QTN detection and candidate gene identification for improved eating and cooking quality in rice using GWAS and PLS regression analysis

We performed GWAS for starch properties and eating, cooking and appearance quality characteristics traits in rice and then used PLS regression to show importance of different loci for different food applications. We performed a genome-wide association study for appearance, eating and cooking quality traits in grain of japonica rice cultivars and identified candidate genes for adhesiveness of cooked rice grains, amylopectin composition and β-glucanase activity in rice endosperm among a total of 525 quantitative trait nucleotide (QTN) loci. The study used 1,054,635 single-nucleotide polymorphisms (SNPs) based on genome sequence data of 150 rice cultivars and 89 grain appearance, eating and cooking quality traits. These included grain shape, protein content, amylose content, amylopectin chain length, starch viscosity properties, starch degradation enzyme activities, and physicochemical characteristics of cooked rice grains analyzed in three years. Cluster regions of genetic loci on rice chromosomes 1, 4, 5, 6, 8, 9, 10 and 11 were detected, with several regions co-located with starch biosynthesis and degradation genes. Partial least squares (PLS) regression analysis revealed that the QTN genotypes were unevenly distributed in subpopulations of rice cultivars classified by their primary application. We could therefore select and accumulate these QTNs to improve grain quality in further breeding programs by developing novel rice cultivars with appropriate phenotypes for each food usage: as high eating quality cooked rice, staple food rice, sushi rice, sake brewing rice, and high-yielding rice cultivars.