Ewa Szczurek

Our lab focuses on artificial intelligence, in particular probabilistic graphical models and deep generative models, and their applications in computational medicine. Specific applications of our research include oncology, pulmonology and the AI-driven design of antimicrobial peptides, a work for which we were recently awarded the ERC Consolidator grant. Have a glimpse on our example projects:

• Antimicrobial resistance: variational autoencoders to the rescue! We work out specialized deep generative models to produce synthetic antimicrobial peptides that can kill antibiotic-resistant bacteria. Methods: deep learning, generative models.
• Modeling tumor microenvironment. What is the spatial organization of the tumor and its neighborhood? How do they interact? These interesting questions are solved in our lab by analyzing spatial transcriptomics, digitized tumor imaging, or imaging mass spectrometry data. This work is in collaboration with colleagues from the Oncology Bioinformatics department at Merck, Germany, and with an international consortium called IMMUcan. Methods: probabilistic graphical models, deep learning, machine learning models.
• Drawing the genealogical trees of tumors. Which cancer mutations come first? Which events and how drive tumor evolution: single nucleotide, or copy number variants?  How does drug resistance appear in cancer? These and many more questions about the family history of tumor cells we find very exciting! Methods: probabilistic graphical models.

Deep learningGenerative models  Probabilistic graphical models

Machine learning models