SUMMARY Human brain development represents perhaps the pinnacle of complex organ specification, and an ideal model system for understanding 1) how normal development can produce all the cell types necessary for human cognition and 2) how genetic variation can perturb this process and lead to disease. We will generate large-scale single cell data sets to develop accurate models capable of predicting the effects of both genetic changes to regulatory elements and perturbations to trans-acting regulatory factors on gene expression during the complex developmental process of human brain development. We will study two highly medically relevant, human, in vitro, temporally dynamic differentiation systems that faithfully recapitulate fetal differentiation patterns: hiPSC- derived cerebral cortical and spinal cord organoids. For each of these differentiation trajectories, we will work in distinct aims toward mapping, perturbing, modeling, validating, and learning: Mapping: we will generate systematic, single cell multi-omic (RNA-seq, ATAC-seq, and protein quantification) data to map regulatory elements, chromatin contacts, RNA polymerase, protein binding, and gene expression through differentiation of hiPSCs to brain tissue. Perturbing: We will use CRISPR-based methods to comprehensively identify TFs required for differentiation and map the single-cell gene regulatory and expression impact of perturbing a subset of these factors at multiple time points across these differentiation trajectories. Modeling: We will develop multi- input nucleotide-resolved neural networks to learn dynamic gene regulatory networks using these mapping and perturbation data sets. These models will aim to understand the changing landscape of regulation and grammars of transcription factor motifs over differentiation time, and will predict both chromatin and gene expression effects expected from genetic perturbations. Validating: We will apply our network models to identify, investigate, and experimentally test perturbations relevant to understanding disease variation, by knocking down transcription factors, perturbing regulatory elements, and editing disease-associated noncoding variants. Learning and comparing: Finally, we will extract and test molecular properties of transcription factor function from validated models, and compare experimental and modeling approaches to better understand accuracy, advantages, and disadvantages. Successful completion of our project will provide mechanistic interpretations for how genetic variants may impact development (by disrupting regulatory element that in turn disrupt gene expression) in brain development. Our Stanford team comprises a diverse team of investigators with a history of productive collaboration, and with expertise in genomics methods development (Greenleaf, Engreitz), single cell methods and analysis (Greenleaf, Pasca), 3D cellular models of human brain (Pasca), and deep learning for genomic data sets (Kundaje). The output of t...