I’m a Zurich-based computational biologist working at the intersection of machine learning and multi-omics.

My research interests lie in developing machine-learning methods for bulk, single-cell, spatial transcriptomics and imaging data. I am especially interested in generative and multimodal deep learning to uncover disease-associated cell states. I aim to build robust, interpretable models that can inform both biology and clinical decisions.

After training in computer science and bioinformatics at MIMUW, I did my Ph.D. with Prof. Altuna Akalin at MDC-BIMSB, combining machine learning and omics to study how epigenetic signals are measured and change in disease. As first author, I showed that HOT regions, sites that appear bound by dozens of transcription factors, are likely ChIP-seq artifacts linked to R-loops, and used sequence-based classifiers to identify GC/CpG-rich features they share across species. I also contributed to DNA-methylation biomarker discovery in cardiovascular disease, co-first-authored a review of statistical methods for bisulfite sequencing, and co-developed the Bioconductor package genomation and the PiGx BS-seq pipeline. I have also worked with experimental groups at ETH Zurich (incl. Laboratory of Nutrition and Metabolic Epigenetics) and the Functional Genomics Center Zurich.

In my industry projects, I applied deep learning to biomedical and sensor data. My work includes a LSTM-VAE (PyTorch) for water-safety monitoring, a VAE (scVI) with transfer learning (scArches) to classify endometriosis from menstrual-blood single-cell data, therapeutic-target prioritization in oncology (survival analysis, classification, AE), and multi-omics and Enformer-based enhancer analysis in Alzheimer’s disease. I’ve collaborated with research, clinical, and data science teams. Explore more of my deep-learning projects on the Independent Projects page. I’ve collaborated with teams at:

Feel free to reach out at katwre (at) gmail.com.