利用主成分分析多性狀全基因組關聯分析尋找慢性腎臟病的新風險位點

Abstract

慢性腎臟病是全球各國的重要醫療議題，因為其影響醫療資源甚鉅。已經有許多流行病學研究尋找出慢性腎臟病傳統的危險因子，在許多這類的研究，慢性腎臟病通常被操作定義為腎絲球過濾率估值(estimated glomerular filtration rate; eGFR) < 60ml/min/1.73m2。然而許多慢性腎臟病第三至第五期的人，仍然可以保有穩定的腎功能，所以事實上腎功能的變化軌跡比慢性腎臟病的現下分期還來的重要，而其中白蛋白尿(albuminuria)的程度又是影響腎功能變化的主因。近年National Kidney Foundation也已經建議針對臨床試驗案，可以在相對較短期的追蹤年限使用白蛋白尿和eGFR下降速率(eGFR slope)來替代傳統長期追蹤到腎功能大幅下降30%或是進入腎衰竭等等定義。藉由執行白蛋白尿、eGFR以及eGFR slope個別的全基因組關聯分析(genome-wide association study; GWAS)可以尋找腎臟功能惡化進而發展慢性腎臟病的遺傳風險因子。近年在GWAS的領域，亦提倡藉由多性狀研究的方式來增加GWAS的檢力。比較廣為人知的是Multi-Trait Analysis of GWAS (MTAG)，其方式為對數個性狀的 GWAS summary應用廣義逆方差加權法統合分析 (generalized inverse-variance-weighted meta-analysis)。從另一個面向來看，由於以上三種性狀因為都是連續變項， 我們也可以用主成分分析將這些數量性狀先轉化成不同的主成分(principal component; PC)，再對這些PC進行GWAS (PC-GWAS)。本研究執行了白蛋白尿、eGFR和eGFR slope的單性狀GWAS，以PC-GWAS。藉由比較不同方法之間的結果，以及對比文獻之後可以發現PC-GWAS找到20個新位點，其中有些鄰近的基因從文獻上由其功能描述或是小鼠實驗結果已經可以推估和腎臟功能或疾病相關；另外有4個位點在腎臟表現數量性狀基因座資料庫發現其具顯著的影響。未來設計功能分析研究來這些新位點是否和腎臟功能確實相關仍然是目前檢視GWAS結果的黃金標準。藉由本研究的經驗，PC-GWAS的方式也值得在適合的疾病和族群應用和推廣。

Chronic kidney disease (CKD) is an important issue global-wide, as it possesses tremendous medical resource burden。There are numerous epidemiologic studies finding conventional risk factors for CKD, and the operational definition for CKD patients in these studies is usually individuals with estimated glomerular filtration rate (eGFR) <60ml/min/1.73m2. However, many patients with CKD stage 3-5 have stable renal function for years, and thus the change of renal function with time is more important than the current CKD stage. Proteinuria/albuminuria is the major determinant of eGFR trajectory. National Kidney Foundation has suggested change in albuminuria and eGFR slope fulfill criteria for surrogate end points in clinical trials for chronic kidney disease progression. We can apply genome-wide association study (GWAS) on complex diseases such as CKD to find its heritable factors. Only few individuals of Taiwan biobank have eGFR below 60ml/min/1.73m2, and thus a case-control GWAS may not be appropriate. In addition, evaluating the risk factor for eGFR change and albuminuria is even more important for renal prognosis. We have performed quantitative GWASs regarding eGFR per se, eGFR slope and albuminuria to find risk loci for CKD progression. There were also several multi-trait methods developed in recent years to enhance the power of GWAS. We used principal component analysis to convert these quantitative traits in to pseudotraits (principal components; PC), and then perform the GWAS of the pseudotraits (PC-based GWAS). A comparison of result of different methods and literature review were done. Twenty novel risk loci associated with kidney traits were identified by PC-based GWAS. Subsequent post-GWAS analyses and functional analyses will be done accordingly in future.