We are working within the intersection of data sciences/data analytics and life sciences. Big data plays an increasing role in modern biomedical and biopharmaceutical research. Examples include -omics data (genomics, transcriptomics, methylomics, proteomics, ...) reflecting biological phenomena at the intracellular level, data from clinical and electronic medical records (EMR), bioimages, data from mobile devices and video recordings. A key questions is, how to utilize these data in order to a) better understand disease mechanisms and b) to optimally stratify patients according to their expected treatment success, disease prognosis and early disease onset. The latter approach is also called personalized medicine. In order to address both questions suitable data mining and machine learning algorithms are crucial and need to be developed and applied. In that context one has to take the specific aspects of the employed data (e.g. high noise level, high dimensionality, large heterogeneity) into account. A critical factor for success and acceptance by medical and life scientists is typically the appropriate consideration and integration of background domain knowledge.
The focus of our research lies on development and application of algorithms for the questions outlined above. Historically, a major focus has been on -omics data, but also other data types (e.g. bioimages, health claims records) occasionally play a role and are of interest.
The current research specifically concentrates on:
- statistical learning methods in personalized / precision medicine (e.g. biomarker signatures)
- computational systems biology: structure learning and modeling of molecular interaction networks
- data integration: learning models from multiple data sources
- knowledge integration: integration of biological background knowledge (e.g. biological networks) and disease mechanisms into predictive models for personalized medicine