Recent advances in genome sequencing illustrate the complexity, heterogeneity and plasticity of cancer genomes. In leukemia - a group of blood cancers affecting 300,000 new patients every year – we know over 100 driver mutations. This genetic complexity poses a daunting challenge for the development of targeted therapies and highlights the urgent need for evaluating them in combination. One gene class that has recently emerged as highly promising target space are chromatin regulators, which maintain aberrant cell fate programs in leukemia. The dependency on altered chromatin states is thought to provide great therapeutic opportunities, since epigenetic aberrations are reversible and controlled by a machinery that is amenable to drug modulation. However, the precise mechanisms underlying these dependencies and the most effective and safe targets to exploit them therapeutically remain unknown.Here we propose an innovative approach combining genetically engineered leukemia mouse models and advanced in-vivo RNAi technologies to explore chromatin-associated vulnerabilities at an unprecedented level of depth. Following a first screen in MLL-AF9;Nras-driven AML, which led to the discovery of BRD4 as a promising therapeutic target, we aim to (1) construct a knockdown-validated shRNA library targeting 520 chromatin regulators and use it to comparatively probe chromatin-associated dependencies in diverse leukemia subtypes; (2) explore the mechanistic basis of response and resistance to suppression of BRD4 and new chromatin-associated targets; and (3) pioneer a system for multiplexed combinatorial RNAi screening and use it to identify synergies between established and new chromatin-associated targets. We envision that this ERC-funded project will generate a comprehensive functional-genetic dataset that will greatly complement ongoing genome and epigenome profiling studies and ultimately guide the development of targeted therapies for leukemia and, potentially, other cancers.