Abstract With the advent of molecular marker techniques in the past two decades, genome-wide association study (GWAS) was proved to be an effective tool to reveal genetic architecture of complex traits in human, animal and plants. GWAS typically focuses on associations between genetic markers and quantitative traits in natural populations and takes advantage of recombination events in the evolutionary history. In human, more than 6 000 variant loci were discovered to associate with > 500 quantitative traits and complex diseases. In animals, GWAS was conducted specially on economically important traits, genetic defect diseases and other complex diseases of the main livestock and poultries. In plants, GWAS has been applied to study flowering time, developmental traits and agronomical traits of Arabidopsis, rice, maize and cotton.
Despite the initial success of GWAS that has been achieved, the uncovered associated loci usually have small effects on phenotype and only account for very limited phenotypic variation. The remaining unexplained genetic variance is the socalled “missing heritability”. Three possible factors were responsible for the failure of detecting the cause loci. First, the efficiency of detecting the smalleffect loci is very low and more smalleffect loci are undiscovered. Most GWASs proceed on the base of the assumption that common phenotypic variation is caused by common genetic variation. The power to detect the cause loci is a function of allele frequency, thus it is difficult to identify the functional variants at low frequency though they have larger effects on the phenotype. Second, GWAS was unable to deal with the phenotypic variances caused by structural variation (i.e. copy number variation). Third, current GWASs pay little attention to the interactions among the genetic variances and ones between genetic and environmental factors, which have been affirmed by the results of linkage analysis.
New strategies for GWAS were discussed. The package GMDR-GPU was developed to analyze epistasis effects, and the software QTXNetwork could simultaneously research single locus effect, digenic epistasis effect and their environment interactions in a full genetic model. The unbiased prediction of genetic effects could be obtained.
GWAS would make breakthrough in two aspects for the foreseeable future, due to the increasing availability of high throughput genome sequencing for human and plants. First, the increased advances in “omics” technology (transcriptomics, proteomics and metabolomics) will provide an opportunity to study the association of phenotypic variations with mRNA, protein or metabolite, which position the omics loci linked to the interested traits. Second, multitrait GWAS will improve statistical power for identifying genes contributing to complex traits.
Duan Zhongqu, Zhu Jun. Research progress of genome-wide association study. Journal of Zhejiang University (Agriculture and Life Sciences), 2015, 41(4): 385-393.
全基因组关联分析(genomewide association study，GWAS)是近年来兴起的遗传分析方法，在人类和动植物复杂性状遗传研究中已取得初步成果.本文论述了GWAS研究的基本原理、主要分析方法及常用软件，在人类和动植物复杂性状研究中的应用；分析了GWAS研究中“丢失遗传率”的主要影响因素；介绍了上位性分析的新策略和基于GPU并行计算和混合线性模型的分析软件QTXNetwork；展望了GWAS研究的发展方向.
