Abstract：
BACKGROUND: Polygenic risk scores (PRS) have been widely applied in research studies, showing how population groups can be stratified into risk categories for many common conditions. As healthcare systems consider applying PRS to keep their populations healthy, little work has been carried out demonstrating their implementation at an individual level.
METHODS: We performed a systematic curation of PRS sources from established data repositories, selecting 15 phenotypes, comprising an excess of 37 million SNPs related to cancer, cardiovascular, metabolic and autoimmune diseases. We tested selected phenotypes using whole genome sequencing data for a family of four related individuals. Individual risk scores were given percentile values based upon reference distributions among 1000 Genomes Iberians, Europeans, or all samples. Over 96 billion allele effects were calculated in order to obtain the PRS for each of the individuals analysed here.
CONCLUSIONS: Our results highlight the need for further standardisation in the way PRS are developed and shared, the importance of individual risk assessment rather than the assumption of inherited averages, and the challenges currently posed when translating PRS into risk metrics.
METHODS: We performed a systematic curation of PRS sources from established data repositories, selecting 15 phenotypes, comprising an excess of 37 million SNPs related to cancer, cardiovascular, metabolic and autoimmune diseases. We tested selected phenotypes using whole genome sequencing data for a family of four related individuals. Individual risk scores were given percentile values based upon reference distributions among 1000 Genomes Iberians, Europeans, or all samples. Over 96 billion allele effects were calculated in order to obtain the PRS for each of the individuals analysed here.
CONCLUSIONS: Our results highlight the need for further standardisation in the way PRS are developed and shared, the importance of individual risk assessment rather than the assumption of inherited averages, and the challenges currently posed when translating PRS into risk metrics.
摘要：
背景：多基因风险评分（PRS）已广泛应用于研究中，显示了如何将人口群体分层为许多常见条件的风险类别。随着医疗保健系统考虑应用PRS来保持人口健康，在个人层面上展示其实施的工作很少。
方法：我们对已建立的数据存储库的PRS源进行了系统的管理，选择15个表型，包括超过3700万个与癌症相关的SNP，心血管,代谢和自身免疫性疾病。我们使用全基因组测序数据测试了四个相关个体的家族的选定表型。基于1000个伊比利亚人基因组中的参考分布，给出了个体风险评分的百分位值，欧洲人,或所有样品。计算了超过960亿个等位基因效应,以获得此处分析的每个个体的PRS。
结论：我们的研究结果强调了在开发和共享PRS的方式上进一步标准化的必要性。个人风险评估的重要性，而不是继承平均值的假设，以及将PRS转化为风险指标时当前面临的挑战。
方法：我们对已建立的数据存储库的PRS源进行了系统的管理，选择15个表型，包括超过3700万个与癌症相关的SNP，心血管,代谢和自身免疫性疾病。我们使用全基因组测序数据测试了四个相关个体的家族的选定表型。基于1000个伊比利亚人基因组中的参考分布，给出了个体风险评分的百分位值，欧洲人,或所有样品。计算了超过960亿个等位基因效应,以获得此处分析的每个个体的PRS。
结论：我们的研究结果强调了在开发和共享PRS的方式上进一步标准化的必要性。个人风险评估的重要性，而不是继承平均值的假设，以及将PRS转化为风险指标时当前面临的挑战。
