DESCRIPTION (provided by applicant): Reconstructing the circuits that control how cells detect environmental triggers and adopt specific fates is a fundamental challenge across all areas of biology. Genomic research on circuitry has initially used observational approaches that infer regulation from correlations in molecular profiles, but cannot distinguish correlation from causation. Our Center has developed and successfully demonstrated an approach that uses single perturbations to determine the function of individual components. However, because interactions in circuits are non-linear, we cannot predict how the circuit will function simply by summing up these individual effects. What is needed is a massive combinatorial analysis: perturbing multiple genes simultaneously, with a compatible genomic readout. To take on this apparently intractable problem, we need to radically boost the type and scale of our experimental and analytic methods. Several advances from our groups and others provide an unprecedented framework for such massively parallel, high order combinatorial circuit analysis based on millions of experiments. First, the CRISPR/Cas9 system enables large-scale pooled, multi-locus gene perturbation in mammalian cells. Second, massively-parallel single cell genomics and proteomics, based on combinatorial bead barcoding and gel droplet microfluidics, allow global readouts from hundreds of thousands to millions of cells. Third, the mathematical theories of random matrices and compressive sensing justify substantial reduction in the sampling of an otherwise enormous combinatorial space under biological realistic and testable hypotheses. Here, we will develop a set of Massively Parallel Combinatorial Perturbation (MCPP) assays, as cost-effective methods to measure genomic profiles in individual cells commensurate with the scale required for high order combinatorial pooled perturbation screens (Aim 1). To analyze data generated with these methods, we will develop methods to: generate combinatorial genetic models from an under-sampled high-order combinatorial space, infer molecular mechanisms that explain the genetic models, and tackle the scale and noise of multiple types of single cell measurements (Aim 2). We will perform massive combinatorial perturbations and profiling to derive a genetic model of the transcriptional response to pathogens in dendritic cells, and then develop a dynamic molecular model that integrates the genetic model with high- resolution measurements of diverse molecular changes together with the RNA and protein life cycle (Aim 3). We will apply similar approaches to study cell fate transitions and maintenance in developing embryoid bodies, to build a combinatorial genetic model of how transcription and chromatin factors drive, stabilize or resist cell differentiation inan inherently heterogeneous population (Aim 4). Our studies will develop broadly-applicable methods for large-scale pooled combinatorial genetic pert...