Beyond PheWAS: Recognition of Phenotype Patterns for Discovery and Translation

项目概要 基因组医学为改进诊断方法和更有效地治疗患者带来了希望 具体疗法。全基因组关联研究（GWAS）阐明了遗传标记 提高对许多疾病和病症的风险和机制的临床理解， 这可能最终指导针对特定患者的诊断和治疗。前两 这项工作的周期（2011-2014 年和 2014-2018 年）引入了全表型关联研究 （PheWAS）作为识别新疾病变异关联的系统且有效的方法 并利用电子健康记录 (EHR) 发现多效性。该提案将开发新颖的 使用表型风险评分根据表型模式识别关联的方法 (PheRS) 系统性研究孟德尔疾病变异影响的方法 关于常见病。通过这样做，它还创建了一种评估罕见变异致病性的方法， 并将识别患有未确诊孟德尔病的最高风险的患者。该项目是 由大型 DNA 生物库与 EHR 的去识别副本相结合来实现。这个项目有四个 具体目标。首先，我们将开发并验证用于评估变异致病性的 PheRS 通过在其预测算法中利用计费代码、实验室数据和 NLP 功能。的 第二个目标是在庞大的人群中应用 PheRS，以创建一个强大的稀有变异存储库 不同人群中的关联（eMERGE 网络和大型国家队列，这可以 使用基因型数据接近 200 万人）。第三个目标是评估孟德尔疾病 外显率并评估 PheRS 作为识别未确诊孟德尔风险患者的工具 疾病。第四个目标是使这些工具和资源广泛可用，以帮助变体 解释并帮助其他人运行 PheRS。该项目生成的工具将 验证解释罕见变异功能的新方法，改进基础 对孟德尔病的认识，大大提高了我们对孟德尔病的认识的贡献 孟德尔疾病变异常见疾病和特征，并提供了一种潜在的方法 确定新疗法可能带来益处的患者亚群。

Project Summary Genomic medicine offers hope for improved diagnostic methods and for more effective, patient specific therapies. Genome-wide associated studies (GWAS) elucidate genetic markers that improve clinical understanding of risks and mechanisms for many diseases and conditions and that may ultimately guide diagnosis and therapy on a patient-specific basis. The previous two cycles of this effort (2011-2014 and 2014-2018) introduced the phenome-wide association study (PheWAS) as a systematic and efficient approach to identify novel disease-variant associations and discover pleiotropy using electronic health records (EHRs). This proposal will develop novel methods to identify associations based on patterns of phenotypes using a phenotype risk score (PheRS) methodology to systematically search for the influence of Mendelian disease variants on common disease. By doing so, it also creates a way to assess pathogenicity for rare variants, and will identify patients at highest risk of having undiagnosed Mendelian disease. The project is enabled by large DNA biobanks coupled to de-identified copies of EHR. This project has four specific aims. First, we will develop and validate PheRS for assessment of variant pathogenicity by leveraging billing codes, laboratory data, and NLP features in its predictive algorithms. The second aim is to apply PheRS in huge populations to create a robust repository of rare variant associations in diverse populations (eMERGE Network and large national cohorts, which could approach 2 million people with genotype data). The third aim is to assess Mendelian disease penetrance and evaluate PheRS as a tool to identify patients at risk for undiagnosed Mendelian disease. The fourth aim is make these tools and resources broadly available to aid in variant interpretation and facilitate others running PheRS. The tools generated from this project will validate new approaches to interpreting the function of rare variants, improve basic understanding of Mendelian disease, greatly enhance our understanding of the contribution of Mendelian disease variants to common disease and traits, and offers a potential approach to identify subpopulations of patients for whom new therapies may offer benefit.