing on few members of a community)
- subset_type="fleshy" performs analyses on the subset of networks in which all plants are fleshy-fruited (as opposed to networks where some plants produce dry fruit, which may have different coevolutionary relationships with their consumers)
- subset_type="quant" performs analyses on the quantitative subset of networks, in which the number of interactions between each species pair is reported (as opposed to binary reports of whether interactions were observed or not)
Different analyses make use of the run_model() function, which is defined in the models.R depending on the modeltype and subset_type variables. To see specifications for each model or to make changes, consult the models.R file. The available model types are combinations of the model dependent variable ("degree", "partnerdegree", or partner evolutionary distinctiveness ("partnerED")), the focal taxa of the model ("bird" or "plant"), and for plant-focal degree or partner degree models, whether data should include all networks regardless of frugivore taxa ("alltaxa") or only those in which birds are the frugivores ("birdtaxa").
- "degree_bird", "partnerdegree_bird","degree_plant_alltaxa", "degree_plant_birdtaxa", "partnerdegree_plant_alltaxa", "partnerdegree_plant_birdtaxa", "partnerED_bird," and "partnerED_plant"
Other functions used in 3_analysis.R are defined in network_functions.R. Some of these functions were taken from Fricke and Svenning 2020, and are indicated as such in the file.
Access information
Data was derived from the following sources:
- Fricke and Svenning 2020 (https://doi.org/10.5061/dryad.44j0zpcbx)
- Zanne et al. 2014 (https://doi.org/10.5061/dryad.63q27)
- Jetz et al. 2014 (birdtree.org)
ZanneTreeTrim.tre and EricsonBirdTreeTrim.tre are trimmed versions of phylogenies from Zanne et al. 2014 and Jetz et al. 2014, respectively. Phylogenies were pruned to include only species represented in the Fricke and Svenning (2020) dataset.