Kyle Harrington

@kyleharrington.com

Images, AI, ML, and ALife. Former: @hampshirecolg @brandeisuniversity.bsky.social @harvardmed.bsky.social @uidaho.bsky.social @hhmijanelia.bsky.social @mdc-berlin.bsky.social @ORNL. Now @biohub.org. Opinions are mine.

📣 One of the best parts of #SciPy2026 is seeing the Scientific Python ecosystem evolve in real time. Today's Tools Plenary featured updates from the teams behind Zarr, Xarray, Tiled, Pangeo, scikit-image, Matplotlib, and napari, showcasing new features & future directions 🫶

BildBildBildBild

🚨Open call to the Cell Tracking crowd out there🚨! Loïc's team at @biohub.org opened our Kaggle Challenge for 3D cell tracking, with $60k of prize money💰. Spread the news, delve into our gigabytes of imaging (zarr3) and cell tracking (geff) data of developing zebrafish embryos, and start tracking 🤩

Loïc A. Royer 💻🔬🧪@loicaroyer.bsky.social · last mo.

1/ 🧬🔬 How does one cell become an entire animal? We can now film it in 3D 🐟, but turning those movies into accurate cell tracks is brutally hard. Enter our new #Kaggle challenge: $60K prize pool, to help crack 3D+time cell tracking at scale! 🧵👇 www.kaggle.com/competitions... @biohub.org

6/ 💻 You don't need to be a biologist. If you're into computer vision, object detection, segmentation, or graph optimization 📊, this is your problem. It's a code competition (notebooks, GPU/CPU ≤12h). Starter data and a sample submission are live now.

5/ 🧠 Why it matters: accurate single-cell lineages power developmental biology, immunology, and disease research 🩺, letting us ask where every cell comes from and where it ends up as a body takes shape. Better tracking = faster, more reproducible science.

4/ 🗺️ This is the problem behind Zebrahub (the Biohub's "Google Earth" of zebrafish development) and Ultrack 📄, our method in Nature Methods 2025 that topped the Cell Tracking Challenge on dense 3D embryos. Now it's your turn to push the state of the art!

3/ ⚠️ Why it's hard: high cell density, imaging noise, and irregular shapes break automated trackers, and small per-frame errors compound into broken lineages. 🪢 Today, scientists fix this by hand, with countless hours of manual tracking. That's the bottleneck.

2/ 🧫 The data: 3D+time light-sheet movies of fluorescently labeled zebrafish embryo cells (~88 GB training data, stored as Zarr). Thousands of similar-looking nuclei moving, deforming, and dividing. Biggest Dataset EVER! Your task: detect every cell, link it across time, and catch every division. ⏱️

Bild

Cryo-ET is often framed as a tool for in situ protein structure. But what if the real revolution is contextualization? I explore how #teamtomo is redefining what "local" means in organelle biology, revealing membrane states rather than just protein structures. tinyurl.com/localmembrane

Think globally, act locally: Redefining organellar membrane environments through cryo-electron tomography

Early enthusiasm for the “cellular revolution” in cryo-electron tomography (cryo-ET) was largely driven by the promise of resolving protein structures…

sciencedirect.com

In a series of 3 papers and preprints, we’re thrilled to share with you the working laser phase plate. In collaboration with research led by Holger Müller at UC Berkeley, this is a huge innovation in imaging to make small and faint objects inside cells visible. bit.ly/4vK9LVn

Waves of laser light coming from four different directions and meeting in the middle brightly inside the cavity of the laser phase plate

Proteins are the machinery of life. Billions of their sequences have been cataloged—but the biology behind most is still unknown. Today we're releasing ESMFold2, ESMC and ESM Atlas to help change that. Open source, open to every scientist. biohub.ai/esm/protein