ABSTRACT The overall objective of Project 01 is to improve our ability to identify genetic factors relevant to OSA by studying upper airway anatomy and OSA extreme phenotypes. There are known anatomic risk factors (reduced craniofacial skeleton and enlarged soft tissue structures, including tongue fat) for OSA that have been shown to be heritable. We postulate that certain OSA genetic risk variants operate via changes in anatomy, leading to heterogeneous and extreme OSA. A foundational aspect of this hypothesis is that studying genetic associations with intermediate anatomical phenotypes known to cause OSA will facilitate identification of genetic factors in the presence of heterogeneous etiologies. To address the overarching hypothesis, we have three Specific Aims. In Aim 1 we apply automatic, large-scale, high-throughput and advanced machine-learning techniques to clinically-obtained MR (magnetic resonance) and CT (computed tomography) images of the head and neck to quantify upper airway anatomical risk factors for OSA in patients with linked biobank data. Using these phenotypes, we will then perform genome-wide association studies (GWAS) to identify variants related to anatomy which, in turn, are expected to influence risk for OSA. These data will be used to generate enhanced polygenic risk scores for OSA that incorporate genetic predictors of anatomic risk factors (e.g., tongue fat, mandibular length). In Aim 2 we will perform in silico analyses of genetic loci to identify core biological mechanisms and prioritize likely causal variants and genes underlying association signals. These analyses will provide insights into the significance of GWAS loci and be directly complemented by downstream analyses in cell-based and model systems being performed in Project 04. Finally, in Aim 3 we will use an extreme phenotype design to identify novel anatomical associations and rare genetic variants in genes prioritized in Aims 1-2, which will be utilized to further enhance polygenic risk scores and understanding of disease mechanisms. This proposal has a very strong investigative team, with expertise in both OSA anatomy and genetic analysis, and uses innovative strategies including deep anatomic phenotyping, novel machine learning algorithms, and cutting-edge analysis approaches for identifying and interrogating OSA-susceptibility loci. Findings from this proposal will result in a greater understanding of the impact of genetics and upper airway anatomy on OSA heterogeneity, the downstream clinical impact of these genomic alterations, and their biological underpinnings. This deep dive into the genetic underpinnings of quantitative anatomic intermediate traits for OSA will significantly move the field forward by connecting genetic variation to biological mechanisms, enhance the development of polygenic risk scores that have wide-ranging applications from early detection and treatment to screening and case identification of OSA in the electronic medical reco...