Grain size is an important yield-related trait in rice. Intensive artificial selection for grain size during domestication is evidenced by the larger grains of most of today's cultivars compared with their wild relatives. However, the molecular genetic control of rice grain size is still not well characterized. Here, we report the identification and cloning of Grain Size 6 (GS6), which plays an important role in reducing grain size in rice. A premature stop at the +348 position in the coding sequence (CDS) of GS6 increased grain width and weight significantly. Alignment of the CDS regions of GS6 in 90 rice materials revealed three GS6 alleles. Most japonica varieties (95%) harbor the Type I haplotype, and 62.9% of indica varieties harbor the Type II haplotype. Association analysis revealed that the Type I haplotype tends to increase the width and weight of grains more than either of the Type II or Type III haplotypes. Further investigation of genetic diversity and the evolutionary mechanisms of GS6 showed that the GS6 gene was strongly selected in japonica cultivars. In addition, a "ggc" repeat region identified in the region that encodes the GRAS domain of GS6 played an important historic role in the domestication of grain size in rice. Knowledge of the function of GS6 might aid efforts to elucidate the molecular mechanisms that control grain development and evolution in rice plants, and could facilitate the genetic improvement of rice yield.

The developmental regulation of grasses lemma and palea and their relationship to the floral organs in dicots had been variously explicated and extensively debated. Here, we characterized a triangular hull mutant th1-1 from EMS-mutagenized Oryza sativa ssp. indica cv. 93-11. The th1-1 mutant exhibited obviously triangular hull with tortuous and slender lemma/palea. Using a map-based cloning strategy, the TH1 gene was narrowed down to a 60-kb region on the long arm of chromosome 2. Sequence verification revealed that the th1-1 mutant harbored 1-bp deletion in exon 2 of LOC_Os02g56610 which resulted in a frame-shift mutation. The RNA-interference transgenic plants of LOC_Os02g56610 displayed a similar phenotype to the th1 mutant. Consequently, LOC_Os02g56610 was identified as the TH1 gene which encoded 248 amino acids and contained a DUF640 domain. RT-PCR analysis and GUS staining showed that the transcripts of TH1 mainly accumulated in young inflorescence, lemma and palea of spikelet. These results suggested that TH1 was an important gene controlling the lemma and palea development in rice.

Seed morphology and quality of cultivated soybean (Glycine max) have changed dramatically during domestication from their wild relatives, but their relationship to selection is poorly understood. Here, we describe a semi-dominant locus, ST1 (Seed Thickness 1), affecting seed thickness and encoding a UDP-D-glucuronate 4-epimerase, which catalyses UDP-galacturonic acid production and promotes pectin biosynthesis. Interestingly, this morphological change concurrently boosted seed oil content, which, along with up-regulation of glycolysis biosynthesis modulated by ST1, enabled soybean to become a staple oil crop. Strikingly, ST1 and an inversion controlling seed coat colour formed part of a single selective sweep. Structural variation analysis of the region surrounding ST1 shows that the critical mutation in ST1 existed in earlier wild relatives of soybean and the region containing ST1 subsequently underwent an inversion, which was followed by successive selection for both traits through hitchhiking during selection for seed coat colour. Together, these results provide direct evidence that simultaneously variation for seed morphology and quality occurred earlier than variation for seed coat colour during soybean domestication. The identification of ST1 thus sheds light on a crucial phase of human empirical selection in soybeans and provides evidence that our ancestors improved soybean based on taste.

Quantitative trait locus (QTL) mapping for seed longevity is essential for breeding modern cultivars with resistance to deterioration during postharvest storage. The inbred lines X178 and I178 showed large differences in seed vigour after artificial aging treatment, while they had similar performances in terms of most agronomic traits. An F2:3 population and a recombinant inbred line (RIL) population were generated to map QTL after 5 days under artificial aging conditions. Positive correlations were observed among all investigated traits including the aging germination rate, relative aging germination rate, aging simple vigour index, aging primary root length, aging shoot length and aging total length. Thirteen QTL were identified to locate on five chromosome regions: Chr.1:297 Mb (chromosome 1 region 297 Mb), Chr.3:205 Mb, Chr.4:240 Mb, Chr.5:205 Mb and Chr.7:155 Mb, with 2 to 4 QTL co‐located on a region. In each region, 3–8 previously identified aging‐related QTL were located, confirming the importance of these regions for controlling seed longevity in different maize populations. Taken together, the results of this work provide a foundation for further QTL fine mapping and the molecular‐assisted breeding of aging tolerant maize.