Tanevski Lab

Institute for Computational Biomedicine
Heidelberg University Hospital
Im Neuenheimer Feld 130.3, 69120 Heidelberg, Germany


Our group focuses on problem driven development of approaches to data exploration, hypothesis generation and computational scientific discovery to facilitate translational biomedicine.

Our approaches are based on representation learning and supervised analysis of highly multiplexed spatial omics data. We develop new and extend existing explainable, scalable and readily deployable methods for multi-view learning, graph neural networks, metaheuristic optimization and optimal transport to:

  • Identify clinically relevant regions and interactions by explanatory modeling and optimization of global and local tissue/condition specific persistent multicellular patterns.
  • Learn higher order structural and functional organization to form taxonomical models of tissues for comparative analyses and generation of in-silico samples.
  • Integrate multiomics data with databases of prior knowledge to discover context specific mechanistic insighs spanning multiple omics layers.

Our interest is to address questions of structure-function relationships in disease, progression and response to treatment.

We value collaborations with clinical, experimental biology groups and groups working on the development of novel methods for the acquisition of spatially resolved data. We welcome synergistic collaborations with computational groups towards the construction of more robust theoretical and computational frameworks for the analysis of all aspects of biomedical data and beyond.


The Multiview Intercellular SpaTial modeling framework (MISTy) is an explainable machine learning framework for knowledge extraction and analysis of single-cell, highly multiplexed, spatially resolved data. MISTy facilitates an in-depth understanding of intra- and intercellular relationships. MISTy is a flexible framework able to process a custom number of views describing a different spatial or functional context such as intracellular or broader tissue structure, cell-type composition, functional footprints or anatomical regions.

Kasumi is a method for the identification of spatially localized neighborhoods of intra- and intercellular relationships, persistent across samples and conditions. Kasumi learns compressed explainable representations of spatial omics samples while preserving relevant biological signals that are readily deployable for data exploration and hypothesis generation, facilitating translational tasks.

DOT is a method for transferring cell features from a reference single-cell RNA-seq data to spots/cells in spatial omics. It operates by optimizing a combination of multiple objectives using a Frank-Wolfe algorithm to produce a high quality transfer. Apart from transferring cell types/states to spatial omics, DOT can be used for transferring other relevant categorical or continuous features from one set of omics to another, such as estimating the expression of missinng genes or transferring transcription factor/pathway activities.

Tanevski, J., Vuillard, L., Hartmann, F., Saez-Rodriguez, J. Learning tissue representation by identification of persistent local patterns in spatial omics data. bioRxiv:2024.03.06.583691 (2024).
Rahimi, A., Vale-Silva, L.A., Faelth Savitski, M., Tanevski, J., Saez-Rodriguez, J. DOT: a flexible multi-objective optimization framework for transferring features across single-cell and spatial omics. Nature Communications (2024).
Dimitrov, D., Schäfer, P.S.L., Farr, E., et al. LIANA+: an all-in-one cell-cell communication framework. Nature Cell Biology (2024).
Laury, A. R., Zheng, S., Aho, N. et al. Opening the black box: spatial transcriptomics and the relevance of AI-detected prognostic regions in high grade serous carcinoma. Modern Pathology 100508 (2024).
Paton, V., Gabor, A., Ramirez Flores, R.O. et al. Assessing the impact of transcriptomics data analysis pipelines on downstream functional enrichment results. Nucleic Acids Research (2024).
Heumos, L., Schaar, A.C., Lance, C. et al. Best practices for single-cell analysis across modalities. Nature Reviews Genetics 24, 550–572 (2023).
Tanevski, J., Ramirez Flores, R.O., Gabor, A. et al. Explainable multiview framework for dissecting spatial relationships from highly multiplexed data. Genome Biology 23, 97 (2022).
Kuppe, C., Ramirez Flores, R.O., Li, Z. et al. Spatial multi-omic map of human myocardial infarction. Nature 608, 766–777 (2022).
Gabor, A., Tognetti, M., Driessen, A., Tanevski, J. et al. Cell‐to‐cell and type‐to‐type heterogeneity of signaling networks: insights from the crowd. Molecular Systems Biology, 17(10), e10402 (2021).
Schwabenland, M., Salié, H., Tanevski, J. et al. Deep spatial profiling of human COVID-19 brains reveals neuroinflammation with distinct microanatomical microglia-T-cell interactions. Immunity 54(70), 1594-1610.e11 (2021)
Holland, C.H., Tanevski, J., Perales-Patón, J. et al. Robustness and applicability of transcription factor and pathway analysis tools on single-cell RNA-seq data. Genome Biology 21, 36 (2020).
Tanevski, J., Nguyen, T., Truong, B. et al. Gene selection for optimal prediction of cell position in tissues from single-cell transcriptomics data. Life Science Alliance 3 (11), e202000867 (2020).
Tanevski, J., Todorovski, L., Džeroski, S. Combinatorial search for selecting the structure of models of dynamical systems with equation discovery. Engineering Applications of Artificial Intelligence 89, 103423 (2020).
Tanevski, J., Todorovski, L., Džeroski, S. Process-based design of dynamical biological systems. Scientific Reports 6, 34107 (2016).
Tanevski, J., Todorovski, L., Džeroski, S. Learning stochastic process-based models of dynamical systems from knowledge and data. BMC Systems Biology 10, 30 (2016).

We are looking for a highly motivated and talented PhD student to join the newly created Translational Spatial Profiling Center at the Heidelberg University Hospital, managed by the Institute for Computational Biomedicine and the Institute for Pathology.

In a dynamic and multidisciplinary environment, the candidate will work on the development of novel methods for analysis of state-of-the-art spatially resolved data. The candidate will explore the landscape of explainable machine-learning and optimization approaches and create novel approaches that can be applied in a clinical setting. The work will be motivated by problems arising from a range of translational applications, initially focusing on cancer research, and will be directly supported by data generated within the center.

The candidate should:
For all PhD and Postdoc postions we offer:

To apply please submit a letter of motivation tailored to the position (1 page), CV and a list of references with contact details to contact<at>tanevskilab.org.