Genome-wide association study and candidate gene prediction of kernel starch content in maize

Maize is an important food crop worldwide, and about 70% of its grain weight comes from starch. Starch is not only the main energy resource for human and other animals but also an important raw material for chemical industries. In this study, an association panel including 711 maize inbred lines was used for the examination of both wet-base starch content and dry-base starch content of maize kernel from two environments combined with 2799 single nucleotide polymorphism (SNP) markers spaning the whole genome of maize, genome-wide association study (GWAS) was carried out using FarmCPU model. 67 significant SNPs were identified, of which 23 highly reliable significant SNPs (HRS-SNPs) could be repeatedly associated in different environments. Three HRS-SNPs were reported for the first time by our study, and the remaining 20 HRS-SNPs were either located within the known quantitative trait loci (QTLs) or within 1 Mb of known SNPs associated with mazie kernel starch content. Through gene function annotation, gene ontology (GO) analysis and gene expression analysis, a total of 45 important candidate genes, involving starch biosynthesis, carbohydrate metabolism, sugar metabolism, hormone metabolism, and other pathways were identified within 200 kb regions around the HRS-SNPs. And two genes Ae1 and Pin1 which had been reported to regulate the maize kernel starch content were also detected. Furthermore, elite alles of the nine major SNPs was identified by allelic variation effect analysis. Our study provides new genetic information for further disecting the genetic mechanism of maize kernel starch content and provides important gene resources for accelerating the breeding of new maize varieties with high yield and quality.