Understanding the contribution of copy number variants (CNVs) to human pathology remains challenging, in part because of the complex and varied influence of loci within them. In some cases, a single dosage-sensitive locus drives pathology. In other examples, multiple genes contribute to phenotypes through additive effects and through complex genetic interactions. In the previous funding period, we focused on deletion syndromes and reciprocal CNVs and asked whether we could develop tools to aid the identification of driver genes. Progress in that work has informed several areas of pathology. For example, the identification of SIN3B as a driver of a non-recurrent 230 kb deletion on 19p13.1 that causes a complex neurodevelopmental phenotype reinforced the role of the Sin3/HDAC transcriptional corepressor complex in neurodevelopment and facilitated the identification of Mendelian mutations that phenocopy the CNV. In the case of the 17q24.2 CNV, we learned about the synthetic CNV phenotype likely caused by dosage imbalance of both BPTF and PSMD12. Finally, modeling of multiple phenotypes associated with the 16p11.2 CNV showed how: (a) key neurodevelopmental and craniofacial pathologies were conserved from humans, to mice, to rats, to zebrafish; (b) demonstrated discrete epistatic interactions between loci; and (c) generated a host of surrogate model organisms in which to perform both pathway analysis and drug discovery. Overall, the genetic and functional dissection of CNV exemplars has underscored three observations: (a) not all genes in a CNV are dosage sensitive (suspected, but now supported by experimental evidence); (b) some genes are dosage sensitive in one direction (gain, loss, but not both); and (c) non-driver genes can participate in distinct genetic interactions. The knowledge and experimental tools gained affords us the opportunity to expand our ambition. We still do not understand how intra-CNV combinatorial dosage influences phenotype, and whether the lessons learned from prior work are broadly generalizable. Moreover, given the non-trivial contribution of CNVs to disease burden, it is crucial that we pay attention to the development of therapeutic avenues. We propose three Aims: (1) We will dissect a unique set of CNVs, in which the phenotype is driven only by duplication to identify triplosensitive genes and ask whether disease susceptibility is driven by a threshold effect as a means to understanding why only a subset of human genes are dosage sensitive. (2) We will harness our existing zebrafish models to characterize the molecular properties of epistatic effects using the neuronal and facial cartilage phenotypes of the 16p11.2 CNV and ask whether we can detect epistasis at the level of transcriptional regulation. (3) We will combine our extensive repertoire of zebrafish models of CNV drivers with the Connectivity Map (CMap) Touchstone dataset and a newly developed safety/efficacy protocol to identify candidate therapeutics. Together...