Development and assessment of fully automated and globally transitive geometric morphometric methods, with application to a biological comparative dataset with high interspecific variation

Automated geometric morphometric methods are promising tools for shape analysis in comparative biology: they improve researchers abilities to quantify biological variation extensively (by permitting more specimens to be analyzed) and intensively (by characterizing shapes with greater fidelity). Although use of these methods has increased, automated methods have some notable limitations: pairwise correspondences are frequently inaccurate or lack transitivity (i.e., they are not defined with reference to the full sample). In this study, we reassess the accuracy of two previously published automated methods, cPDist [1] and auto3Dgm [2], and evaluate several modifications to these methods. We show that a substantial fraction of alignments and pairwise maps between specimens of highly dissimilar geometries were inaccurate in the study of Boyer et al. [1], despite a taxonomically sensitive variance structure of continuous Procrustes distances. We also show these inaccuracies can be remedied by utilizing a globally informed methodology within a collection of shapes, instead of only comparing shapes in a pairwise manner (c.f. [2]). Unfortunately, while global information generally enhances maps between dissimilar objects, it can degrade the quality of correspondences between similar objects due to the accumulation of numerical error. We explore a number of approaches to mitigate this degradation, quantify the performance of these approaches, and compare the generated pairwise maps (as well as the shape space characterized by these maps) to a \"ground truth\" obtained from landmarks manually collected by geometric morphometricians. Novel methods both improve the quality of the pairwise correspondences relative to cPDist, and achieve a taxonomic distinctiveness comparable to auto3Dgm.