Life Science Practical Magic

A new AI can spot every single protein inside a human cell using just a few basic landmarks.

April 3, 2026

Original Paper

Generative machine learning unlocks the first proteome-wide image of human cells

Sun, H.; Kahnert, K.; Hansen, J. N.; Leineweber, W. D.; Li, M.; Feng, W.; Ballllosera Navarro, F.; Axelsson, U.; Ouyang, W.; Lundberg, E.

bioRxiv · 10.64898/2026.03.31.715748

AI-generated illustration

The Takeaway

Mapping all 12,000+ proteins in a cell would normally take years of experiments, but this generative model can reconstruct the entire spatial map instantly. It can even detect how drugs change a cell's internal structure just by looking at the cell's outer shape.

From the abstract

The spatial organization of proteins within cells governs virtually all cellular functions. Yet, current imaging technologies can simultaneously visualize only tens of proteins, orders of magnitude below the thousands that populate a single human cell. Here, we present ProtiCelli, a deep generative model that simulates microscopy images for 12,800 human proteins from just three cellular landmark stains. Trained on 1.23 million images from the Human Protein Atlas, ProtiCelli outperforms existing