As biomedical data are made available in large quantities and variety, we face a unique opportunity and obligation: to use the data to accelerate the discovery of tailored treatments for individuals or small groups thereof. The ability to integrate and analyze heterogeneous biomedical data has so far proved to be a challenging undertaking. We are therefore in search for methods and tools that will address this urgent need. One effort in this direction is the research project iASiS, intermediate results of which I will present in this talk. iASiS proposes two levels of data analysis: first at the stage of homogenous datasets, such as medical images or clinical notes, and then at a level that combines insights from heterogeneous sources. In order to facilitate the latter type of analysis, the results of the former type are brought together in a large knowledge graph. In the talk, I will present the structure of the iASiS knowledge graph, the methods that are used for the different types of analysis, and examples of potentially interesting findings.