Data from: Inferring the locomotor ecology of two of the oldest fossil squirrels: Influence of operationalisation, trait, body size, and machine learning method
Data files
Oct 03, 2024 version files 908.36 MB
-
README.md
2.94 KB
-
Woelfer_Lionel_2024_R_data_and_code.zip
908.36 MB
Abstract
Correlations between morphology and lifestyle of extant taxa are useful for predicting lifestyles of extinct relatives. Here, we infer the locomotor behaviour of Palaeosciurus goti from the middle Oligocene and P. feignouxi from the lower Miocene of France using their femoral morphology and different machine learning methods. We used two ways to operationalise morphology, in the form of a geometric morphometric shape dataset and a multivariate dataset of eleven femoral traits. The predictive models were built and tested using more than half (180) of the extant species of squirrel relatives. The neural network model had the greatest predictive power, sometimes outperforming more traditional methods such as linear discriminant analysis. However, the predictive power also depended on the operationalisation and the femoral traits used to build the model. We also found that predictive power tended to slightly improve with increasing body size. Contrary to previous suggestions, the older species, P. goti, was most likely arboreal, whereas P. feignouxi was more likely terrestrial. This provides further evidence that arboreality was already the most common locomotor ecology among the earliest squirrels, while a predominantly terrestrial locomotor behaviour evolved shortly afterwards, before the vast establishment of grasslands in Europe.
README: R data and code
https://doi.org/10.5061/dryad.c866t1gdb
Inferring the locomotor ecology of two of the oldest fossil squirrels: influence of operationalisation, trait, body size, and machine learning method
Jan Wölfer1* & Lionel Hautier2
1Humboldt-Universität zu Berlin, Philippstraße 13, 10115 Berlin
2Institut des Sciences de l’Évolution de Montpellier, UMR 5554, Univ de Montpellier, CNRS, IRD, Montpellier Cedex 5, France
*corresponding author: jan.woelfer@gmx.de
Article DOI: 10.1098/rspb.2024.0743
Here you find the R code and data together presented as an R project. It contains all data necessary to reproduce the geometric morphometric analysis of the landmark data, the generation of the multivariate dataset, the machine learning methods for classification/testing/prediction, and the generation of the phylogeney with ancestral character estimation. After unzipping the folder, you will find the following subfolders and main files:
- ply subfolder: contains all bone surface files as .ply
- pts subfolder: contains all landmarks stored in the .pts format
- Template subfolder: contains the .ply and .pts files of the template for projecting the surfice sliding semi-landmarks onto the target specimens
- Ident_fossils.xlsx: grouping variables in terms of lifestyle (see the R script for more information), sex, body mass [g] and an indicator variable of the fossils.
- linear_data_fossils.xlsx: variables which where measured directly on surface files in Geomagic (patellar grove width, anteroposterior and mediolateral midshaft diameters) with three measurements each.
- sciuromorpha_phylogeny.nex: Phylogenetic information in nexus format
- Femur.Rproj: R project file
- .RData: The R objects generated during the final analysis
- resamplingsV2.R: R code used to downsample the curve sliding semi-landmarks. This file is based on a previous version (kindly provided by Anne-Claire Fabre), which was published as supplementary material to the following publication: Botton-Divet, Léo, et al. "Morphological analysis of long bones in semi-aquatic mustelids and their terrestrial relatives." Integrative and Comparative Biology 56.6 (2016): 1298-1309.
- Script: R script for analysis
In addition, you find various .R objects which store interim results: atlas.R, final_specimens_fossils_all.R, linear_df_fossils_all.R, Proc_fit_fossils_all.R, Proc_fit2_fossils_all.R, slid_specimens_fossils_all.R. When going through the script, these are loaded by default into R to ensure that the results of the analysis remain the same despite potential changes in the R functions since the generation of these files.
To reproduce all results and raw figures, open the Femur.Rproj file and the R script called 'Script' (if not loaded automatically). The script is annotated and contains all steps of the analysis from top to bottom.