With the increasing interest in personalized and precision medicine, molecular profiling of a patient’s tumor material is becoming an essential step for prognosis and treatment planning of many cancer types. However, molecular profiling is complicated in many cancer types by the heterogeneous nature of the carcinogenic process and is restricted because of budgetary reasons. Now, the digitization of histopathology glass slides to Whole Slide Images (WSIs) presents unprecedented opportunities for cost-efficient analyses of tumor properties. Namely, WSIs are available without additional cost as they are obtained in routine clinical practice for diagnostic purposes. Even though they intrinsically offer morphological rather than molecular information, recent studies have shown the potential of deep learning techniques to extract hidden morphological features from WSIs that associate with molecular properties. Hence, deep learning models on WSIs present the potential to gain fundamental insights on the presence of omics properties and their spatial (co)location, and as a result to guide prognosis. Specifically, in this project we will investigate the potential of using WSIs as a cost efficient proxy for spatially resolved molecular profiling of tumor properties. Hereto, we will develop an innovative deep learning framework for the analysis of WSIs which will be generic and efficient to fine-tune on different omics prediction tasks in various cancer types.