Genome-wide Association Analysis Method For Crop Seed Traits

Crop seeds,as primary source of human staple food,animal feed and food industrial raw materials,which grow and develop on the maternal plant.Maternal plant plays a pivotal role in the development stage of seed endosperm,so the genetic expression of seed endosperm traits could be regulated by maternal plant genotypes or endosperm genotypes and mostly by both simultaneously.Understanding the genetic mechanism of seed endosperm traits is a great challenge because of its complex mechanism of multiple genetic systems,and the influence of epistasis and gene-environment interaction.With the large-scale resequencing of crops and the availability of a large number of SNP markers,genome wide association study has been a powerful tool for dissecting genetic architecture of complex agronomic traits.However,current gene mapping methods for seed trait are QTL mapping methods based on specific designed mapping populations and conventional genome-wide association analysis for diploid traits without taking account of the large difference in genetics between seed traits and common agronomic traits.On the basis of the genetic features of seed traits,a genome-wide association analysis method was proposed for dissecting genetic architecture of crop seed traits,which incorporated maternal and offspring effects into one mapping framework based on natural population,The mixed linear model was used to evaluate the maternal additive and dominance effects,endosperm(embryo)additive and dominance effects,epistatic effects,and gene-environment interaction effects.The parental genotypes were needed to infer the probability distribution of the endosperm or embryo genotypes,then to construct expectation coefficient matrix for endosperm or embryo additive and dominant respectively,the phenotypic mean of the seed quantitative trait was used to conduct genome-wide association analysis.Monte Carlo simulations investigated the influence of different models,different heritability,and different sample sizes on parameter estimation and statistic power.Based on the proposed method,corresponding software was developed for genome-wide association analysis of seed traits.As a worked example,the new model and analysis method was applied in conducting genome-wide association analysis of the cottonseed oil traits,12 significant SNP and 5 pairs of epistatic SNP were identified,indicating the effectiveness and reliability of the proposed method.