Biological membranes define cellular compartments and orchestrate signaling cascades, all while undergoing constant remodeling. We can resolve their structure from imaging data, but the analysis path is fragmented. One tool for segmentation, another for meshing, something else for protein localization, yet another for setting up physical simulations. Export, convert, hope nothing breaks.
Mosaic fixes that. One graphical interface for the entire workflow.
Ask questions and get answers immediately. What is the membrane curvature where proteins cluster? How does protein density vary across the surface? How would this system behave under different conditions? Explore interactively and transition seamlessly from observation to quantitative hypothesis testing to data-driven physical simulations of real geometries.
Import tomograms, segment membranes, build surface meshes, localize proteins, measure geometry, export simulation-ready systems. Adjust parameters and watch results update in real time.
Complete influenza A virus analysis: from tomogram to simulation-ready model
See it work: The IAV tutorial walks through a real analysis start to finish. About 30 minutes hands-on.
Need to prep for MD? Export Martini-compatible coarse-grained systems with positioned proteins. Run HMFF to simulate biomembranes using both your experimental density and physics. Send the result straight to GROMACS.
Mosaic interface
Once your workflow is dialed in, use the pipeline builder to scale it to large datasets:
mosaic-pipeline config.jsonWorking on a remote server or composing custom workflows? The Mosaic shell exposes most GUI operations as commands you can run interactively or replay from a script:
mosaic-shell workflow.shMosaic requires Python 3.11 or higher. Install with pip:
pip install mosaic-gui
mosaic &The first launch takes a bit longer than subsequent starts. A short interactive tour then walks through the basics, about five minutes. Re-run any time with mosaic --onboard basics.
See the installation guide for full instructions.
If Mosaic contributes to your research, please cite:
@article{Maurer2025,
author = {Maurer, Valentin J. and Siggel, Marc and Jensen, Rasmus K. and
Mahamid, Julia and Kosinski, Jan and Pezeshkian, Weria},
title = {Helfrich Monte Carlo Flexible Fitting: physics-based,
data-driven cell-scale simulations},
journal = {bioRxiv},
year = {2025},
doi = {10.1101/2025.05.24.655915}
}Mosaic and HMFF were developed jointly to bridge structural data and physical simulations. HMFF uniquely integrates experimental volumetric data into membrane simulations, enabling data and physics to jointly determine membrane conformation. HMFF is implemented in FreeDTS and integrated into Mosaic.
Mosaic is developed by the Kosinski Lab at the European Molecular Biology Laboratory (EMBL Hamburg).
License: GPL-2.0 (see LICENSE for details)
- Documentation: https://kosinskilab.github.io/mosaic/
- PyPI Package: https://pypi.org/project/mosaic-gui/
- Source Code: https://github.com/KosinskiLab/mosaic
- Issues & Support: https://github.com/KosinskiLab/mosaic/issues