Understanding the complexity of cancer requires making sense of equally complex biological data. In our lab, we develop probabilistic models for the integration of multi-omics datasets to uncover hidden structure in the data by capturing both shared and modality-specific sources of variation.