Overview

The multi-part objects in this example are partial wing skeletons comprised of humerus, radius and ulna. These partial wing skeletons are from 15 extant species of penguin and five fossil species of penguin, which together constitute a dataset of 60 wing bones (Table 1).

   Shape data from the multiple-part objects are provided as landmark configurations. Three sets of landmarks were produced for each of the digital replicas so that the replicates could be averaged to mitigate effects of placement error. The nine sets of landmark configurations (i.e. three sets from each of the three bones) were separately read into R using readPts with gpa = FALSE (i.e. generalized Procrustes transformation not performed) and are available by calling morphoBlocks. The function readPts is a wrapper for the read.pts and cSize functions from Morpho (Schlager 2017). First, load the dataset:
library(morphoBlocks)
data(penguinWings)

 

The analysis in this example will use the mean values of the three replicate landmark configurations from each part. Creating these averaged values requires averaging the configurations stored in the @raw term of their respective block-class objects.

# Extract and average the landmark configurations

hum_av <- (hum1@raw + hum2@raw + hum3@raw)/3
rad_av <- (rad1@raw + rad2@raw + rad3@raw)/3
uln_av <- (uln1@raw + uln2@raw + uln3@raw)/3

 

Prepare data blocks for analysis

Each set of averaged landmarks can now be formatted into a data block with the formatBlock function. Here, we will use the @curves term from one of the replicates, and will set gpa = TRUE to perform generalized Procrustes transformation on the landmark configurations. generalized Procrustes transformation is performed using gpagen from geomorph (Adams and Otarola-Castillo, 2013), which is called by formatBlock.

# Format the averaged landmark configurations into data blocks

block1 <- formatBlock(hum_av, curves = hum1@curves, k = 3, gpa = TRUE)
block2 <- formatBlock(rad_av, curves = rad1@curves, k = 3, gpa = TRUE)
block3 <- formatBlock(uln_av, curves = uln1@curves, k = 3, gpa = TRUE)

 

The three data blocks of Procrustes-transformed configurations (humerus, radius, ulna) are organised into a list of data blocks and scaled using the normalised weighted centroid size method from Collyer et al. (2020).

# Scale and combine data blocks into a single list of blocks

blocklist <- combineBlocks(blocks = c(block1, block2, block3))

 

Analyse list of blocks

The scaled data blocks are analysed with regularised consensus principal component analysis (RCPCA) in mode 2 using the rgcca function from RGCCA (Tenenhaus and Guillemot, 2017), which is called by analyseBlocks when option = "rcpca".

# Analyse the list of data blocks using RCPCA

result <- analyseBlocks(blocklist, ncomp = 10)

 

Plot results

Use scoresPlot to show the consensus space from the analysis, which here represents a morphospace for the partial-wing skeleton.

# Setup colour vector to show different ages of fossil penguins. Paleocene (brown),
# stem-lineage penguins from the Oligocene (light brown), and extant penguins (white).

pcol <- c("#ffffff", "#ffffff", "#ffffff", "#ffffff", "#e6b481", "#ffffff", "#ffffff", "#ffffff",
    "#ffffff", "#ffffff", "#ffffff", "#ffffff", "#ffffff", "#ffffff", "#feebd3", "#feebd3",
    "#ffffff", "#e6b481", "#feebd3", "#ffffff")


# Plot consensus space showing global component one (GC1) and global component two (GC2)

scoresPlot(result, pcol = pcol)

 

Use loadingsPlot to show the loadings for global component one of the analysis. Global component loadings are visualised by colouring the mean position of each landmark in each block. Stronger orange colours represent landmarks with larger loadings (i.e. greater amounts of variation), and stronger blue colours represent landmarks with weaker loadings. Each data block is plotted as a separate panel.

# Plot loadings for global component one (GC1)

loadingsPlot(result, cex.3d = 15)

   

Data sources

Table 1. 3D digital replicas produced from bones of modern and fossil penguins.
Accession Genus Species Age Humerus Radius Ulna Institution
AV19569 Aptenodytes forsteri Extant Right Left Right CM
2013-1-257 Aptenodytes patagonicus Extant Left Right Right CM
OR.019303 Eudyptes filholi Extant Right Right Right NMNZ
OR.017176 Eudyptes pachyrhynchus Extant Right Left Left NMNZ
OR.030196 Eudyptes robustus Extant Right Right Right NMNZ
OR.027787 Eudyptes schlegeli Extant Right Right Right NMNZ
OR.025560 Eudyptes sclateri Extant Right Right Right NMNZ
OR.018965 Eudyptula minor Extant Left Left Left NMNZ
OR.026794 Megadyptes antipodes Extant Right Right Right NMNZ
AV38751 Pygoscelis adeliae Extant Left Left Left CM
AV32995 Pygoscelis antarctica Extant Right Right Right CM
AV38694 Pygoscelis papua Extant Right Right Right CM
OR.022800 Spheniscus demersus Extant Right Right Right NMNZ
CU2138 Spheniscus humboldti Extant Right Right Right UC
OR.027329 Spheniscus magellanicus Extant Right Right Right NMNZ
OR.01451 Platydyptes novaezeelandiae 27.3-21.7 Ma Left Right Left NMNZ
GL407 Archaeospheniscus lowei 29.8-25.2 Ma Right Right Right OM
OU22065 Kairuku grebneffi 29.8-25.20 Ma Right Right Right OU
CM 2016.6.1 Sequiwaimanu rosieae 61 Ma Left Left Right CM
OU12651 Muriwaimanu tuatahi 61.3-56.0 Ma Right Left Left OU
Note:
Institution abbreviations: CM, Canterbury Museum, Christchurch, New Zealand. NMNZ, Museum of New Zealand Te Papa Tongarewa, Wellington, New Zealand. OM, Otago Museum, Dunedin, New Zealand. UC, University of Canterbury, Christchurch, New Zealand (specimen held at OU, Geology Museum, University of Otago, Dunedin, New Zealand). Specimen ages from or compiled by Slack et al. (2006), Ksepka and Ando (2011), and Ksepka et al. (2012) and literature reviewed therein.

 

References

  • Adams, D. C., Otarola-Castillo, E. (2013). Geomorph: An R package for the collection and analysis of geometric morphometric shape data. Methods in Ecology and Evolution, 4, 393-399. https://doi.org/10.1111/2041-210x.12035

  • Collyer, M. L., Davis, M. A., Adams, D. C. (2020). Making heads or tails of combined landmark configurations in geometric morphometric data. Evolutionary Biology, 47, 193-205. https://doi.org/10.1007/s11692-020-09503-z

  • Ksepka, D. T., Ando, T. (2011). Penguins past, present, and future: Trends in the evolution of the Sphenisciformes. In G. Dyke, G. Kaiser (Ed.), Living Dinosaurs: The Evolutionary History of Modern Birds, First Edition (pp. 155-186). John Wiley and Sons.

  • Ksepka, D. T., Fordyce, R. E., Ando, T., Jones, C. M. (2012). New fossil penguins (Aves, Sphenisciformes) from the Oligocene of New Zealand reveal the skeletal plan of stem penguins. Journal of Vertebrate Paleontology, 32, 235-254. https://doi.org/10.1080/02724634.2012.652051

  • Schlager, S. (2017). Morpho and Rvcg - Shape Analysis in R. In G. Zheng, S. Li and G. Szekely (Ed.), Statistical shape and deformation analysis (pp. 217-256). Academic Press.

  • Slack, K. E., Jones, C. M., Ando, T., Harrison, G. L., Fordyce, R. E., Arnason, U., Penny, D. (2006). Early penguin fossils, plus mitochondrial genomes, calibrate avian evolution. Molecular Biology and Evolution, 23, 1144-1155. https://doi.org/10.1093/molbev/msj124

  • Tenenhaus, A., Guillemot, V. (2017). RGCCA: Regularized and Sparse Generalized Canonical Correlation Analysis for multiblock data 2.1.2. https://CRAN.R-project.org/package=RGCCA.