We focus on chromosomal instability and somatic copy-number alterations and aim to understand how somatic genetic heterogeneity arises and how it affects cellular and patient phenotypes in space and time. To this end, we develop tailored machine learning (ML) algorithms and statistical methods and work closely with clinical partners on the analysis on large-scale patient cohorts.
The methods we develop help us to overcome three major challenges in cancer genomics: (i) knowledge transfer from dedicated high-resolution experiments on cell lines and model organisms to clinical patient datasets, (ii) integration of heterogenous multi-omics data through consistent, interpretable and robust machine learning models and (iii) dealing with the intrinsic cell-type heterogeneity and noise inherent to clinical patient samples.
We develop and apply these methods on large clinical cohorts and actively pursue interdisciplinary and collaborative endeavours with clinical research groups. We are members of the ), the consortium, the consortium and the Human Cell Atlas Lung Biological Network, where our methods have found wide acceptance. We further hold close ties to the , the , the and to local clinical research groups at the and .