Department of Systems and Computational Biology, Albert Einstein College of Medicine, Bronx, New York, United States of America.
PLoS One. 2012;7(11):e49093. doi: 10.1371/journal.pone.0049093. Epub 2012 Nov 9.
In this work we show that in genome-wide association studies (GWAS) there is a strong bias favoring of genes covered by larger numbers of SNPs. Thus, we state here that there is a need for correction for such bias when performing downstream gene-level analysis, e.g. pathway analysis and gene-set analysis. We investigate several methods of obtaining gene level statistical significance in GWAS, and compare their effectiveness in correcting such bias. We also propose a simple algorithm based on first order statistic that corrects such bias.
在这项工作中,我们表明在全基因组关联研究(GWAS)中,存在强烈的偏向,有利于由更多数量的 SNPs 覆盖的基因。因此,我们在这里指出,在进行下游基因水平分析时,例如途径分析和基因集分析,需要对此类偏差进行校正。我们研究了几种在 GWAS 中获得基因水平统计显著性的方法,并比较了它们校正这种偏差的效果。我们还提出了一种基于一阶统计的简单算法来校正这种偏差。