The Inaugural Gordon Research Symposium on Volume Electron Microscopy

Reported by Odara Medagedara

Chair, Inaugural Gordon Research Symposium for Volume Electron Microscopy 2025

The first Gordon Research Symposium (GRS) on Volume Electron Microscopy (vEM), held the weekend before the 2025 Gordon Research Conference (GRC) on vEM, brought together early-career researchers from across the field to share their work and connect with peers. As the first GRS for the vEM field, it provided a valuable space for emerging scientists to present research, exchange ideas, and engage in discussions in a welcoming setting.

Chaired by myself and my co-chair, Kara Fulton (Harvard Medical School), the symposium was designed to highlight innovative research, facilitate deep technical and conceptual discussions, and create space for early-career voices. As a third-year PhD student organizing my first conference, this experience was a deeply rewarding challenge. I’m incredibly grateful for the enthusiasm, generosity, and insight everyone brought to the weekend.

The Symposium began with an engaging keynote on high-resolution and high-throughput 3D EM delivered by Professor Harald Hess. What followed was a diverse program of talks spanning a wide range of vEM applications, including synaptic connectivity in the brain, ultrastructural changes in cancer, imaging of marine plankton biodiversity, and advances in automated analysis and data integration. Breaking up the talks were poster sessions which were lively and interactive, and it was exciting to see how engaged everyone was with each other’s work.

One standout highlight was our career development panel, which brought together a fantastic group of speakers to discuss pathways in vEM – from core facilities to academic research. The conversation was honest, eye-opening, and full of practical advice. Many attendees shared that this session was one of the most impactful moments of the Symposium. For this reason, I would very much like to thank all our panelists for their openness and thoughtfulness.

Another encouraging sign of the enthusiasm within this community was the election for the next GRS (vEM) chairs. We had six early-career researchers step forward to run – a fantastic show of commitment to keeping this momentum going. We’re excited to announce that Yuxin Zhang (The Francis Crick Institute, United Kingdom) and Chandni Bhickta (European Molecular Biology Laboratory Heidelberg, Germany) will co-chair the next GRS (vEM) in 2027. We’re already looking forward to seeing how they shape the next edition of the Symposium.

I also want to extend heartfelt thanks to everyone who participated and contributed – our speakers, discussion leaders, poster presenters, and every attendee who brought their curiosity and energy. The collaborative spirit of this community made the 2025 GRS (vEM) not just successful, but truly memorable.

I’m excited to see where this momentum takes the vEM community next!

empanada new version released!

By Kedar Narayan

Thank you for your use of empanada (and feedback) - we are at >4K downloads and counting. Here's something for you: empanada v1.2 is released! 

https://empanada.readthedocs.io/en/latest/ 

1. Two new models: DropNet and NucleoNet

• DropNet for instance segmentation of lipid droplets. DropNet is a panoptic deep lab (PDL) model trained on lipid droplets from various tissues and sample prep protocols (including typical artifacts arising from them). It is also trained to segment round secretory vesicles from acinar cells (for example salivary gland) but to ignore islet granules, i.e. from alpha beta and delta pancreatic cells, as well as empty round vesicles. It does well on reasonably “typical” lipid droplets but struggles with extremely large droplets or previously unseen/rare artifacts.
• NucleoNet for instance, segmentation of nuclei. NucleoNet is a base PDL model trained on a heterogenous dataset of nuclei. It does well especially on “typical” nuclei in adherent, non-adherent and tissue samples. NucleoNet is not trained on dividing cells or dead/dying cells. If you have very unusual or highly convoluted nuclei it may struggle on account of the large size of the nuclei and some inherent limitations with PDL; however, it will likely benefit from fine-tuning in this case. 
• Please consider using the “pick patches” module to capture unusual nuclei or lipid droplets where the models underperform (please also create GT for the features - we don’t know your biology!) and contact us at nci-cmm@mail.nih.gov

New video! - AI Segmentation in Volume Electron Microscopy

Segmentation is a critical step in analyzing and visualizing 3D volume electron microscopy data. Learn how semantic, instance, and panoptic segmentation help isolate cellular structures such as mitochondria, nuclei, and cytoplasm within complex biological volumes. Video created by Falconieri Visuals.

🎥 Watch the video here. Feel free to use it in your talks.

The Microscopists Podcasts are now featured on Volume EM Website

Global vEM Facility Map:

Tell Us Where You Are!

Submit your vEM Facility details to be included on the international map. Include the techniques available, web site, and facility manager's information and get connected with the global vEM community.