# Tanevski Lab

> Tanevski Lab is a computational biomedical discovery group at Heidelberg University and Heidelberg University Hospital. The lab develops interpretable, scalable machine learning and optimization approaches for single-cell and spatial omics data to turn tissue complexity into mechanistic insight and clinical prediction.

## Research

The lab works at the intersection of AI and biomedicine to understand the principles governing tissue organization and plasticity. Core research directions:

- Formalizing spatial tissue organization by learning representations that integrate single-cell, spatial multiomics, and imaging data into a unified view of tissue architecture
- Modeling tissue plasticity as a structured dynamical process to reconstruct how spatial programs evolve across disease stages and under therapeutic pressure
- Linking spatial organization to clinical outcomes via explainable frameworks that establish the molecular architecture of tissue as predictive of disease progression and therapeutic response

Affiliated with the Institute for Computational Biomedicine, the Translational Spatial Profiling Center, and the Interdisciplinary Center for Scientific Computing at Heidelberg University.

## Tools

- [MISTy](https://github.com/jtanevski/mistyR): Multiview Intercellular SpaTial modeling framework. Explainable ML for learning intra- and intercellular spatial relationships from highly multiplexed spatially resolved data. R package. Publication: https://doi.org/10.1186/s13059-022-02663-5
- [Kasumi](https://github.com/jtanevski/kasumi): Identification of spatially localized neighborhoods of intra- and intercellular relationships persistent across samples. Learns compressed, explainable representations of spatial omics samples. R package. Publication: https://doi.org/10.1038/s41467-025-59448-0
- [DOT](https://saezlab.github.io/DOT/): Cell feature transfer from scRNA-seq to spatial omics via multi-objective Frank-Wolfe optimization. Available in R and Python. Publication: https://doi.org/10.1038/s41467-024-48868-z
- [TOAST](https://github.com/cecca46/TOAST): Topography-aware Optimal Alignment of Spatially-resolved Tissues using Fused Gromov-Wasserstein. Python. Publication: https://doi.org/10.1101/2025.04.15.648894
- [ContextFlow](https://github.com/santanurathod/ContextFlow): Context-aware flow matching for inferring structural tissue dynamics from longitudinal spatially resolved omics data, integrating ligand-receptor cell-cell communication patterns. Python. Publication: https://doi.org/10.48550/arXiv.2510.02952
- [SpaCEy](https://github.com/saezlab/SpaCEy): Explainable graph neural network uncovering organizational tissue patterns predictive of clinical outcomes by modeling tissues as spatial graphs of cells and their interactions. Python. Publication: https://doi.org/10.64898/2025.12.12.693857

## Key Publications

- [Explainable multiview framework for dissecting spatial relationships from highly multiplexed data](https://doi.org/10.1186/s13059-022-02663-5): Introduces MISTy. Genome Biology 23, 97 (2022).
- [Learning tissue representation by identification of persistent local patterns in spatial omics data](https://doi.org/10.1038/s41467-025-59448-0): Introduces Kasumi. Nature Communications 16, 4071 (2025).
- [DOT: a flexible multi-objective optimization framework for transferring features across single-cell and spatial omics](https://doi.org/10.1038/s41467-024-48868-z): Introduces DOT. Nature Communications 15, 4994 (2024).
- [Topography-aware optimal transport for alignment of spatial omics data](https://doi.org/10.1016/j.crmeth.2026.101373): Introduces TOAST. Cell Reports Methods (2026).
- [ContextFlow: Context-Aware Flow Matching For Trajectory Inference From Spatial Omics Data](https://doi.org/10.48550/arXiv.2510.02952): Introduces ContextFlow. arXiv:2510.02952 (2025).
- [SpaCEy: Discovery of Functional Spatial Tissue Patterns by Association with Clinical Features Using Explainable Graph Neural Networks](https://doi.org/10.64898/2025.12.12.693857): Introduces SpaCEy. bioRxiv (2025).
- [Pathology-oriented multiplexing enables integrative disease mapping](https://doi.org/10.1038/s41586-025-09225-2): Nature 644, 516–526 (2025).
- [LIANA+ provides an all-in-one framework for cell–cell communication inference](https://doi.org/10.1038/s41556-024-01469-w): Nature Cell Biology 26, 1613–1622 (2024).

## People

- Jovan Tanevski (Group Leader) — computational biomedical discovery, spatial omics ML, explainable AI, cell-cell communication
- Chang Lu (Postdoc)
- Yunfan Bai (Postdoc)
- Sebastian Gonzalez Tirado (PhD student)
- Aroj Hada (PhD student)
- Leoni Zimmermann (PhD student)
- Alessandro Greco (Scientific programmer)
- Philipp Schäfer (Associated PhD student)
- Robin Fallegger (Associated PhD student)
- Chiara Schiller (Associated PhD student)
- Francesco Ceccarelli (Associated PhD student)

## Contact

- Email: contact@tanevskilab.org
- Address: Institute for Computational Biomedicine, Im Neuenheimer Feld 130.3, 69120 Heidelberg, Germany
- GitHub: https://github.com/tanevskilab
- Google Scholar: https://scholar.google.com/citations?user=F47ynYgAAAAJ
- Bluesky: https://bsky.app/profile/tanevski.bsky.social
- X/Twitter: https://x.com/tanevski