ALAndrew H. Lee

Andrew H. Lee

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Andrew H. Lee has a total of 77 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Deep learning software and revised 2D model to segment bone in micro-CT scans

    by Andrew H. Lee, Ganesh Talluri, Manan Damani, Brandon Vera Covarrubias, Helena Hanna, Jeremy Chavez, Julian M. Moore, Jacob Baradarian, Michael Molgaard, Beau Nielson, et al.

    Deep learning (DL) enables automated bone segmentation in micro-CT datasets but can struggle to generalize across developmental stages, anatomical regions, and imaging conditions. We present BP-2D-03, which is a revised 2D Bone-Pores segmentation model. It was fitted to a dataset comprising 20 micro-CT scans spanning five mammalian species and 142,960 image patches. To manage the substantially larger and more varied dataset, we developed a DL software interface with modules for training (“BONe DLFit”), prediction (“BONe DLPred”), and evaluation (“BONe IoU”). These tools resolve prior issues such as slice-level data leakage, high memory usage, and limited multi-GPU support. Model performance was evaluated through three analyses. First, 5-fold cross-validation with three seeds per fold evaluated baseline robustness and stability. The model showed generally high mean Intersection-over-Union (IoU) with minimal variation across seeds, but performance varied more across folds related to differences in scan composition. These findings show that the baseline model is stable overall but that predictivity can decline for atypical scans. Second, 30 benchmarking experiments tested how model architecture, encoder backbone, and patch size influence segmentation IoU and computational efficiency. U-Net and UNet++ architectures with simple convolutional backbones (e.g., ResNet-18) achieved the highest IoU values, approaching 0.97. Third, cross-platform experiments confirmed that results are consistent across hardware configurations, operating systems, and implementations (Avizo 3D and standalone). Together, these analyses demonstrate that the BONe DL software delivers robust baseline performance and reproducible results across platforms.

  • Open Access

    Segmentation of cortical bone, trabecular bone, and medullary pores from micro‐ CT images using 2D and 3D deep learning models

    by Andrew H. Lee, Julian M. Moore, Brandon Vera Covarrubias, Leigha M. Lynch

    Computed tomography (CT) enables rapid imaging of large‐scale studies of bone, but those datasets typically require manual segmentation, which is time‐consuming and prone to error. Convolutional neural networks (CNNs) offer an automated solution, achieving superior performance on image data. In this methodology‐focused paper, we used CNNs to train segmentation models from scratch on 2D and 3D patches from micro‐CT scans of otter long bones. These new models, collectively called BONe (Bone One‐shot Network), aimed to be fast and accurate, and we expected enhanced results from 3D training due to better spatial context. Contrary to expectations, 2D models performed slightly better than 3D models in labeling details such as thin trabecular bone. Although lacking in some detail, 3D models appeared to generalize better and predict smoother internal surfaces than 2D models. However, the massive computational costs of 3D models limit their scalability and practicality, leading us to recommend 2D models for bone segmentation. BONe models showed potential for broader applications with variation in performance across species and scan quality. Notably, BONe models demonstrated promising results on skull segmentation, suggesting their potential utility beyond long bones with further refinement and fine‐tuning.


  • Open Access

    Chimerism in specimens referred to Saurophaganax maximus reveals a new species of Allosaurus (Dinosauria, Theropoda)

    by Andrew Danison, Mathew Wedel, Daniel Barta, Holly Woodward, Holley Flora, Andrew Lee, Eric Snively

    Saurophaganax maximus is the designation of a massive theropod dinosaur recovered from the Kenton Member of the Upper Jurassic Morrison Formation in Oklahoma. The theropod was originally given the name Saurophagus maximus but was later revised to Saurophaganax maximus as the former did not conform to ICZN standards. Several autapomorphies were described for S. maximus including the postorbital lacking a postorbital boss, the atlas vertebra lacking facets for a proatlas, cervical vertebrae with nearly vertical postzygapophyses, unique accessory laminae on the holotype neural arch, pneumatized post-pectoral dorsal centra, laterally bowed femora, tibiae without an astragalar buttress and more prominent distomedial crest, less distally divergent fourth metatarsals, and craniocaudally expanded chevrons. However, our re-evaluation of these specimens shows that some of this material, including the holotype, is more parsimoniously referred to diplodocid sauropods found in the same quarry rather than an allosaurid. Because saurischian material referred to S. maximus is likely chimeric but diagnostic, we propose taxonomic revision and hypothesize dual validity. The name Saurophaganax maximus should refer to a previously undiagnosed diplodocid, instead of the large theropod recovered from the Kenton 1 Quarry. The presence of at least one skeletally mature theropod was confirmed through paleohistology of a fourth metatarsal. Despite the similarity of the theropod material to known species of Allosaurus, some decisively theropod specimens feature subtle autapomorphies that suggest a separate species, which we describe as Allosaurus anax sp. nov.


  • Open Access

    Development of limb bone laminarity in the homing pigeon (Columba livia)

    by Rylee S. McGuire, Raffi Ourfalian, Kelly Ezell, Andrew H. Lee

    Background: Birds show adaptations in limb bone shape that are associated with resisting locomotor loads. Whether comparable adaptations occur in the microstructure of avian cortical bone is less clear. One proposed microstructural adaptation is laminar bone in which the proportion of circumferentially-oriented vascular canals (i.e., laminarity) is large. Previous work on adult birds shows elevated laminarity in specific limb elements of some taxa, presumably to resist torsion-induced shear strain during locomotion. However, more recent analyses using improved measurements in adult birds and bats reveal lower laminarity than expected in bones associated with torsional loading. Even so, there may still be support for the resistance hypothesis if laminarity increases with growth and locomotor maturation.Methods: Here, we tested that hypothesis using a growth series of 17 homing pigeons (15–563 g). Torsional rigidity and laminarity of limb bones were measured from histological sections sampled from midshaft. Ontogenetic trends in laminarity were assessed using principal component analysis to reduce dimensionality followed by beta regression with a logit link function.Results: We found that torsional rigidity of limb bones increases disproportionately with growth, consistent with rapid structural compensation associated with locomotor maturation. However, laminarity decreases with maturity, weakening the hypothesis that high laminarity is a flight adaptation at least in the pigeon. Instead, the histological results suggest that low laminarity, specifically the relative proportion of longitudinal canals aligned with peak principal strains, may better reflect the loading history of a bone.


  • Open Access

    The effects of growth rate and biomechanical loading on bone laminarity within the emu skeleton

    by Amanda L. Kuehn, Andrew H. Lee, Russell P. Main, Erin L.R. Simons

    The orientation of vascular canals in primary bone may reflect differences in growth rate and/or adaptation to biomechanical loads. Previous studies link specific canal orientations to bone growth rates, but results between different taxa are contradictory. Circumferential vascular canals (forming laminar bone) have been hypothesized to reflect either (or both) rapid growth rate or locomotion-induced torsional loading. Previous work on the hindlimb biomechanics in the emu shows that the femur and tibiotarsus experience large shear strains, likely resulting from torsional loads that increase through ontogeny. Here, we test how growth rate and biomechanical loading affect bone laminarity in wing and hindlimb elements from growing emu (2–60 wks). If laminar bone is an adaptation to torsion-induced shear strains, it should increase from juveniles to adults. Alternatively, if bone laminarity reflects rapid growth, as has been shown previously in emu, it should be abundant in fast-growing juveniles and decrease with age. Transverse mid-shaft histological sections from the limb bones (femur, tibiotarsus, humerus, ulna, and radius) were prepared and imaged. Growth rates were measured using fluorescent bone labels. Vascular canal orientation was quantified using laminarity index (proportion of circumferential canals). Principal components analysis was performed to convert highly correlated variables (i.e., mass, age, growth rate, and shear strain) into principal components. Random-intercept beta regression modeling determined which principal components best explained laminarity. The fastest growth rates were found in young individuals for all five skeletal elements. Maximum growth rate did not coincide with peak laminarity. Instead, in the femur and tibiotarsus, elevated laminarity is strongly correlated with adult features such as large size, old age, and modest growth rate. This result is contrary to predictions made based on a previous study of emu but is consistent with results observed in some other avian species (penguin, chicken). Shear strain in the caudal octant of the femur and tibiotarsus is positively correlated with laminarity but has a weaker effect on laminarity relative to mass, age, and growth rate. Laminarity in the wing elements is variable and does not correlate with ontogenetic factors (including mass, age, and growth rate). Its presence may relate to relaxed developmental canalization or a retained ancestral feature. In conclusion, ontogeny (including growth rate) is the dominant influence on vascular canal orientation at least in the hindlimb of the emu.


  • Open Access

    Phylogeny and herbivory are related to avian cecal size

    by Andrew Hunt, Layla Al-Nakkash, Andrew H. Lee, Heather F. Smith

    Avian ceca, a pair of blind sacs arising from the junction of the ileum and colon, are homologous to the cecum in mammals. Cecal size is hypothesized to depend on dietary proclivities and pressures, with faunivorous species having short ceca, whereas herbivorous species have long ceca. Previous tests of this hypothesis, however, did not account for phylogenetic pseudoreplication among closely related taxa. We collated published data on cecal length, dietary category, flying ability, and body mass from 155 avian taxa. Character states were mapped onto a phylogenetic framework, and the permutation tail probability test was used to detect phylogenetic signal in each character. Phylogenetic signal is significant among the characters. As with the cecoappendicular complex in mammals, closely-related birds tend to have similar cecal length. To account for phylogenetic pseudoreplication, we performed phylogenetic generalized least squares regression on cecal length and body mass with dietary category, superordinal-level clade, and flying ability as cofactors. The best-fitting regression model supports the dietary hypothesis for the avian cecum. Among sampled birds of comparable body mass, mean cecal length is significantly longer in herbivorous species than in carnivorous ones (p = 0.008), presumably allowing the extraction of nutrients without the burden of fermenting bulky masses of dietary fiber. Exceptions to this trend, however, suggest that avian ceca are functionally complex and may have additional roles in water balance and nitrogen recycling.


  • Open Access

    Biomechanical evolution of solid bones in large animals: a microanatomical investigation

    by Alexandra Houssaye, Katja Waskow, Shoji Hayashi, Raphaël Cornette, Andrew H. Lee, John R. Hutchinson


  • Open Access

    Wing bone laminarity is not an adaptation for torsional resistance in bats

    by Andrew H. Lee, Erin L.R. Simons

    Torsional loading is a common feature of skeletal biomechanics during vertebrate flight. The importance of resisting torsional loads is best illustrated by the convergence of wing bone structure (e.g., long with thin walls) across extant bats and birds. Whether or not such a convergence occurs at the microstructural level is less clear. In volant birds, the humeri and ulnae often contain abundant laminar bony tissue in which primary circumferential vascular canals course concentrically about the long axis of the bone. These circumferential canals and the matrix surrounding them presumably function to resist the tissue-level shear stress caused by flight-induced torsion. Here, we assess whether or not laminar bone is a general adaptive feature in extant flying vertebrates using a histological analysis of bat bones. We sampled the humeri from six adult taxa representing a broad phylogenetic and body size range (6-1,000 g). Transverse thick sections were prepared from the midshaft of each humerus. Bone tissue was classified based on the predominant orientation of primary vascular canals. Our results show that humeri from bats across a wide phylogenetic and body size range do not contain any laminar bone. Instead, humeri are essentially avascular in bats below about 100 g and are poorly vascularized with occasional longitudinal to slightly radial canals in large bats. In contrast, humeri from birds across a comparable size range (40-1,000 g) are highly vascularized with a wide range in bone laminarity. Phylogenetically-informed scaling analyses reveal that the difference in vascularity between birds and bats is best explained by higher somatic relative growth rates in birds. The presence of wing bone laminarity in birds and its absence in bats suggests that laminar bone is not a necessary biomechanical feature in flying vertebrates and may be apomorphic to birds.


  • Open Access

    Diversity of Limb-Bone Safety Factors for Locomotion in Terrestrial Vertebrates: Evolution and Mixed Chains

    by Richard W. Blob, Nora R. Espinoza, Michael T. Butcher, Andrew H. Lee, Angela R. D’amico, Faraz Baig, K. Megan Sheffield

    During locomotion over land, vertebrates’ limb bones are exposed to loads. Like most biological structures, limb bones have a capacity to withstand greater loads than they usually experience, termed a safety factor (SF). How diverse are limb-bone SFs, and what factors correlate with such variation? We have examined these questions from two perspectives. First, we evaluated locomotor SF for the femur in diverse lineages, including salamanders, frogs, turtles, lizards, crocodilians, and marsupials (opossums). Comparisons with values for hind-limb elements in running birds and eutherian mammals indicate phylogenetic diversity in limb-bone SF. A high SF (∼7) is primitive for tetrapods, but low magnitudes of load and elevated strength of bones contribute to different degrees across lineages; moreover, birds and eutherians appear to have evolved lower SFs independently. Second, we tested the hypothesis that SFs would be similar across limb bones within a taxon by comparing data from the humerus and femur of alligators. Both in bending and in torsion, we found a higher SF for the humerus than for the femur. Such a “mixed chain” of different SFs across elements has been predicted if bones have differing variabilities in load, different costs to maintain, or high SF values in general. Although variability in load is similar for the humerus and femur, a high SF may be less costly for the humerus because it is smaller than the femur. The high SFs of alligators also might facilitate differences in SF among their limb bones. Beyond these specific findings, however, a more general implication of our results is that evaluations of the diversity of limb-bone SFs can provide important perspective to direct future research. In particular, more complete understanding of variation in SF could provide insight into factors that promoted the evolutionary radiation of terrestrial locomotor function in vertebrates.


  • Open Access

    Microanatomical and Histological Features in the Long Bones of Mosasaurine Mosasaurs (Reptilia, Squamata) – Implications for Aquatic Adaptation and Growth Rates

    by Alexandra Houssaye, Johan Lindgren, Rodrigo Pellegrini, Andrew H. Lee, Damien Germain, Michael J. Polcyn

    During their evolution in the Late Cretaceous, mosasauroids attained a worldwide distribution, accompanied by a marked increase in body size and open ocean adaptations. This transition from land-dwellers to highly marine-adapted forms is readily apparent not only at the gross anatomic level but also in their inner bone architecture, which underwent profound modifications. The present contribution describes, both qualitatively and quantitatively, the internal organization (microanatomy) and tissue types and characteristics (histology) of propodial and epipodial bones in one lineage of mosasauroids; i.e., the subfamily Mosasaurinae. By using microanatomical and histological data from limb bones in combination with recently acquired knowledge on the inner structure of ribs and vertebrae, and through comparisons with extant squamates and semi-aquatic to fully marine amniotes, we infer possible implications on mosasaurine evolution, aquatic adaptation, growth rates, and basal metabolic rates. Notably, we observe the occurrence of an unusual type of parallel-fibered bone, with large and randomly shaped osteocyte lacunae (otherwise typical of fibrous bone) and particular microanatomical features in Dallasaurus, which displays, rather than a spongious inner organization, bone mass increase in its humeri and a tubular organization in its femora and ribs. The dominance of an unusual type of parallel-fibered bone suggests growth rates and, by extension, basal metabolic rates intermediate between that of the extant leatherback turtle, Dermochelys, and those suggested for plesiosaur and ichthyosaur reptiles. Moreover, the microanatomical features of the relatively primitive genus Dallasaurus differ from those of more derived mosasaurines, indicating an intermediate stage of adaptation for a marine existence. The more complete image of the various microanatomical trends observed in mosasaurine skeletal elements supports the evolutionary convergence between this lineage of secondarily aquatically adapted squamates and cetaceans in the ecological transition from a coastal to a pelagic lifestyle.