High-throughput phenomics of global ant biodiversity

Julian Katzke, F. Javier García, Jacob J. Relle, Fumika Azuma, Tomáš Faragó, Lazzat Aibekova, Alexandre Casadei-Ferreira, Shubham Gautam, Adrian Richter, Evropi Toulkeridou, Sabine Bremer, Elias Hamann, Jenny Hein, Janes Odar, Chandan Sarkar, Jacobus J. Boomsma, Rodrigo M. Feitosa, Lukas Schrader, Guojie Zhang, Sándor Csősz, Minsoo Dong, Olívia Evangelista, Georg Fischer, Brian L. Fisher, Jaime A. Florez-Fernandez, Serge Aron, Abel Bernadou, Martín Bollazzi, Raphaël Boulay, Sylvia Cremer, Heike Feldhaar, Foitzik Susanne, Erik T. Frank, Jürgen Gadau, Daniele Giannetti, Stéphane De Greef, Heikki Helanterä, Ana Ješovnik, Andrew V. Suarez, Bálint Markó, David R. Nash, Jérôme Orivel, Jes Søe Pedersen, Frédéric Petitclerc, Stephen A. Rehner, Helen Sindre, András Tartally, Kazuki Tsuji, Irène Villalta, Herbert C. Wagner, Kiko Gómez, Donató A. Grasso, Benoit Guénard, Peter G. Hawkes, R. Roy Johnson, Roberto A. Keller, Rasmus Stenbak Larsen, Timothy A. Linksvayer, Cong Liu, Arthur Matte, Masako Ogasawara, Hao Ran, Juanita Rodríguez, Enrico Schifani, Schultz Ted R., Jonathan Z. Shik, Jeffrey Sosa‐Calvo, Chao Tong, Leonardo Tozetto, Seonwoo Yoon, Masashi Yoshimura, Jie Zhao, Tilo Baumbach, Evan P. Economo, Thomas van de KampNature Methods 2026 · DOI

TL;DR

Researchers have created "Antscan," a massive, publicly accessible database of 3D ant models generated using high-throughput synchrotron X-ray microtomography. This resource democratizes the study of ant morphology, enabling unprecedented comparative analyses of ant evolution, structure, and function, and paving the way for similar initiatives in other organism groups.

What Problem Does This Paper Address?

The "big data" revolution has transformed many areas of biology, but the study of organismal form (phenomics) has lagged behind due to limitations in imaging and computational throughput. Digitizing entire organisms in 3D offers immense potential for understanding biodiversity, evolution, and ecological roles. However, existing methods for capturing detailed 3D morphological data are often slow, require extensive specimen preparation, and result in data that is difficult to compare across studies. This bottleneck has hindered the large-scale documentation of morphological diversity, especially for species-rich groups like insects, which play crucial roles in ecosystems.

Methods

The core of this work is the "Antscan" initiative, which leverages high-throughput synchrotron X-ray microtomography (micro-CT) to overcome the throughput limitations of traditional methods. Key aspects of their approach include:

  • Synchrotron Micro-CT: Utilizing the high flux density of synchrotron X-rays allows for rapid scanning of specimens with high resolution. This technique also enables phase contrast imaging, which is crucial for visualizing soft tissues without the need for staining, thus preserving specimens.
  • Automated Workflow: A standardized protocol involving robotic sample exchange, automated scanning, and GPU-based tomographic reconstruction was implemented to maximize efficiency.
  • Data Standardization: Scans were acquired with standardized parameters across multiple beamlines, ensuring comparability of datasets. This is critical for enabling automated analysis and machine learning applications.
  • Data Processing Pipeline: A multi-step pipeline was developed for processing the raw tomographic data. This includes merging scans from multiple height steps (for larger specimens), automated segmentation to remove background, and generation of 3D surface models.
  • Open Science Platform: The Antscan data is made publicly available through an interactive online repository (Antscan.info and Biomedisa.info), with associated metadata and downloadable datasets under a CC BY 4.0 license.

Datasets

The Antscan initiative has generated a new, large-scale dataset:

  • Antscan Database: This database contains 2,193 whole-body 3D ant datasets.
  • Taxonomic Coverage: The datasets represent 212 genera and 792 species, covering 14 out of 16 recognized ant subfamilies and over 90% of all described ant species.
  • Specimen Types: Includes 1,671 workers, 291 queens, and 220 males, with some specimens identified as media or major workers.
  • Associated Data: 186 scanned species are linked to genomic data, with 157 scans from the same nest series as sequenced specimens.
  • Outgroups: 32 scans of non-ant wasps are included for comparative purposes.
  • Existing Collections: The study utilized specimens from museums and personal collections worldwide, primarily ethanol-preserved.

Key Insights and Ideas

  • Democratizing Phenomics: By making high-resolution 3D morphological data freely available, Antscan significantly lowers the barrier to entry for researchers, educators, and enthusiasts interested in ant morphology and evolution.
  • Bridging Genomics and Morphology: The direct pairing of phenomic data with genome sequencing projects (e.g., Global Ant Genomics Alliance) creates a powerful resource for understanding the genetic basis of morphological diversity and adaptation.
  • Scalability of Synchrotron Micro-CT: The success of Antscan demonstrates that high-throughput synchrotron micro-CT is a viable and scalable approach for digitizing large collections of small organisms, opening doors for similar projects in other insect groups and beyond.
  • Unlocking Hidden Patterns: Even simple screening of the Antscan data can reveal patterns in morphological and ecological adaptations that were previously difficult to discern. The example of biomineral armor distribution highlights this potential.
  • Long-Term Data Preservation: The use of standardized protocols and open-access repositories ensures the long-term availability and usability of the generated data.

Key Results

  • Massive 3D Dataset: Creation of a database with 2,193 3D ant datasets, representing a significant portion of global ant diversity.
  • High Throughput: The automated synchrotron micro-CT workflow achieved a scanning pace of approximately 25 scans per hour, a dramatic improvement over conventional micro-CT.
  • Comprehensive Phylogeny: The dataset covers a broad phylogenetic range of ants, enabling evolutionary studies across major lineages.
  • Data Accessibility: All datasets, metadata, and associated code are publicly available, promoting open science and reproducibility.
  • Demonstrated Utility: The paper showcases exemplary use cases, including detailed segmentation of anatomical features and comparative screening for traits like biomineral armor, demonstrating the immediate value of the dataset.

Limitations and Future Work

  • Specimen Preservation: While ethanol-preserved specimens are generally suitable, some showed signs of shrinkage and decay, which could affect certain morphometric analyses.
  • Partial Truncation: Due to the manual positioning of specimens and limitations in field of view, some datasets exhibit partial truncation of appendages. Future work aims to improve robotic setups and imaging pipelines for autonomous recognition and readjustment.
  • Data Processing Bottlenecks: While automation is advancing, processing large volumes of 3D data remains computationally intensive. Continued investment in computing infrastructure and workflow automation is needed.
  • Need for Advanced Bioinformatics: Unlocking the full potential of micro-CT data requires further development of bioinformatics tools, particularly in automated image analysis and segmentation.
  • Expansion to Other Organisms: The Antscan workflow is presented as a scalable approach that can be adapted for digitizing other lineages of small organisms, suggesting future applications beyond ants.

Summary updated

Abstract

The big data era in biology is underway, but the study of organismal form has been slow to capitalize on advances in imaging and computation. Imaging approaches can digitize whole organisms, but low throughput has limited the effort to document morphological diversity. Here, within the open science initiative 'Antscan', we applied high-throughput synchrotron X-ray microtomography to capture phenotypes across a diverse and ecologically dominant insect group: ants. At https://www.antscan.info , we provide 2,193 whole-body three-dimensional ant datasets from 212 genera and 792 species to broadly cover the ant phylogeny with a global scope, also pairing phenomic data with genome sequencing projects. Scans acquired with standardized parameters facilitate automated analysis, and free access to data can broaden the audience and incentivize methods development. Antscan presents a scalable approach to create libraries of diverse anatomies, heralding an era of studies on the evolution, structure and function of organismal phenotypes.
BibTeX
@inproceedings{katzke2026high,
  author = {Julian Katzke and F. Javier García and Jacob J. Relle and Fumika Azuma and Tomáš Faragó and Lazzat Aibekova and Alexandre Casadei-Ferreira and Shubham Gautam and Adrian Richter and Evropi Toulkeridou and Sabine Bremer and Elias Hamann and Jenny Hein and Janes Odar and Chandan Sarkar and Sabine Bremer and Jacobus J. Boomsma and Rodrigo M. Feitosa and Lukas Schrader and Guojie Zhang and Sándor Csősz and Minsoo Dong and Olívia Evangelista and Georg Fischer and Brian L. Fisher and Jaime A. Florez-Fernandez and Serge Aron and Abel Bernadou and Martín Bollazzi and Raphaël Boulay and Sylvia Cremer and Heike Feldhaar and Foitzik Susanne and Erik T. Frank and Jürgen Gadau and Daniele Giannetti and Stéphane De Greef and Heikki Helanterä and Ana Ješovnik and Andrew V. Suarez and Bálint Markó and David R. Nash and Jérôme Orivel and Jes Søe Pedersen and Frédéric Petitclerc and Stephen A. Rehner and Helen Sindre and András Tartally and Kazuki Tsuji and Irène Villalta and Herbert C. Wagner and Fede García and Kiko Gómez and Donató A. Grasso and Stéphane De Greef and Benoit Guénard and Peter G. Hawkes and R. Roy Johnson and Roberto A. Keller and Rasmus Stenbak Larsen and Timothy A. Linksvayer and Cong Liu and Arthur Matte and Masako Ogasawara and Hao Ran and Juanita Rodríguez and Enrico Schifani and Schultz Ted R. and Jonathan Z. Shik and Jeffrey Sosa‐Calvo and Chao Tong and Leonardo Tozetto and Seonwoo Yoon and Masashi Yoshimura and Jie Zhao and Tilo Baumbach and Evan P. Economo and Thomas van de Kamp},
  title = {{High-throughput phenomics of global ant biodiversity}},
  booktitle = {Nature Methods},
  year = {2026},
  doi = {10.1038/s41592-026-03005-0},
}
Figure