OMERO-Screen
One reproducible workflow from a 96-well plate to the figure in your paper.
AI / Agents
Developers
Community
Meta
Requires: Python >=3.12,<3.15
Provides-Extra: gui
Package Info
End to end high content bio-image analysis powered by AI
OMERO-Screen takes a 96-well immunofluorescence plate from an OMERO server to a publication figure: segmentation, feature extraction, cell-cycle assignment, storage, classification and plotting as one reproducible workflow. It is built for cell biologists running high-content screens, not for image-analysis specialists.
Active development. OMERO-Screen is being developed alongside ongoing research, so expect frequent changes and occasional breaking ones. A stable release is planned for later in 2026. Pin a commit if you need an analysis to stay reproducible.
What makes it different
Rapid generation and adaptation of new models
Deep learning models for segmentation, tracking and classification have revolutionised bioimage analysis, but remain difficult for experimentalists to access. OMERO-Screen builds on Cellpose and Trackastra, and adds a rapid generation-and-deployment loop for classifiers. Its modular architecture means models are selected from the command line (--model, --track <model>), so an improved model can be swapped into the pipeline without touching the analysis code.
4i / cyclic IF and automated tracking, without the coding
More advanced image analysis usually demands coding skills that most cell biology labs do not have. OMERO-Screen integrates automated registration and tracking pipelines, generating highly multiplexed single-cell data from a few simple commands.
Agentic AI integration
Because OMERO-Screen is driven entirely by terminal commands, it is directly accessible to AI agents. We ship agent skills that run the whole pipeline — from plate analysis through quantification to figure production — from natural language prompts.
Data integration and management
High-content IF data tends to grow exponentially and end up in chaos: thousands of images, analysis spread across multiple tools and scattered files, and a final figure that nobody can reproduce exactly. OMERO-Screen keeps raw data, masks and analysis files together on an OMERO server. Screen images and segmentation masks are viewed through local Zarr stores and napari plugins, and single-cell data are queried from a local DuckDB database. This keeps terabytes of high-content image analysis data in one well-organised, queryable form, and keeps published results reproducible.
Start here
- Install and configure — prerequisites, hardware, OMERO credentials
- Analyse your first plate — one continuous journey from a plate ID to a figure
- Try it — run it on your own machine
- How it works — the pipeline stages, the data model, and the six packages
Source, issues and pull requests live on GitHub. OMERO-Screen is developed in the Hochegger lab at the University of Sussex.