Approved Research

Machine learning for risk gene identification and personal disease risk prediction

Lay summary

Artificial intelligence and machine learning have boosted our ability to understand complex data sets. The human genome is currently among the most complex kinds of data. It holds valuable information on hereditary common diseases such as autoimmune and cardiovascular diseases. In this 3 year project, our first aim is to develop novel machine learning methods to understand which parts of the genome are predisposing to disease and by which molecular mechanisms they cause disease. For this purpose we aim to integrate as much prior knowledge into the models as possible. Based on these results our second aim is to predict individuals at risk of disease. Both aims have the potential to improve public health in the long run. Identification of disease genes forms the basis for the design of treatment strategies. For example, identified disease genes can serve as a starting point to develop new drugs.  Personal genetic risk prediction can improve healthcare in particular by providing early diagnosis and prevention strategies for high risk individuals.