Cardiovascular diseases are leading global causes of death and disability, presenting as interrelated phenotypes of atherosclerotic vascular disease, heart failure, and arrhythmias. They arise from interactions between environmental factors and common and rare genetic variants, including relatively common Mendelian lipid disorders, cardiomyopathies, and arrhythmias that collectively occur in at least 1/100 individuals. The availability of genetic sequencing is altering clinical management, but a major barrier to the widespread application of this practice is that the function of the vast majority of variants in key cardiovascular disease genes is unknown. Variant effect maps that define function for nearly all missense variants in a target sequence offer a way forward. This project brings together scientists at the forefront of variant effect mapping in diverse cellular systems, illuminating underlying cardiovascular biology, establishing relationships between variant function and human phenotypes, and working with others in multi-institutional collaborations. Our CardioVar team will generate a comprehensive atlas of variant effect maps for key cardiovascular disease genes. In Aim 1, we will develop, optimize, and validate a range of high-throughput cellular assays. We will use a range of generalizable (e.g. surface abundance) and bespoke (e.g. electrophysiological, lipoprotein uptake) assays to directly measure variant function in disease-relevant context. Assays will be assessed by their ability to discriminate pathogenic from benign variants. In Aim 2, we will use in situ targeted mutagenesis or insertion of variant constructs at a safe harbor site to generate pools of cells capturing all single-nucleotide changes in target genes. We will then deploy existing validated assays and those emerging from Aim 1 to generate and validate variant effect maps at scale. Functional scores and uncertainty estimates will be derived and evaluated, both by performance on pathogenic and benign variants and on correlation with discrete and quantitative phenotypes in clinical cohorts. In Aim 3, we will derive biological and clinical insights from variant effect maps. Discordant cases, where variant scores diverge from clinical annotation, will be further investigated in zebrafish, iPSC-cardiomyocytes, and automated patch clamping systems. Through a combination of hypothesis-driven analysis and machine learning models, we will reveal relationships among variant effects, protein structure, protein function, and human phenotypes. To optimize use of the atlas, we will provide a portal serving as a variant-centric decision support system for evaluating functional evidence of pathogenicity. We will release variant effect map data pre- publication via MaveDB (that we co-developed) and share all renewable variant assay reagents. The CardioVar atlas of missense variant effects, covering key cardiovascular disease genes, will be an essential and interpretable communit...