Breadcrumb Navigation

Research Group Gunnar Rätsch

Keywords

Cancer Genomics, Medical Informatics, Machine Learning

Summary & Mission statement

The Biomedical Informatics group’s research at ETH lies at the interface between methods and translational research in cancer genomics, medical informatics, and machine learning. We develop new state-of-the-art computational methods for medical applications to bring them into clinical practice.

Overview

With the increasing digitalization of medicine, including Electronic Health Records, patient data is very comprehensive, allowing access to rich clinical data sets. Powerful computational tools are already transforming medicine on many fronts. A solid methodological basis in machine learning and statistical data analysis is necessary to develop AI techniques suitable for solving clinical problems. A hallmark of my group is tackling highly relevant biomedical problems that require non-trivial technical solutions.

My group has specialized in developing advanced algorithms for

  • large-scale learning and data integration,
  • transcriptome analysis based on deep genome-, RNA- and ribosome footprinting data,
  • new analysis approaches of electronic health records using text mining and probabilistic machine learning methods, and
  • predictive models of clinical state utilizing molecular and clinical data.

The BMI Lab is part of the Tumor Profiler (TuPro) study, involving patients from the University Hospitals in Zurich and Basel. The TuPro consortium performs a detailed investigation into the molecular and functional properties of tumors. It aims to help physicians determine better which treatment will best match every patient’s cancer and thus be most effective. Additionally, I serve as co-leader of the RNA-Seq technical working group of the ARGO project of the International Cancer Genome Consortium.

Publications