MODERNISING AND EXPANDING BRAINGLOBE’S NEUROANATOMICAL ATLAS INFRASTRUCTURE

The BrainGlobe Initiative aims to facilitate the creation of a community-driven software ecosystem for computational neuroanatomy. This is underpinned by the BrainGlobe Atlas API. The BrainGlobe Atlas API gives standardised, programmatic access to 200 neuroanatomical atlases, spanning a range of imaging modalities, developmental time points and species. The core development team has recently worked on improving the Atlas API and its related software tools and data.

We have moved BrainGlobe’s atlas data to a commercial cloud provider, making downloads faster and more reliable. We also converted the atlas data to OME-Zarr, a next-generation chunked file format for image data tailored to cloud applications. OME-Zarr allows the BrainGlobe Atlas API to fetch individual atlas components on-the-fly, making it more flexible and more reliable. Together, these changes have ultimately improved the user experience. The API itself has remained backwards-compatible.

To facilitate mapping data between coordinate spaces defined by various atlases, we have developed “brainglobe-ccf-translator”. This tool currently provides an API to transform data across ten different mouse brain atlases.

Beyond this, we have created novel tools for anatomical template construction, and applied these to make brain atlases for emerging model organisms as well as for the rat using two-photon serial tomography.

Together, these tools underpin anatomical research software in systems neuroscience.