Department of Orthopedic Surgery, Kyoto University Graduate School of Medicine, Kyoto, Japan.
Center for Genomic Medicine, Kyoto University Graduate School of Medicine, Kyoto, Japan.
Arthritis Res Ther. 2023 Jun 12;25(1):103. doi: 10.1186/s13075-023-03082-y.
Polygenic risk score (PRS) analysis is used to predict disease risk. Although PRS has been shown to have great potential in improving clinical care, PRS accuracy assessment has been mainly focused on European ancestry. This study aimed to develop an accurate genetic risk score for knee osteoarthritis (OA) using a multi-population PRS and leveraging a multi-trait PRS in the Japanese population.
We calculated PRS using PRS-CS-auto, derived from genome-wide association study (GWAS) summary statistics for knee OA in the Japanese population (same ancestry) and multi-population. We further identified risk factor traits for which PRS could predict knee OA and subsequently developed an integrated PRS based on multi-trait analysis of GWAS (MTAG), including genetically correlated risk traits. PRS performance was evaluated in participants of the Nagahama cohort study who underwent radiographic evaluation of the knees (n = 3,279). PRSs were incorporated into knee OA integrated risk models along with clinical risk factors.
A total of 2,852 genotyped individuals were included in the PRS analysis. The PRS based on Japanese knee OA GWAS was not associated with knee OA (p = 0.228). In contrast, PRS based on multi-population knee OA GWAS showed a significant association with knee OA (p = 6.7 × 10, odds ratio (OR) per standard deviation = 1.19), whereas PRS based on MTAG of multi-population knee OA, along with risk factor traits such as body mass index GWAS, displayed an even stronger association with knee OA (p = 5.4 × 10, OR = 1.24). Incorporating this PRS into traditional risk factors improved the predictive ability of knee OA (area under the curve, 74.4% to 74.7%; p = 0.029).
This study showed that multi-trait PRS based on MTAG, combined with traditional risk factors, and using large sample size multi-population GWAS, significantly improved predictive accuracy for knee OA in the Japanese population, even when the sample size of GWAS of the same ancestry was small. To the best of our knowledge, this is the first study to show a statistically significant association between the PRS and knee OA in a non-European population.
No. C278.
多基因风险评分（PRS）分析用于预测疾病风险。尽管 PRS 已被证明在改善临床护理方面具有巨大潜力，但 PRS 准确性评估主要集中在欧洲血统人群。本研究旨在使用多人群 PRS 为日本人群开发一种准确的膝关节骨关节炎（OA）遗传风险评分，并利用多特征 PRS。
我们使用 PRS-CS-auto 计算 PRS，该方法源自日本人群（同一种族）的膝关节 OA 全基因组关联研究（GWAS）汇总统计数据和多人群。我们进一步确定了 PRS 可以预测膝关节 OA 的风险因素特征，随后基于 GWAS 的多特征分析（MTAG）开发了一个综合 PRS，其中包括遗传相关的风险特征。在接受膝关节放射学评估的长滨队列研究参与者（n=3279）中评估 PRS 性能。将 PRS 与临床风险因素一起纳入膝关节 OA 综合风险模型。
共有 2852 名个体纳入 PRS 分析。基于日本膝关节 OA GWAS 的 PRS 与膝关节 OA 无关（p=0.228）。相比之下，基于多人群膝关节 OA GWAS 的 PRS 与膝关节 OA 呈显著相关（p=6.7×10，每标准偏差的优势比（OR）=1.19），而基于多人群膝关节 OA 的 MTAG 的 PRS，以及体重指数 GWAS 等风险因素特征，与膝关节 OA 的相关性更强（p=5.4×10，OR=1.24）。将该 PRS 纳入传统风险因素可提高膝关节 OA 的预测能力（曲线下面积，74.4%至 74.7%；p=0.029）。
本研究表明，基于 MTAG 的多特征 PRS，结合传统风险因素，并使用大样本量的多人群 GWAS，显著提高了日本人群膝关节 OA 的预测准确性，即使同一种族的 GWAS 样本量较小也是如此。据我们所知，这是第一项在非欧洲人群中显示 PRS 与膝关节 OA 之间存在统计学显著关联的研究。
No. C278。
Zotero 插件
