James M. Mulqueeney

The fossil record provides the most powerful evidence of large-scale biodiversity change on Earth, but it does so at coarse and often idiosyncratic temporal scales. One critical problem that arises concerns the evolutionary consequences of individual environmental experience. Individuals respond to their environment instantaneously, whereas the resolution of most fossil records aggregates multiple paleoenvironments over time scales beyond individual lifespans. Therefore, the presence of phenotypic plasticity in deep time and the extent of its influence on macroevolution remain poorly understood. Using coupled computed tomography and laser ablation inductively coupled plasma mass spectrometry protocols, we studied the environmental dependence of developmental trajectories across three sister species of macroperforate planktonic foraminifera. A foraminiferal shell preserves all stages of the individual’s ontogeny, as well as the environmental state experienced throughout its lifetime. Generalized additive mixed effect (GAMM) models show that somatic growth rates differ among the three Menardella species and that these are inversely correlated with calcification temperature, as reconstructed from Mg/Ca measurements through ontogeny. This environmental dependence varies among species: The thermal sensitivity of individual chamber-to-chamber growth rates of Menardella limbata and Menardella pertenuis is double that seen in Menardella exilis . In contrast, no such environmental signal was recovered for architectural shape traits. Our integrated approach is widely applicable and demonstrates that detecting developmental plasticity in the fossil record is feasible. Extrapolating these techniques in deep time promises to revolutionize our understanding of the ways in which environmentally associated trait variation drove the diversification of life on Earth.

The study of phenotypic evolution has been transformed in recent decades by methods allowing precise quantification of anatomical shape, in particular 3D geometric morphometrics. While this effectiveness of geometric morphometrics has been demonstrated by thousands of studies, it generally requires manual or semi-automated landmarking, which is time-consuming, susceptible to operator bias, and limits comparisons across morphologically disparate taxa. Emerging automated methods, particularly landmark-free techniques, offer potential solutions, but these approaches have thus far been primarily applied to closely related forms. In this study, we explore the utility of automated, landmark-free approaches for macroevolutionary analyses. We compare an application of Large Deformation Diffeomorphic Metric Mapping (LDDMM) known as Deterministic Atlas Analysis (DAA) with a high-density geometric morphometric approach, using a dataset of 322 mammals spanning 180 families. Initially, challenges arose from using mixed modalities (computed tomography (CT) and surface scans), which we addressed by standardising the data by using Poisson surface reconstruction that creates watertight, closed surfaces for all specimens. After standardisation, we observed a significant improvement in the correspondence between patterns of shape variation measured using manual landmarking and DAA, although differences emerged, especially for Primates and Cetacea. We further evaluated the downstream effects of these differences on macroevolutionary analyses, finding that both methods produced comparable but varying estimates of phylogenetic signal, morphological disparity and evolutionary rates. Our findings highlight the potential of landmark-free approaches like DAA for large scale studies across disparate taxa, owing to their enhanced efficiency. However, they also reveal several challenges that should be addressed before these methods can be widely adopted. In this context, we outline these issues, propose solutions based on existing literature, and identify potential avenues for further research. We argue that by incorporating these improvements, the application of landmark-free analyses could be expanded, thereby enhancing the scope of morphometric studies and enabling the analysis of larger and more diverse datasets.

The segmentation of fine-grained and complex structures from volumetric data, such as 3D biomedical images, is a manually intensive process, with performance hindered by limited training data and the difficulty of adapting AI models for specialised datasets. Here, we introduce SPROUT, a user-friendly and interpretable segmentation framework that leverages domain-specific priors, enabling experts to translate their knowledge into reproducible, high-quality segmentations across diverse imaging modalities without the need for training data. Its adaptive design facilitates parameter transferability and improves generalisation across similar datasets. Implemented as scripts and a napari plugin, it supports interactive editing and scalable batch processing, lowering the technical barrier for domain experts. We applied SPROUT to datasets spanning different imaging modalities, anatomical complexities, postures, and target structures, producing high-quality segmentations across 2D and 3D tasks. Quantitative comparisons with other methods on a representative dataset showed SPROUT achieved results comparable to expert-corrected interpolation while requiring substantially less manual input. In scenarios where SPROUT achieved high-quality results, supervised models often struggled to reach similar accuracy, highlighting the challenge of deep learning methods in complex domains. We also explored integration with foundation models to accelerate segmentation in high-contrast datasets, illustrating potential for hybrid workflows.

The rapid and repeatable characterization of individual morphology has advanced automated taxonomic classification. The most direct study of evolutionary processes is, however, not from taxonomic description, but rather of the evolution of the traits that comprise individuals and define species. Repeatable signatures of individual morphology are crucial for analysing the response to selection at scale, and thus tracking evolutionary trajectories through time and across species boundaries. Here, we introduce our 3DKMI—an open‐source MATLAB package designed for the study of morphology using three‐dimensional (3D) Krawtchouk moment invariants. The volumetric features derived from the 3D images remain stable under translation, scaling and rotation and, for an image of size 128 × 128 × 128 can be computed in less than 0.1 s. We applied our package as a case study on a collection of 300 X‐ray computed tomography scans of planktonic foraminifera specimens across five species to (1) assess the invariance of the features under different transformations and (2) analyse morphological differences among species based on the extracted characteristics. We show that 3DKMI has the capacity to efficiently and repeatedly characterize the signatures of individual morphology. In the future, we hope that the 3D feature extraction technique 3DKMI will be widely applied to digital collections to advance research in ecology and evolution.

Deep learning has emerged as a robust tool for automating feature extraction from three-dimensional images, offering an efficient alternative to labour-intensive and potentially biased manual image segmentation methods. However, there has been limited exploration into the optimal training set sizes, including assessing whether artficial expansion by data augmentation can achieve consistent results in less time and how consistent these benefits are across different types of traits. In this study, we manually segmented 50 planktonic foraminifera specimens from the genus Menardella to determine the minimum number of training images required to produce accurate volumetric and shape data from internal and external structures. The results reveal unsurprisingly that deep learning models improve with a larger number of training images with eight specimens being required to achieve 95% accuracy. Furthermore, data augmentation can enhance network accuracy by up to 8.0%. Notably, predicting both volumetric and shape measurements for the internal structure poses a greater challenge compared with the external structure, owing to low contrast differences between different materials and increased geometric complexity. These results provide novel insight into optimal training set sizes for precise image segmentation of diverse traits and highlight the potential of data augmentation for enhancing multivariate feature extraction from three-dimensional images.

The study of phenotypic evolution has been transformed by methods allowing three-dimensional quantification of anatomical form. The present state-of-the-art 3D geometric morphometrics, which relies heavily on manual landmarking, is time-consuming, prone to operator bias, and cannot effectively compare disparate shapes. Emerging automated approaches, notably landmark-free techniques, offer promise but have only been applied to closely related species. Here, we compare landmark-free approaches with high-density geometric morphometric methods on 322 mammals across 180 families. Leveraging the benefits of Poisson meshes, which combine open and closed projections, we show how landmark-free methods have greater power to differentiate major taxonomic groups. Although there is coarse correspondence in shape variation between the two methods, the finer features of the landmark-free approach likely reflect its broader sampling of the surface structure. Our study underscores the robustness of landmark-free methods for large-scale comparative analysis and helps propel morphometrics into a new era of bigger data.

Deep learning has emerged as a robust tool for automating feature extraction from 3D images, offering an efficient alternative to labour-intensive and potentially biased manual image segmentation methods. However, there has been limited exploration into the optimal training set sizes, including assessing whether artificial expansion by data augmentation can achieve consistent results in less time and how consistent these benefits are across different types of traits. In this study, we manually segmented 50 planktonic foraminifera specimens from the genus Menardella to determine the minimum number of training images required to produce accurate volumetric and shape data from internal and external structures. The results reveal unsurprisingly that deep learning models improve with a larger number of training images and that data augmentation can enhance network accuracy by up to 8.0%. Notably, predicting both volumetric and shape measurements for the internal structure poses a greater challenge compared to the external structure, due to low contrast between different materials and increased geometric complexity. These results provide novel insight into optimal training set sizes for precise image segmentation of diverse traits and highlight the potential of data augmentation for enhancing multivariate feature extraction from 3D images. Subject Category Life Sciences – Computer Science Subject Areas computational biology, Artificial Intelligence

Synopsis: Artificial intelligence (AI) is poised to revolutionize many aspects of science, including the study of evolutionary morphology. While classical AI methods such as principal component analysis and cluster analysis have been commonplace in the study of evolutionary morphology for decades, recent years have seen increasing application of deep learning to ecology and evolutionary biology. As digitized specimen databases become increasingly prevalent and openly available, AI is offering vast new potential to circumvent long-standing barriers to rapid, big data analysis of phenotypes. Here, we review the current state of AI methods available for the study of evolutionary morphology, which are most developed in the area of data acquisition and processing. We introduce the main available AI techniques, categorizing them into 3 stages based on their order of appearance: (1) machine learning, (2) deep learning, and (3) the most recent advancements in large-scale models and multimodal learning. Next, we present case studies of existing approaches using AI for evolutionary morphology, including image capture and segmentation, feature recognition, morphometrics, and phylogenetics. We then discuss the prospectus for near-term advances in specific areas of inquiry within this field, including the potential of new AI methods that have not yet been applied to the study of morphological evolution. In particular, we note key areas where AI remains underutilized and could be used to enhance studies of evolutionary morphology. This combination of current methods and potential developments has the capacity to transform the evolutionary analysis of the organismal phenotype into evolutionary phenomics, leading to an era of “big data” that aligns the study of phenotypes with genomics and other areas of bioinformatics.