Open access to annotated cryoET tomograms

Currently, annotating tomograms from cryo-electron tomography (cryoET) experiments is a tedious, time-consuming, and often manual process. Our goal is to accelerate this process by catalyzing the development of sophisticated machine-learning methods for automatic annotation, helping researchers find scientific insights faster.

The portal provides biologists and developers open access to high-quality, standardized, annotated data they can readily use to retrain or develop new annotation models and algorithms. Currently, the portal contains 15,138 tomograms from 298 datasets contributed by the groups of Julia Mahamid, Jürgen Plitzko, David Agard, John Briggs, Abhay Kotecha, Ben Engel, Danielle Grotjahn, and Grant Jensen.

All tomograms include rich standardized metadata such as data tree structure and naming conventions. Most groups have already provided annotations for their data, and there is a structure to add new annotations to existing tomograms.

We are actively growing the number of annotated datasets on the portal and encourage researchers to share their data. We are working with EMPIAR to host the data and support annotation.

Ultimately, our vision is to contribute to developing a large, open-access database of annotated and validated 3D structural information for cells that researchers can use to gain new insights into cellular and structural biology.