遺傳相關研究中的基因劑量分析：潛在的偏差及其校正方法

Abstract

探討基因與疾病相關性的病例對照研究中，許多研究進行基因劑量分析。作者發現其中約有三分之一的研究，其定義高危險基因型(高危險基因)的方式是依資料本身該基因型頻率(基因頻率)病例組大於對照組而定。作者進行蒙地卡羅模擬，發現此種定義方式下即使各基因座與疾病無相關，亦會呈現偽陽性的基因劑量關係。作者提出排列校正方法，可以成功地校正各偏差。本研究所提方法執行容易且校正效果良好，可望成為基因劑量效應研究之標準程序。

Many case-control association studies performed the “gene-dose analysis” to investigate the association between multiple loci and the disease. The authors found that the researchers in about one-third of these studies defined the high-risk genotypes (high-risk alleles) as those that have higher genotype frequencies (allele frequencies) in the case group than in the control in the study. The authors performed Monte-Carlo simulation and found that even though there is no association between each locus and the disease, this kind of definition will show false-positive gene-dose relationships. The authors proposed a permutation correction method which is easy to perform and can correct the biases very effectively. The permutation correction method may become a standard procedure for a valid gene-dose analysis in the days to come.