Identification and Validation of SNPs Associated with a Maize Yield Component Trait via GWAS

Advances in next generation sequencing (NGS) technologies and the development of appropriate populations and statistical methods enable the genome-wide dissection of the genetic determinants of traits. Several yield component traits were collected from four types of crosses associated with the Nested Association Mapping (NAM) population: NAM recombinant inbred lines (RILs, N=5,000), B73 x NAM RILs (N=700), Mo17 x NAM RILs (N=300) and a partial diallel of the NAM parents (N=250). A set of 4.9 million genic SNPs was identified via the analysis of RNA-seq data from four tissues from each of the NAM parents. These genic SNPs were combined with maize HapMap1 and HapMap2 SNPs to form a set of ~13 million SNPs that were projected onto the four cross type populations based on NAM parental and RIL genotypes. A joint composite QTL analysis was conducted using ~1,000 tagging SNPs from the Maize HapMap project. Next, three different approaches were used to conduct GWAS: 1) Bayesian-based simultaneous model fitting, 2) stepwise regression and 3) one-by-one SNP scanning. Approximately 20 trait associated SNPs (TASs) were identified by at least two methods. These SNPs, combined with an additional ~200 method-specific TASs were selected for validation using three unrelated populations. As compared to control SNPs, many of the TASs were associated with variation for yield component traits in one or more of three unrelated populations, thereby validating these TASs.