Search

SummaryDespite extensive research, internal tumor heterogeneity presents enormous challenges to achieve complete therapeutic responses. Changes in protein expression are central determinants of cancer phenotypes that reflect potential therapeutic targets. However, previous proteomic studies did not address internal heterogeneity, therefore, masked the necessary spatial resolution to achieve a comprehensive understanding of cancer complexity. Here we present the first large-scale multi-focal breast cancer proteomic study of 330 tumor regions which associated cancer cell function, pathological parameters, and spatial localization of each tumor region. We found marked internal proteomic heterogeneity even within tumors presenting homogeneous receptor expression. Additionally, analysis of the internal heterogeneity, based on coexisting receptor expression or histological patterns in single tumors, showed significant functional differences between homogeneous and heterogeneous tumors related to cancer metabolism, immunogenicity, and proliferation. We anticipate that this study will serve as a starting point towards the development of improved cancer therapy and diagnostics.

Detecting cancer dependencies is key to disease treatment. Recent efforts have mapped gene dependencies and drug sensitivities in hundreds of cancer cell lines. These data allow us to learn for the first time models of tumor vulnerabilities and apply them to suggest novel drug targets. Here we devise novel deep learning methods for predicting gene dependencies and drug sensitivities from gene expression measurements. By combining dimensionality reduction strategies, we are able to learn accurate models that outperform simpler neural networks or linear models.