Study of genome-wide association in contrasting environments

Breeding tools such as Marker-assisted Selection (MAS), Genomic selection (GS), Genome-wide association studies (GWAS), and quantitative trait loci (QTL) have played a significant role in enabling the great leap and results experienced from breeding programs. GWAS has been widely applied in maize breeding due to the genetic variability, distinct population, and availability of marker information from maize studies. To understand how different environments affect the performance of the maize variety, a study of the association between the trait under study and molecular markers in a multi-environment trial is essential. This study aimed to explore the effect of different environments identifying markers associated with maize yield in multi-site trials using the GWAS tool. The 2018 maize dataset from the genome-2-field (G2F) initiative (available at https://www.genomes2fields.org/home/) was analyzed for yield trait marker association. The choice of locations to include for this study was possible after evaluating the available environmental data by performing the multivariate exploratory technique of principal components (PC). Three sites were selected based on a more significant divergence among environmental variables (soil temperature, temperature, dew point, solar radiation, photoperiod, soil moisture, and rainfall). To determine the potential association between SNPs and phenotypic variants, we carried out statistical analysis for the yield trait with each SNP across the maize genome for each site. GWAS analysis was performed by implementing the fixed and random model circulating probability unification (FarmCPU). All the analyses were carried out within the R environment. The first two PCs were selected, explaining 89.25% of the variance. About the sites chosen, the sites NYH2 (New York State) and IAH3 (Wisconsin State) show specific properties concerning the evaluated variables, especially for variables with strong associations in each PC. GAH2 (Georgia State) presented similarities with most other sites evaluated in the PC. The study identified significant SNPs on chromosomes 1, 2, 3, 7, and 8. In site GAH2, 62 significant SNPs were observed on chromosomes 1 and 8. In contrast, from site IAH3, we observed 247 significant SNPs. From site NYH2, we observed 12 significant SNPs. Several studies have demonstrated the superior performance of FarmCPU in GWAS compared to other methods, especially for traits affected by population structure and relatedness. The present study concluded that the number of significant SNPs identified varied across sites. These differences could have been attributed to the environmental conditions such crops were subjected to in the different sites. Thus, different environments affected the identifying markers associated with the yield in maize and should be considered for future genomic association studies.