Data and Scripts from: Bayesian prediction of multivariate ecology from phenotypic data yields novel insights into the diets of extant and extinct taxa
Nations, Jonathan; Wisniewski, Anna; Slater, Graham (2022), Data and Scripts from: Bayesian prediction of multivariate ecology from phenotypic data yields novel insights into the diets of extant and extinct taxa, Dryad, Dataset, https://doi.org/10.5061/dryad.pc866t1rg
An organism’s phenotype often relates to its ecology in a well-characterized manner, enabling prediction of ecology for taxa that lack direct ecological information, such as fossils. Diet is a particularly important component of a species' ecology; however, in order to predict diet it must first be codified, and establishing metrics that effectively summarize dietary variability without excessive information loss remains challenging. We employed a dietary item relative importance coding scheme to derive multivariate dietary classifications for a sample of extant carnivoran mammals, and then used Bayesian multilevel modeling to assess whether these scores could be predicted from a set of dental metrics. There is no ``one size fits all" model for predicting dietary item importance; different topographical features best predict different foods, and model-averaged estimates perform especially well. We also show how models derived from living taxa can be used to provide novel insights into the dietary diversity of extinct carnivoran species. Our approach need not be limited to diet as an ecological trait of interest, to these phenotypic traits, or to carnivorans. Rather, this framework serves as a general approach to predicting multivariate ecology from phenotypic traits.
The dataset contains dental topographic metrics (aDNE, OPCr, RFI, and RLGA) and body mass estimates for 113 living and extinct carnivorans (Order Carnivora). It also contains ordinal dietary rankings (1-4, with 1 being none in the diet and 4 being a primary food item) for 13 food items for 92 of the extant species. We use these dietary rankings to generate predictive models of multivariate diet, then estimate the multivariate diets for fossil and understudied extant taxa. Running the models yields outputs which are then used to generate predictions. The data also contains a phylogenetic tree of Carnivora used in the models.
The Dryad repository contains all of the input and output data from this project. The README.md file details each of the 13 data objects (mostly csv files). As each of the outputs is called as an input in downstream analyses, we included these.
The Zenodo repo linked to this Dryad repo contains the entirety of the GitHub repo for this project. It is written in such a way where, once all the packages have been installed, it can be downloaded and run in it's enitety to reproduce all of the results from this project. It contains the scripts, input data, plotting commands, and its own detailed README.md file.
The Zenodo repo also contains the Supplemental Figures and the Supplemental table, all in .pdf format.
National Science Foundation, Award: DBI-2010756