AI/ML for Biomedicine

This group, led by Prof. Marianna Rapsomaniki, focuses on developing AI/ML approaches tailored to complex and multi-modal biomedical data to elucidate cancer mechanisms, from the intra-cellular to the patient level. Topics of special interest include understanding 3D genome regulation, modeling the tumor microenvironment using single-cell omics (transcriptomics, metabolomics, protemics), and developing AI-driven precision oncology approaches based on predicting the effects of drug perturbations ex vivo.

• Ovchinnikova, K., Born, J., Chouvardas, P., Rapsomaniki, M.* and Kruithof de Julio, M.* (2024). "Overcoming limitations in current measures of drug response may enable AI driven precision oncology." npj Precision Oncology, 8, 95.
• Gossi, F., Pati P., Chouvardas, P., Martinelli, A. L., Kruithof-de Julio, M., Rapsomaniki, M. A. (2023). "Matching single cells across modalities with contrastive learning and optimal transport." Brief Bioinform 24(3).
• Martinelli, A. L., Rapsomaniki, M. A. (2022). "ATHENA: analysis of tumor heterogeneity from spatial omics measurements." Bioinformatics 38(11): 3151-3153.
• Rapsomaniki, M. A., Maxouri, S., Nathanailidou, P., Garrastacho, M. R., Giakoumakis, N. N., Taraviras, S., Lygeros, J., Lygerou, Z. (2021). "In silico analysis of DNA re-replication across a complete genome reveals cell-to-cell heterogeneity and genome plasticity." NAR Genom Bioinform 3(1): lqaa112.
• Wagner, J., Rapsomaniki, M. A., Chevrier, S., Anzeneder, T., Langwieder, C. , Dykgers, A., Rees, M., Ramaswamy, A., Muenst, S., Soysal, S. D., Jacobs, A., Windhager, J., Silina, K., van den Broek, M., Dedes, K. J., Rodriguez Martinez, M., Weber, W. P., Bodenmiller, B. (2019). "A Single-Cell Atlas of the Tumor and Immune Ecosystem of Human Breast Cancer." Cell 177(5): 1330-1345 e1318.
• Rapsomaniki, M. A., Lun, X. K.,Woerner, S., Laumanns, M., Bodenmiller, B., Martinez, M. R. (2018). "CellCycleTRACER accounts for cell cycle and volume in mass cytometry data." Nat Commun 9(1): 632.

• AI for precision oncology: We create artificial intelligence (AI) tools to better understand the complexity of cancer. With these tools, we can analyze complex tumor data to see how cancer cells evolve and respond to treatments. Our goal is to discover valuable information that will lead to more effective, personalized treatments for each patient.
• Generative AI for medical research: We develop AI technologies capable of reproducing costly and time-consuming medical experiments. Using existing data, these technologies can predict the results of unperformed experiments or help test new drugs quickly and at lower costs.
• PROMETEX: Metabolically-instructed therapy selection for prostate cancer. We combine AI with patient-derived cancer cell models (organoids) to create an atlas of responses to different drugs. This atlas helps us choose the most effective treatments for each patient, taking individual differences into account. PROMETEX is a collaboration with the Urogenus lab at the University of Bern and the Alexandrov lab at EMBL (European Molecular Biology Laboratory) and is supported by the Swiss National Science Foundation (SNSF).