Introduction: Rice (Oryza sativa L.) production is being challenged by global warming. Identifying new loci and favorable alleles associated with heat tolerance is crucial to developing rice heat-tolerant varieties.Methods: We evaluated the heat tolerance at the seedling stage using 620 diverse rice accessions. A total of six loci associated with heat tolerance were identified by a genome-wide association study (GWAS) with ~2.8 million single nucleotide polymorphisms (SNPs). Results: Among the six detected loci, qHT7 harbored the strongest association signal and the most associated SNPs. By comparing the transcriptomes of two representative accessions with contrasting heat tolerance, LOC_Os07g48710 (OsVQ30) was selected as a promising candidate gene in qHT7 due to the significant difference in its expression level between the two accessions. Haplotype 4 (Hap4) of LOC_Os07g48710 was determined as the favorable haplotype for heat tolerance via the gene-based haplotype analysis. The heat-tolerant haplotype LOC_Os07g48710Hap4 is highly enriched in the tropical Geng/Japonica accessions, and its frequency has decreased significantly during the improvement process of rice varieties. Discussion: Based on the GWAS and transcriptomics integrated results, a hypothetical model modulated by qHT7 in response to heat stress was proposed. Our results provide valuable candidate genes for improving rice heat tolerance through molecular breeding.

The narrow base of genetic diversity of modern rice varieties is mainly attributed to the overuse of the common backbone parents that leads to the lack of varied favorable alleles in the process of breeding new varieties. Introgression lines (ILs) developed by a backcross strategy combined with marker-assisted selection (MAS) are powerful prebreeding tools for broadening the genetic base of existing cultivars. They have high power for mapping quantitative trait loci (QTLs) either with major or minor effects, and are used for precisely evaluating the genetic effects of QTLs and detecting the gene-by-gene or gene-by-environment interactions due to their low genetic background noise. ILs developed from multiple donors in a fixed background can be used as an IL platform to identify the best alleles or allele combinations for breeding by design. In the present paper, we reviewed the recent achievements from ILs in rice functional genomics research and breeding, including the genetic dissection of complex traits, identification of elite alleles and background-independent and epistatic QTLs, analysis of genetic interaction, and genetic improvement of single and multiple target traits. We also discussed how to develop ILs for further identification of new elite alleles, and how to utilize IL platforms for rice genetic improvement.

Whole grain rice is recommended in a natural healthy diet because of its high nutritional and healthful benefits compared to polished or white rice. The whole grain contains the pericarp with many assorted colors (such as brown, red, and black) associated with taste and commercial quality. The color attributes of whole grain or brown rice are usually undesirable and need to be improved. To decipher the genetic basis of color variation in the whole grain rice, we conducted a genome-wide association analysis of three parameters of grain colors (brightness, redness, and yellowness) in a panel of 682 rice accessions. Twenty-six loci were identified for the color parameters, implying that grain color is under polygenic control. Among them, some major-effect loci were co-localized with the previously identified genes such as Rc and Rd. To eliminate the possible mask of Rc on other loci influencing grain color, we performed the association analysis in a subset of the panel that excluded the pigmented (red and black) rice. Eighteen loci or SNPs were detected to be associated with grain color in the subpopulation, many of which were not reported before. Two significant peak SNP regions on chromosomes 1 and 9 were validated using near-isogenic lines. Based on differential expression analysis of annotated genes within the SNP regions and metabolic analysis of pooled extreme samples, we found at least three annotated genes as potential candidates involved in the flavonoid metabolic pathway related to pericarp color. These results provide insights into the genetic basis of rice grain color and facilitate genomic breeding to improve appearance and commercial quality of whole grain rice.

Grain quality is a key determinant of commercial value in rice. Efficiently improving grain quality, without compromising grain yield, is a challenge in rice breeding programs. Here we report on the identification and application of a grain quality gene, Chalk7, which causes a slender shape and decreases grain chalkiness in rice. Three allele-specific markers for Chalk7, and two other grain genes (GS3 and Chalk5) were developed, and used to stack the desirable alleles at these loci. The effects of individual or combined alleles at the loci were evaluated using a set of near-isogenic lines, each containing one to three favorable alleles in a common background of an elite variety. We found that the favorable allele combination of the three loci, which rarely occurs in natural rice germplasm, greatly reduces chalky grains without negatively impacting on grain yield. The data for newly developed allele-specific markers and pre-breeding lines will facilitate the improvement of grain appearance quality in rice.

High selenium (Se) and low cadmium (Cd) and arsenic (As) contents in rice grains were good for human health. The genetic basis and relationship of Se, Cd and As concentrations in rice grains are still largely unknown. In the present study, large variations were observed in Se, Cd and As concentrations in brown and milled rice in normal and Se treatment conditions in 307 rice accessions from 3K Rice Genomes Project. Se fertilizer treatment greatly increased Se concentrations but had no obvious changes in concentrations of Cd and As both in brown and milled rice. Total of 237 QTL were identified for Se, Cd and As concentrations in brown and milled rice in normal and Se treatment conditions as well as ratio of concentrations under Se treatment to normal conditions. Only 19 QTL (13.4%) were mapped for concentrations of Se and Cd, Se and As, and Se, Cd and As in the same or adjacent regions, indicating that most Se concentration QTL are independent of Cd and As concentration QTL. Forty-three favorable alleles were identified for 40 candidate genes by gene-based association study and haplotype analysis in 14 important QTL regions. Se-enriched rice variety will be developed by pyramiding favorable alleles at different Se QTL and excluding undesirable alleles at Cd and As QTL, or combining favorable alleles at Se QTL with the alleles at Se-sensitive QTL by marker-assisted selection.