To stay up-to-date with the lab, don't forget to follow us on Twitter (@BodenmillerLab)!

We develop experimental and computational methods to study tumor ecosystems on the single-cell level. In such an ecosystem, many cell types, including tumor, stromal, immune and endothelial cells, interact and communicate in multicellular assemblies. We seek to understand how tumor ecosystems function and ultimately how their properties affect disease. We generate comprehensive single-cell datasets of tumors from many patients, which requires detection of dozens of markers simultaneously. We have pioneered mass cytometry-based methods for simultaneous multiplexed marker detection and analysis in both dissociated tissues and on tissue sections, and continue to develop these methods further (see technology section).

Solid tumors are multicellular ecosystems of diverse cell types that interact to manifest emergent phenotypes which ultimately determine clinical outcome. We combine suspension or imaging mass cytometry with computational techniques to systematically describe cell phenotypes in tumor ecosystems and to examine their distribution across patients and tumor types. We generate atlases of human tumors from large patient cohorts with known clinical outcome, and identify cellular and spatial phenotypes associated with disease progression. These atlases lay the foundation for improved patient stratification and provide potential drug targets for ongoing study. These data are also the foundation for follow-up studies to understand mechanisms of the disease. We are part of several large-scale, multi-center projects bringing together clinicians, research labs and pharmaceutical companies, in which we develop mass cytometry and imaging mass cytometry for precision medicince applications.

Single-cell systems biology analyses of tumor samples yield a wealth of data about cancer biology. To understand the regulatory networks underlying the disease, we use algorithmic and data-driven approaches to model subpopulations of cells and their signaling network structures. Further, using data from imaging mass cytometry, we model how signaling network states spatially couple with those of other cells. As a complement to modeling and associative studies, we use in vitro patient-derived organoid and cell co-culture models to conduct small molecule screens and carry out perturbation studies aimed at understanding mechanistic aspects of tumor biology.

The Bodenmiller lab is supported by these consortia:

In mass cytometry, we use a mass spectrometer to measure protein and/or transcript levels within cells, using antibodies or RNA probes linked to different metal isotopes. In imaging mass cytometry, we have extended this technology to solid tissue samples such as tumor biopsies, analyzing them spatially and capturing effects of the local microenvironment on tumor cells. We apply IMC in 2D and are also developing it in 3D. Mass cytometry can in principle reliably differentiate over a 100 probes.  We are continuously improving the speed, number of markers, resolution, reliability, sensitivity, biological interpretability and overall quality of high-dimensional single cell analysis.

Deriving relevant biological information from high-dimensional datasets is an ongoing challenge in systems biology. We explore the capabilities of existing statistical and image analysis tools to analyse imaging mass cytometry data in a meaningful way. We also develop software to process, visualize and analyze high dimensional IMC datasets. 

The Bodenmiller GitHub page has code and scripts for many projects.

The IMC workflow provides an overview on IMC data analysis approaches.

The Bodenmiller lab protocols page is in progress.

We are always looking for excellent and motivated students/postdocs with a strong background in biological/biochemical/biomedical research or bioinformatics.

Prospective Ph.D. students can apply to our group via the Life Science Zurich Graduate School. Our group is a member of the Molecular Life Sciences, the Systems Biology and the Cancer Biology programs. For computational students, we offer shared Ph.D. positions with bioinformatics research groups.

Post-doctoral candidates should be highly motivated with a passion for science with great interest in quantitative biology, single-cell analysis and systems biomedicine. Our lab and collaborators provide an excellent and vibrant interdisciplinary environment, including systems (cancer) biology, biochemistry, analytical sciences, computer sciences and biomedical research. Candidates should have a demonstrated record of productivity during their Ph.D. studies, including publications in peer-reviewed journals (see DORA).

Applications should be directly sent to Bernd Bodenmiller, including your CV, Publication List and a short paragraph with your scientific interests and what you hope to achieve during your postdoctoral time.

Search PubMed.

We provide a range of master and semester projects for motivated students with interest in biological/biochemical/biomedical research or bioinformatics. Applications should be sent directly to Bernd Bodenmiller, including your CV and a short paragraph with your scientific interests.

Alternatively, for hands-on experience in the Bodenmiller lab, we offer the block courses BME331 & BME330 in the Spring and Autumn semester respectively.

For details of provided lectures, please see course descriptions at UZH and ETHZ.

Master’s Thesis or Other Credit-Earning Assignment in Deep Learning for Oncology

The Bodenmiller Lab at UZH/ETH is seeking talented and motivated Master’s students with expertise in machine learning to join our team. We offer exciting opportunities to work on cutting-edge deep learning projects focused on tumor-tissue image analysis as part of a Master’s thesis or a credit-earning assignment (e.g. unpaid internship, lab rotation, semester project).

About the Projects

Imaging Mass Cytometry (IMC) is an advanced imaging technology that combines mass spectrometry with microscopy to generate highly multiplexed spatial data at single-cell resolution. Unlike traditional imaging methods that rely on fluorescent markers, IMC uses metal-tagged antibodies to detect dozens of proteins simultaneously within a tissue sample. This enables researchers to study complex cellular environments—such as tumors or immune tissues—with exceptional detail. Currently, the standard approach for analyzing IMC images of tumor tissue involves cell segmentation, where individual cells are identified and treated as the smallest units of analysis. However, this method introduces potential biases and overlooks valuable contextual information in the surrounding tissue. To overcome these limitations, we are developing deep learning approaches that analyze IMC data directly at the pixel level, enabling a more comprehensive and unbiased understanding of tissue architecture and cellular interactions. Potential Projects:

• Optimization of Vision Transformer (ViT) architectures to improve IMC image data analysis
• Enhancing explainability of pixel-based IMC models to better understand deep learning predictions
• Validating pixel-based IMC models by comparing them to single-cell analysis techniques

These projects offer a unique opportunity to apply and develop advanced AI models, work with high-dimensional biological data, and contribute to impactful biomedical research.

What We Offer

• A dynamic and interdisciplinary research environment at the intersection of machine learning and biological sciences
• The chance to work closely with leading researchers in computational biology and biomedical data science
• Hands-on experience applying deep learning to biomedical research
• Flexible start dates

Who Should Apply?

We are looking for students who:

• Are proficient in Python and PyTorch
• Are self-reliant and have strong problem-solving skills
• Have an interest in computer vision and cancer biology

Nice to have:

• A background in computer science, engineering, computational biology or bioinformatics
• Software or machine learning engineering experience
• Experience with high-performance computing environments
• Experience with workflow management systems

How to Apply

Interested students should submit the following documents:

• CV (max. 2 pages)
• A short motivation letter (max. 1)
• Transcript of records (we will focus primarily on relevant coursework rather than overall grades)
• Optional: A link to a GitHub repository showcasing a relevant application

Applications should be sent in ONE email addressed to both

• santiago.castrodau@uzh.ch
• victor.ibanez@uzh.ch