Vermeer: Autoregressive generative modeling of microscopy predicts protein localization
Vermeer is an autoregressive generative model that predicts protein localization microscopy from sequences and cell landmarks, enabling zero-shot transfer across imaging conditions.
Published 2026Sydney Poster Session 4 · Wed, Dec 9, 5:00 PM–8:00 PM local time · Hall 1-4OpenReview ↗
Only vote on papers you've read. Sign in with GitHub to vote.
Vermeer delivers a sharp, sequence-conditioned advance in generative microscopy that targets real antibody bottlenecks, though its claims of substantially improved biological fidelity and zero-shot transfer remain vague without clearer assay metrics and wet-lab validation.
Abstract
Abstract Fluorescent microscopy provides a rich view into how proteins localize within cells, but it remains experimentally infeasible to image human proteins across all of the different factors that can impact localization. We introduce Vermeer , a channel-adaptive autoregressive generative model for in silico generation of microscopy images of protein localization. Vermeer conditions generations on protein sequences and landmark stains showing the morphology of cells, which enables it to generalize to unseen proteins and cell lines. We show that Vermeer, trained on the Human Protein Atlas, can generate images with substantially improved perceptual quality and biological fidelity over previous proposals. Additionally, Vermeer’s autoregressive framework enables flexible generation using varying channel subsets and orderings, enabling zero-shot transfer to data collected under different imaging conditions and channel configurations than those used for training. These results position Vermeer to enable scalable modeling of protein localization and is a step towards generative foundation models that can operate over distinct microscopy datasets. Code is publicly accessible at https://github.com/microsoft/vermeer.git .