基于单核苷酸多态性的基因互作分析方法学进展

Advances in development of gene-gene interaction analysis methods based on SNP data: a review

基于单核苷酸多态性的关联分析已成为当前解析人类常见复杂疾病遗传机制的重要手段之一,然而,目前普遍使用的单位点分析策略仅能发现部分单独效应显著的易感SNP位点,因此遗漏了重要的遗传力组分——基因上位效应或联合效应。识别全基因组多基因间复杂的互作关系已成为全面解析复杂疾病致病分子机制必不可少的一项任务。已有很多方法被应用于全基因组交互作用分析,加深了人类对复杂疾病遗传机制的进一步认识。基于各类方法的理论基础及算法的异同,文章对目前应用较为广泛的基于遗传互作模型的方法、不基于互作模型的方法和数据挖掘类算法3类方法进行了系统地评述,着重介绍了这些方法的主要思想、实现过程及应用中的注意事项等,并指出开展大规模全基因组范围互作检测面临的问题,以期能为相关领域的研究者提供方法学参考。

The SNP-based association analysis has become one of the most important approaches to interpret the under- lying molecular mechanisms for human complex diseases. Nevertheless, the widely-used singe-locus analysis is only capa- ble of capturing a small portion of susceptible SNPs with prominent marginal effects, leaving the important genetic compo- nent, epistasis or joint effects, to be undetectable. Identifying the complex interplays among multiple genes in the ge- nome-wide context is an essential task for systematically unraveling the molecular mechanisms for complex diseases. Many approaches have been used to detect genome-wide gene-gene interactions and provided new insights into the genetic basis of complex diseases. This paper reviewed recent advances of the methods for detecting gene-gene interaction, categorizedinto three types, model-based and model-free statistical methods, and data mining methods, based on their characteristics in theory and numerical algorithm. In particular, the basic principle, numerical implementation and cautions for application for each method were elucidated. In addition, this paper briefly discussed the limitations and challenges associated with detect- ing genome-wide epistasis, in order to provide some methodological consultancies for scientists in the related fields.