基于bin标记的上位性互作全基因组关联分析方法的研究

全基因组关联分析(GWAS)是解码植物基因型-表型关联的常用方法。剖析上位性互作遗传位点，有利于解析植物复杂性状遗传基础，对植物分子育种有指导意义。测序技术快速发展促使海量基因组信息的获得，同时也带来了超高计算量，阻碍了上位性互作GWAS的实施，因此降低标记密度是实现上位性互作GWAS的关键。本研究拟建立基于bin标记的上位性互作GWAS模型（bin-epGWAS）。首先，基于LD原理和方差最小原则创建bin标记分析法，将高密度SNP转化成数目显著降低的bin标记，然后利用bin标记构建可以估计主效应和二维上位性互作效应的bin-epGWAS模型；其次，开发bin-epGWAS R软件包；最后，利用1035份陆地棉材料，剖析棉纤维上位性互作遗传位点，探究棉纤维遗传分子基础。本研究将弥补两步骤上位性互作GWAS方法中遗传信息丢失的不足，为上位性互作GWAS模型的优化改进提供理论和方法参考。

Genome-wide association studies (GWAS) are statistical tools widely used to identify the associations between genotypic variants and phenotypic traits in plants. Detecting the epistasis loci can help us better dissect the genetic architectures of complex traits and thus is very instructive in molecular breeding in plants. A large amount of genomic information is obtained depend on the rapid development of sequencing technology，resulting in the huge computational cost, which is the main difficulty in the implementation of epistatic GWAS. Therefore, the key to achieve epistatic GWAS is to reduce the marker density without compromising the genomic information. In this project, we aim to develop a new bin-based epistatic GWAS model called bin-epGWAS. Firstly, a method named binAnalysis will be developed based on LD (linkage disequilibrium) and the principle of minimum variance to convert the high-dimension SNPs into the low-dimension bins. Then we will use the bins to construct the bin-epGWAS model, which can estimate the additive effects, the dominant effects, and the epistatic interaction effects. Secondly, an R package called bin-epGWAS will be developed. Then the R package will be tested and optimized using a series of simulated data. Finally, a natural cotton population including 1035 varieties will be used for detecting the epistasis interaction loci and exploring the genetic basis. The bin-epGWAS model will be superior to the two-step epistatic GWAS models, which lose some genomic information. This project will provide the theoretical and methodological reference for the optimization and improvement of epistatic GWAS model.