PROJECT SUMMARY/ABSTRACT A major impediment to personalized medicine is the slow pace at which new genetic variants are designated as pathogenic or benign. Genome sequencing projects are rapidly uncovering new genetic variation, but a lack of functional annotation for newly discovered variants means that most are classified as Variants of Uncertain Significance (VUS), even in well-studied protein coding genes. Because of this problem, the NHGRI 2020 Strategic Vision emphasizes the need for new highly parallel technologies that assess the functional impact of genetic variants. One promising approach is Deep Mutational Scanning (DMS), in which large numbers of genetic variants are synthesized and assayed in parallel. The large scale of DMS has the potential to keep pace with efforts in variant discovery. This proposal addresses two problems that limit the reliable use of DMS. First, most published DMS are inappropriate for clinical use because they suffer from low biological reproducibility. Our analysis suggests that this problem is caused by the inability to identify and exclude outlier measurements caused by genetic anomalies that occur during the transfection and integration of variant libraries into cells. We will test whether a barcoding strategy designed specifically to detect these types of outliers improves the reproducibility of DMS. The barcoding strategy will be coupled with an improved landing pad system for genomic integration of DMS libraries. Another issue with DMS is the choice of cell type for functional assays. For most diseases, the relevant in vivo cell types are inaccessible. Investigators must compromise between difficult primary cell models and more tractable cell culture systems that less faithfully reflect in vivo biology. We will address this problem first focusing on DMS of transcription factors (TFs), an important class of disease genes. Because cell-type specific TFs work in concert with other TFs, we will test whether expressing groups of cell type-specific TFs in a tractable cell line allows us to perform a DMS that faithfully represents the in vivo cell type. We propose to test this premise using the Cone-Rod Homeobox (CRX) as a model TF. CRX is specific to photoreceptors in the retina, but several studies show that its activity can be recapitulated in HEK293 cells engineered to co-express groups of photoreceptor-specific TFs. To test the validity of this system for a DMS we will directly compare results from engineered HEK293 lines to those from live intact retinas. We hope to determine whether tractable cell lines engineered to express collections of cell-type specific TFs could be reasonable systems for DMS of TFs. Our overall goal is to improve the functionality and reliability of DMS by improving its methodologies and expanding the range of appropriate cell lines.