We work at the interface between the hard data sciences (computational statistics and machine learning) and biomedicine. Our long-term goal is to deeply integrate the hard data sciences with systems approaches to disease biology and medicine. To this end we have over several years studied key statistical issues that arise in complex, high-dimensional biomedical data. At the DZNE our focus is on prediction, stratification and systems analyses for neurodegenerative diseases. Working with colleagues across the DZNE (in fundamental, population and clinical research) and internationally, we are developing and applying approaches at the statistical frontier to help realize the promise of next-generation biomedicine.

Advances in high-throughput molecular assays and deep phenotyping coupled with systems-level analyses have the potential to transform biomedical research. Such approaches can inform stratification into disease subtypes, allow prediction of disease state and help elucidate relevant biology at a systems level. Our efforts are directed towards developing high-dimensional statistical and machine learning methods to realize this potential. This involves working on novel methods motivated by, and applied to, specific applications but also working with colleagues across research areas to clarify conceptual issues that arise in moving towards truly scalable and data-intensive approaches. In the area of systems biology, we are currently working on principled yet highly scalable approaches by which to build and test global molecular networks that are specific to biological or disease context. Furthermore, we are working on methods for prediction and stratification of neurodegenerative diseases, with an emphasis on integrative analyses using diverse high-dimensional data types.