Polygenic Risk Scores for Diverse Populations - Bridging Research and Clinical Care

摘要 本补充文件将在目前资助的题为“多基因风险评分”的R 01 HL 151152基金的基础上进行扩展 (PRS)为不同人群-桥接研究和临床护理”通过一个集中的多基因风险评分 高血压（HTN）和相关特征的分析，我们所有人的不同数据的合并， 生活方式的影响，这在最初的R 01申请中没有优先考虑。心血管疾病（CVD）和 其风险因素造成重大社会负担，是发病、死亡和残疾的主要原因。 最近的报告记录了40年来CVD死亡率下降的减缓或逆转，这促使人们开始关注心血管疾病。 CVD预防、诊断和治疗的创新转型。基因组研究， 精准医疗，通过应用PRS，风险变量的聚合， 变成一个单一的分数。然而，迄今为止，参与减贫战略研究的绝大多数人都是 欧洲血统（EA）我们已经证明，EA衍生的PRS不能推广到种族， 种族多样的人群，因为EA衍生的PRS预测CVD和CVD风险因素， EA的准确性高于所有其他种族和民族群体。这种缺乏普遍性的现象反映了 预测的准确性和不可预测的偏见的背景下，人口结构的特点是种族和 种族多样的人口。尽管有这些限制，商业化和向患者提供EA- 衍生的PRS正在进行中，即使EA衍生的PRS应用于不同的人群， 很有可能加剧CVD和CVD危险因素中长期存在的种族和民族差异。研究 这使得在种族和民族多样化的情况下能够估计、验证、校准和情境化PRS。 需要向所有人口提供信息，以确保所有人口都能从精准医疗中获益。我们将利用数据 使用遗传学和流行病学（PAGE）研究的人口和结构，旨在充分整合 我们所有人的研究计划进行独立验证。我们建议（一）建立公正的生产者责任制度， 祖先多样性PAGE研究中的HTN相关性状（收缩压和舒张压以及HTN） 人口PRS将在All of Us研究计划中得到验证;（ii）情境化/绩效 在关键的生活方式因素：制定和评估生活方式风险评分（LRS）组成的非遗传临床 和预后数据，以提高PRS的预测性能。PRS+LRS将在All of Us中进行验证 研究计划;和（iii）临床表征：评价HTN相关PRS的临床意义， 利用All of Us研究计划提高对HTN相关疾病的生物学洞察力， 对具有最大临床影响的基因和途径进行随访。我们将进行全现象关联 （PheWAS）分析，以验证预期的高血压相关性并检测新的表型 潜在的临床影响。

ABSTRACT This supplement will expand on the currently funded grant R01 HL151152 entitled “Polygenic Risk Scores (PRS) for Diverse Populations - Bridging Research and Clinical Care” through a focused polygenic risk scores analysis of hypertension (HTN) and related traits, an incorporation of All of Us diverse data, and incorporation of lifestyle effects, which were not prioritized in the original R01 application. Cardiovascular disease (CVD) and its risk factors impose major societal burdens and are leading causes of morbidity, mortality, and disability. Recent reports documenting a slowing or reversal of four decades of declining CVD mortality rates motivate an innovative transformation of CVD prevention, diagnosis, and treatment. Genomic research, a cornerstone of precision medicine, offers such a transformation through the application of PRS, the aggregation of risk variants into a single score. However, the vast majority of participants included in the PRS research to-date have been of European ancestry (EA). We have demonstrated that EA-derived PRS are not generalizable to racially and ethnically diverse populations, as EA-derived PRS predict CVD and CVD risk factors with much greater accuracy in EA than in all other race and ethnicity groups. This lack of generalizability reflects differential prediction accuracy and unpredictable bias in the context of population structure that characterizes racially and ethnically diverse populations. Despite these limitations, efforts to commercialize and provide to patients EA- derived PRS are underway, even though the application of EA-derived PRS to diverse populations has the strong potential to exacerbate longstanding racial and ethnic disparities in CVD and CVD risk factors. Research that enables estimation, validation, calibration, and contextualization of PRS in racially and ethnically diverse populations is needed to ensure that all populations reap the benefits of precision medicine. We will utilize data from the Population and Architecture using Genetics and Epidemiology (PAGE) study and aim to fully integrate the All of Us Research Program for independent validation. We propose (i) the creation of unbiased PRS for HTN-related traits (systolic and diastolic blood pressure and HTN) in the ancestrally diverse PAGE study population. The PRS will be validated in the All of Us Research Program; (ii) Contextualization/Performance across key lifestyle factors: Develop and evaluate a lifestyle risk score (LRS) composed of non-genetic clinical and prognostic data to improve predictive performance of PRS. The PRS+LRS will be validated in the All of Us Research Program; and (iii) Clinical Characterization: Evaluate the clinical significance of HTN-related PRS to improve biologic insight of HTN-related diseases using the All of Us research program to prioritize functional follow-up on genes and pathways with greatest clinical impact. We will conduct phenome-wide association (PheWAS) analyses to validate expected hypertension-related associations and detect novel phenotypic associations of potential clinical impact.