Human genetics reveals underlying biological mechanisms for human disease susceptibility and quantitative trait variability. Random inheritance of parental alleles provides a natural experiment to test genes for a causal role in human phenotypes. Understanding the variants, genes, and mechanisms that underlie human phenotypic variability can have important therapeutic implications – drugs whose targets are supported by human genetic evidence are more likely to be efficacious. Despite improvements in treatments for obesity (particularly newly-approved terzepatide, whose targets both have support from our prior genetic studies), obesity remains a pressing public health need, and the most effective therapies carry significant risks and cost. Our prior and ongoing genome-wide association studies (GWAS) have implicated both known and novel genes for anthropometric traits, including measures of obesity and height (the classical model polygenic trait). However, translating genetic discovery into new biology and actionable mechanistic hypotheses remains challenging. Fortunately, improved tools and new resources now enable more rapid progress on the journey from genetic discovery to biological knowledge. We can benchmark and therefore rationally select powerful computational tools and data sets that, especially in combination, more precisely implicate causal variants, genes, and pathways. Newly available and much larger studies of rare variation, especially coding variation, also help pinpoint relevant genes. Whole genome sequence data sets now allow queries of structural variants whose often large effect sizes can reveal regulatory effects on causal genes. Expanded sample sizes for non- European ancestries will increase power and permit a broader delineation of the phenotypic consequences of causal variants. Functional studies can test hypotheses emerging from genetic and computational analyses, especially if the functional assays are shown to be relevant by benchmarking them against genetic data. This proposal builds on the ongoing, successful and global collaborations we established within the GIANT consortium, which led to discovery of most of the common variants known to be associated with anthropometric traits. We will leverage new collaborations and newly available genetic resources, with larger and more diverse sample sizes. By covering the range of variants - common and rare, coding, noncoding, and structural - we will more completely define the genetic basis of measures of obesity (BMI and WHR) and height. We will use data-driven approaches to benchmark and select computational methods (to turn genetic data into hypotheses about causal variants, genes, and mechanisms) and functional assays (to test these hypotheses). Finally, we will examine, in diverse populations, the phenotypic and metabolic implications of polygenic risk scores composed of associated variants (including meaningfully selected subsets). Our proposal will also address as yet un...