Genome‐Wide Association Study Using Historical Breeding Populations Discovers Genomic Regions Involved in High‐Quality Rice

Abstract

Core Ideas Genome‐wide association study (GWAS) for rice quality was performed in two breeding populations. Twenty‐two putative quantitative trait loci (QTL) were associated to rice quality. A genomic region on chromosome 6 was associated with all quality traits in the tropical japonica population. Markers for favorable haplotypes are ready for immediate use for selection. Rice (Oryza sativa L.) is one of the most important staple food crops in the world; however, there has recently been a shift in consumer demand for higher grain quality. Therefore, understanding the genetic architecture of grain quality has become a key objective of rice breeding programs. Genome‐wide association studies (GWAS) using large diversity panels have successfully identified genomic regions associated with complex traits in diverse crop species. Our main objective was to identify genomic regions associated with grain quality and to identify and characterize favorable haplotypes for selection. We used two locally adapted rice breeding populations and historical phenotypic data for three rice quality traits: yield after milling, percentage of head rice recovery, and percentage of chalky grain. We detected 22 putative quantitative trait loci (QTL) in the same genomic regions as starch synthesis, starch metabolism, and cell wall synthesis‐related genes are found. Additionally, we found a genomic region on chromosome 6 in the tropical japonica population that was associated with all quality traits and we identified favorable haplotypes. Furthermore, this region is linked to the OsBEI gene that codes for a starch branching enzyme I, which is implicated in starch granule formation. In tropical japonica, we also found two putative QTL linked to OsBEII, OsDEP1, and OsDEP2. Our study provides an insight into the genetic basis of rice grain chalkiness, yield after milling, and head rice, identifying favorable haplotypes and molecular markers for selection in breeding programs.

Integrating Molecular Markers and Environmental Covariates To Interpret Genotype by Environment Interaction in Rice (Oryza sativaL.) Grown in Subtropical Areas

“…The indica population consisted of 327 elite breeding lines evaluated over three years (2010-2012), and the field design consisted in a randomized complete block design with two or three replications. Trait correlations, heritabilities, and genomic prediction accuracies for this dataset were computed in previous studies (Rosas et al 2017, Monteverde et al 2018, Quero et al 2018). The tropical japonica population consisted of 320 elite breeding lines evaluated over five years (2009-2013).…”

“…The grain quality traits measured were Percentage of Head Rice Recovery (PHR measured in grams, as the weight of whole milled kernels, using a 100g sample of rough rice), and the percentage of Chalky Grain (GC measured as % of chalky kernels in a subsample of 50 g of total milled rice, where the area of chalk -core, white back or white belly- was larger than half the kernel area based on visual inspection). More details about how grain quality traits were measured can be found in Quero et al (2018) and Monteverde et al (2018).…”

“…Imputation of missing data were performed with the FILLIN algorithm implemented in TASSEL 5.0 (Swarts et al 2014) for both datasets separately. The GBS datasets were filtered to retain markers with <50% missing data after imputation, and a minor allele frequency MAF > 0.05, as reported by Quero et al (2018), and Monteverde et al (2018). The final indica and tropical japonica marker dataset consisted of 92,430 and 44,598 SNP markers respectively.…”

“…In a previous study we showed that accounting for heterogeneous covariance parameters between pairs of environments can be beneficial for predicting yield and milling quality performance in Uruguayan rice for untested environments (Monteverde et al 2018). In another study, Quero et al (2018) found a set of QTL for milling yield traits in the same Uruguayan indica and tropical japonica populations. However, none of these studies tested the use of EC to both predict yield and milling quality traits in untested environments, and investigate QTL responses in specific environments.…”

Training Population Optimization for Genomic Selection

“…The average imputation accuracy was approximately 94% for both the indica and tropical japonica datasets. Single nucleotide polymorphism markers that had more than 50% missing data after the imputation along with monomorphic SNPs and SNPs with a minor allele frequency below 5% were removed from the datasets, as reported in Quero et al (2018). The remaining missing data were imputed by the mean to perform principal component analysis.…”