An international research team has built the largest three-dimensional digital library of ants ever assembled, scanning more than 2,000 specimens to produce detailed models of nearly 800 species. The project, called Antscan, used a particle accelerator to image ethanol-preserved insects whole, capturing not just their external anatomy but internal organs, muscles, and nervous systems. Published in Nature Methods, the effort represents a new kind of biological resource, one that treats physical form as data at a scale previously reserved for genomics.
What the Antscan Library Contains
The Antscan database holds 2,193 whole-body scans spanning 792 species and 212 genera from around the world. Each specimen was preserved in ethanol and scanned in toto, meaning the entire body was captured in a single imaging session rather than dissected into parts. The result is a publicly accessible archive where researchers can rotate, slice, and measure ant bodies digitally, examining structures that would normally require destroying the physical specimen.
The 3D images go well beyond surface detail. According to the University of Maryland entomology department, the scans reveal internal structures including muscles, nervous systems, and digestive systems. Researchers have already begun using the database to study the brain, gut, and stinger across different species, turning what was once a painstaking dissection exercise into a digital comparison task.
Because the dataset is standardized, scientists can compare body plans across distant branches of the ant family tree. Subtle differences in mandibles, antennae, or thorax shape that might be missed in two-dimensional photographs become obvious when specimens can be virtually sliced at any angle. The library also preserves rare and fragile museum material in digital form, creating a high-resolution backup of specimens that may be irreplaceable.
How a Particle Accelerator Replaced Lab Scanners
Traditional micro-CT equipment, a smaller and more precise version of a hospital CT scanner, can take up to 15 hours to image a single insect, according to reporting in Science. At that pace, building a library of 2,000 specimens would take years of continuous scanning. The Antscan team sidestepped this bottleneck by using high-throughput synchrotron microtomography, essentially harnessing the intense X-ray beams generated by a particle accelerator to image specimens far faster than any benchtop scanner could manage.
Per a University of Maryland release syndicated through ScienceDaily, the team scanned approximately 2,000 specimens in a single week. That claim sits in tension with the 15-hour-per-specimen figure reported for conventional lab CT, but the discrepancy itself illustrates the point: synchrotron beamlines operate at a fundamentally different throughput tier. Each scan can be completed in minutes rather than hours, and automated sample changers keep the beamline working almost continuously.
Centralizing imaging at a large facility also meant that specimen preparation and scanning protocols could be standardized. Ants from different collections and continents were mounted and imaged under the same conditions, reducing variation that might otherwise complicate cross-species comparisons. The raw image stacks were then reconstructed into three-dimensional volumes and deposited into the public database.
Why Ant Anatomy Has Been Hard to Study at Scale
Biology has entered what the study’s authors call “the big data era,” yet the study of organismal form has been slow to capitalize on advances in imaging, as the peer-reviewed article notes. Genomics can sequence thousands of organisms in parallel, but phenomics, the systematic measurement of physical traits, has lagged behind. Ants are a case study in why. There are more than 14,000 described species, many of them tiny and structurally complex, and comparing their anatomy has traditionally required skilled taxonomists working one specimen at a time under a microscope.
Daniel Kronauer of Rockefeller University captured the difficulty bluntly in comments relayed by Maryland colleagues: “And this problem was a doozy.” Imaging tiny organisms at the resolution needed to distinguish species-level differences, while also preserving internal anatomy, demands equipment and workflows that most entomology labs simply do not have. The Antscan project addressed this gap by centralizing the scanning at a synchrotron facility and then distributing the data openly.
The initiative also depended on extensive museum partnerships. As highlighted in a University of Maryland feature, curators and field biologists contributed specimens representing major ant lineages and ecological guilds, from subterranean predators to canopy specialists. This breadth makes the library useful not just for taxonomy, but for questions about how anatomy tracks behavior, habitat, and social organization across the ant tree of life.
Open Access and the Biomedisa Platform
The datasets are hosted on an interactive repository built on the Biomedisa framework, a platform designed for biomedical image segmentation. A Nature Methods Research Briefing on the project highlights the role of open-source visualization and analysis tools in making the data usable beyond the original research team. Any scientist with a browser can access the scans, rotate them in three dimensions, and run measurements without needing specialized software licenses.
This open-access design matters because it changes who can do comparative anatomy. A researcher studying ant brain evolution in a small university lab no longer needs to secure beamtime at a synchrotron or even own a micro-CT scanner. The curated archive functions as a shared instrument, lowering the barrier to entry for morphological research across entomology, ecology, and evolutionary biology. Students can download volumes to practice segmentation, and collaborators on different continents can work from the same digital specimen instead of shipping fragile material back and forth.
Access is further streamlined through institutional login services, with Nature’s identity platform providing a route for users who authenticate via linked accounts to reach the underlying methods paper, and a similar pathway via Springer Nature credentials for the companion Biomedisa article. While the 3D volumes themselves are made broadly available, these integrations help tie the raw data back to the technical documentation that explains how it was generated.
Limits of the Library
For all its scale, the Antscan collection is not a universal answer. The images are not the final word for every research need, as Science reported. Some questions require live behavior, fine-scale ultrastructure visible only with electron microscopy, or biochemical information that CT imaging cannot provide. Ethanol preservation can also alter tissue contrast, and the voxel size chosen for high-throughput scanning inevitably trades off some resolution for speed.
The species coverage, while unprecedented, still represents only a fraction of global ant diversity. Many lineages and geographic regions remain under-sampled, and within-species variation is only sparsely captured because most taxa are represented by one or a few individuals. Researchers interested in developmental changes or caste differences within a single species will still need targeted imaging campaigns.
There are also analytical challenges. Turning three-dimensional shapes into quantitative traits requires segmentation, landmarking, and statistical modeling that are still active areas of method development. The Antscan team has demonstrated workflows for extracting features such as brain volume or mandible geometry, but scaling those pipelines to hundreds of species will demand both computational resources and community standards.
Nonetheless, the project marks a turning point. By showing that thousands of complex organisms can be imaged quickly and shared openly, Antscan offers a template for similar efforts in other groups, from beetles to fish larvae. In doing so, it nudges morphology toward the same data-rich, collaborative model that has transformed genomics, one ant at a time, but at a scale that would have been hard to imagine just a few years ago.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
