Published December 13, 2022 | Version 1.0
Software Open

Pascal code, input data, and output data for pGLS and RMA (= SMA) analysis of hummingbird morphological traits

  • 1. Federal University of Goiás, Brazil

Contributors

  • 1. Federal University of Goiás

Description

The following files are available for download from this Zenodo repository:

1) PGLS_Inputs.zip:  Input data for pGLS analysis. The file contains 3000 plausible phylogenies, in addition to morphological data for 220 species of hummingbirds, separately by sex and for intersexual averages, as well as elevation above sea level for each species, as detailed in the publication’s Appendix D2: Methods: Elevation.

2) FilterAndPGLS.dpr: Pascal source code. The software (1) inputs the morphological and elevation data, (2) inputs alternative phylogenies, and (3) performs pGLS analysis on (a) each morphological trait, controlling for phylogeny (called “phylo-filtering,” in the publication), and (b) each morphological trait controlling for phylogeny, body size, and elevation (called “triple-filtering,” in the publication). The software then (4) outputs the error term (residuals) of pGLS analysis, which can be interpreted as the morphological trait filtered for phylogeny and other trait values (body size and elevation).

3) RMA.dpr: Pascal source code for reduced major axis regression (RMA) between all pairs of morphological traits (RMA is identical to SMA, as defined in the publication). The software reads the filtered morphological trait values (output of pGLS analysis) and performs RMA. The RMA is performed for each replicate of pGLS, each using a different/alternative phylogenetic hypothesis. (Note: The output from this code is not used in the final version of the publication. Instead, SMA [= RMA] is computed from the means of PGLS residuals, among phylogenies, for each species.) Scattergrams, color-coded for feeding style, are produced from the PGLS residuals for each pair of variables.

4) PGLSAndRMA_Outputs.rar: Includes: (1) "Filtered" morphological traits, which are the outputs of pGLS analysis described above. (2) RMA outputs, as the estimated coefficients of RMA regression each pair of variables. Because RMA is replicated for alternative phylogenies, there are as many RMA coefficients as there are replicates of pGLS analysis (100). (3) Scattergrams, color-coded for feeding style, from the PGLS residuals for each pair of variables, overlaying all 100 sets of PGLS residuals and RMA lines.

NOTES: Delphi RAD Studio 11 Alexandria Release 2 was used to code, compile, and execute the programs FilterAndPGLS.dpr and RMA.dpr. The code can be compiled using Delphi (https://www.embarcadero.com/products/delphi/starter/free-download) or Free Pascal Compiler (https://www.freepascal.org/).

The data and software code were used in the paper "Repeated evolution of unorthodox feeding styles drives a negative correlation between foot size and bill length in hummingbirds" published in The American Naturalist.  

 

Files

PGLS_Inputs.zip

Files (93.6 MB)

Name Size Download all
md5:fcbe7677db8c12f8e0108be029ea25dc
13.8 kB Download
md5:36184b996416346c829ce663d4605498
9.3 MB Download
md5:3beeff97f1ef6d65b2ee250458989539
9.9 MB Preview Download
md5:c461fc358542b647dff0709caf787f4d
74.3 MB Download
md5:e9994b028ec85840d0ff6e4bdec4850b
6.8 kB Download

Additional details

References

  • Colwell et al. Repeated evolution of unorthodox feeding styles drives a negative correlation between foot size and bill length in hummingbirds. American Naturalist
  • Rangel et al. 2015. Phylogenetic uncertainty revisited. Evolution