Primary root response to combined drought and heat stress is regulated via salicylic acid metabolism in maize.

The primary root is the first organ to perceive the stress signals for abiotic stress. In this study, maize plants subjected to drought, heat and combined stresses displayed a significantly reduced primary root length. Metabolic and transcriptional analyses detected 72 and 5,469 differentially expressed metabolites and genes in response to stress conditions, respectively. The functional annotation of differentially expressed metabolites and genes indicated that primary root development was mediated by pathways involving phenylalanine metabolism, hormone metabolism and signaling under stress conditions. Furthermore, we found that the concentration of salicylic acid and two precursors, shikimic acid and phenylalanine, showed rapid negative accumulation after all three stresses. The expression levels of some key genes involved in salicylic acid metabolism and signal transduction were differentially expressed under stress conditions. This study extends our understanding of the mechanism of primary root responses to abiotic stress tolerance in maize.

OsCOMT, encoding a caffeic acid O-methyltransferase in melatonin biosynthesis, increases rice grain yield through dual regulation of leaf senescence and vascular development.

Melatonin, a natural phytohormone in plants, plays multiple critical roles in plant growth and stress responses. Although melatonin biosynthesis-related genes have been suggested to possess diverse biological functions, their roles and functional mechanisms in regulating rice grain yield remain largely unexplored. Here, we uncovered the roles of a caffeic acid O-methyltransferase (OsCOMT) gene in mediating rice grain yield through dual regulation of leaf senescence and vascular development. In vitro and in vivo evidence revealed that OsCOMT is involved in melatonin biosynthesis. Transgenic assays suggested that OsCOMT significantly delays leaf senescence at the grain filling stage by inhibiting degradation of chlorophyll and chloroplast, which, in turn, improves photosynthesis efficiency. In addition, the number and size of vascular bundles in the culms and leaves were significantly increased in the OsCOMT-overexpressing plants, while decreased in the knockout plants, suggesting that OsCOMT plays a positive role in vascular development of rice. Further evidence indicated that OsCOMT-mediated vascular development might owe to the crosstalk between melatonin and cytokinin. More importantly, we found that OsCOMT is a positive regulator of grain yield, and overexpression of OsCOMT increase grain yield per plant even in a high-yield variety background, suggesting that OsCOMT can be used as an important target for enhancing rice yield. Our findings shed novel insights into melatonin-mediated leaf senescence and vascular development and provide a possible strategy for genetic improvement of rice grain yield.

Inference of Chromosome-Length Haplotypes Using Genomic Data of Three or a Few More Single Gametes.

Compared with genomic data of individual markers, haplotype data provide higher resolution for DNA variants, advancing our knowledge in genetics and evolution. Although many computational and experimental phasing methods have been developed for analyzing diploid genomes, it remains challenging to reconstruct chromosome-scale haplotypes at low cost, which constrains the utility of this valuable genetic resource. Gamete cells, the natural packaging of haploid complements, are ideal materials for phasing entire chromosomes because the majority of the haplotypic allele combinations has been preserved. Therefore, compared with the current diploid-based phasing methods, using haploid genomic data of single gametes may substantially reduce the complexity in inferring the donor's chromosomal haplotypes. In this study, we developed the first easy-to-use R package, Hapi, for inferring chromosome-length haplotypes of individual diploid genomes with only a few gametes. Hapi outperformed other phasing methods when analyzing both simulated and real single gamete cell sequencing data sets. The results also suggested that chromosome-scale haplotypes may be inferred by using as few as three gametes, which has pushed the boundary to its possible limit. The single gamete cell sequencing technology allied with the cost-effective Hapi method will make large-scale haplotype-based genetic studies feasible and affordable, promoting the use of haplotype data in a wide range of research.

Exploiting natural variation in crown root traits via genome-wide association studies in maize.

BACKGROUND: Root system architecture (RSA), which is determined by the crown root angle (CRA), crown root diameter (CRD), and crown root number (CRN), is an important factor affecting the ability of plants to obtain nutrients and water from the soil. However, the genetic mechanisms regulating crown root traits in the field remain unclear. METHODS: In this study, the CRA, CRD, and CRN of 316 diverse maize inbred lines were analysed in three field trials. Substantial phenotypic variations were observed for the three crown root traits in all environments. A genome-wide association study was conducted using two single-locus methods (GLM and MLM) and three multi-locus methods (FarmCPU, FASTmrMLM, and FASTmrEMMA) with 140,421 SNP. RESULTS: A total of 38 QTL including 126 SNPs were detected for CRA, CRD, and CRN. Additionally, 113 candidate genes within 50 kb of the significant SNPs were identified. Combining the gene annotation information and the expression profiles, 3 genes including GRMZM2G141205 (IAA), GRMZM2G138511 (HSP) and GRMZM2G175910 (cytokinin-O-glucosyltransferase) were selected as potentially candidate genes related to crown root development. Moreover, GRMZM2G141205, encoding an AUX/IAA transcriptional regulator, was resequenced in all tested lines. Five variants were identified as significantly associated with CRN in different environments. Four haplotypes were detected based on these significant variants, and Hap1 has more CRN. CONCLUSIONS: These findings may be useful for clarifying the genetic basis of maize root system architecture. Furthermore, the identified candidate genes and variants may be relevant for breeding new maize varieties with root traits suitable for diverse environmental conditions.

Incorporation of parental phenotypic data into multi-omic models improves prediction of yield-related traits in hybrid rice.

Hybrid breeding has been shown to effectively increase rice productivity. However, identifying desirable hybrids out of numerous potential combinations is a daunting challenge. Genomic selection holds great promise for accelerating hybrid breeding by enabling early selection before phenotypes are measured. With the recent advances in multi-omic technologies, hybrid prediction based on transcriptomic and metabolomic data has received increasing attention. However, the current omic-based hybrid prediction has ignored parental phenotypic information, which is of fundamental importance in plant breeding. In this study, we integrated parental phenotypic information into various multi-omic prediction models applied in hybrid breeding of rice and compared the predictabilities of 15 combinations from four sets of predictors from the parents, that is genome, transcriptome, metabolome and phenome. The predictability for each combination was evaluated using the best linear unbiased prediction and a modified fast HAT method. We found significant interactions between predictors and traits in predictability, but joint prediction with various combinations of the predictors significantly improved predictability relative to prediction of any single source omic data for each trait investigated. Incorporation of parental phenotypic data into various omic predictors increased the predictability, averagely by 13.6%, 54.5%, 19.9% and 8.3%, for grain yield, number of tillers per plant, number of grains per panicle and 1000 grain weight, respectively. Among nine models of incorporating parental traits, the AD-All model was the most effective one. This novel strategy of incorporating parental phenotypic data into multi-omic prediction is expected to improve hybrid breeding progress, especially with the development of high-throughput phenotyping technologies.

Integrating transcriptome, co-expression and QTL-seq analysis reveals that primary root growth in maize is regulated via flavonoid biosynthesis and auxin signal transduction.

The primary root is critical for early seedling growth and survival. To understand the molecular mechanisms governing primary root development, we performed a dynamic transcriptome analysis of two maize (Zea mays) inbred lines with contrasting primary root length at nine time points over a 12-day period. A total of 18 702 genes were differentially expressed between two lines or different time points. Gene enrichment, phytohormone content determination, and metabolomics analysis showed that auxin biosynthesis and signal transduction, as well as the phenylpropanoid and flavonoid biosynthesis pathways, were associated with root development. Co-expression network analysis revealed that eight modules were associated with lines/stages, as well as primary or lateral root length. In root-related modules, flavonoid metabolism accompanied by auxin biosynthesis and signal transduction constituted a complex gene regulatory network during primary root development. Two candidate genes (rootless concerning crown and seminal roots, rtcs and Zm00001d012781) involved in auxin signaling and flavonoid biosynthesis were identified by co-expression network analysis, QTL-seq and functional annotation. These results increase our understanding of the regulatory network controlling the development of primary and lateral root length, and provide a valuable genetic resource for improvement of root performance in maize.

Genetic control of root plasticity in response to salt stress in maize.

KEY MESSAGE: GWAS identified 559 significant SNPs associated with the remodelling of the root architecture in response to salt, and 168 candidate genes were prioritized by integrating RNA-seq, DEG and WGCNA data. Salinity is a major environmental factor limiting crop growth and productivity. The root is the first plant organ to encounter salt stress, yet the effects of salinity on maize root development remain unclear. In this study, the natural variations in 14 root and 4 shoot traits were evaluated in 319 maize inbred lines under control and saline conditions. Considerable phenotypic variations were observed for all traits, with high salt concentrations decreasing the root length, but increasing the root diameter. A genome-wide association study was conducted to analyse these traits and their plasticity (relative variation). We detected 559 significant single nucleotide polymorphisms, of which 125, 181 and 253 were associated with the control condition, stress condition and trait plasticity, respectively. A total of 168 of 587 candidate genes identified by genome-wide association study were supported by the differentially expressed genes or co-expression networks. Two candidate genes ZmIAA1 and ZmGRAS43 were validated by resequencing. Among these genes, 130 were detected under stress condition or trait plasticity that involved in diverse biological processes including plant hormone signal transduction, phenylpropanoid biosynthesis and fatty acid biosynthesis. Our findings clarify the root remodelling to salinity, and the identified loci and candidate genes may be important for the genetic improvement of root traits and salt tolerance in maize.

Genetic and genomic analysis of the seed-filling process in maize based on a logistic model.

Seed filling is a dynamic process that determines seed size and nutritional quality. This time-dependent trait follows a logistic (S-shaped) growth curve that can be described by a logistic function, with parameters of biological relevance. When compared between genotypes, the filling dynamics variations are explained by the differences of parameter values; as such, the parameter estimates can be considered as "traits" for genetic analysis to identify loci that are associated with the seed-filling process. We carried out genetic and genomic analysis of the seed-filling process in maize, using a recombinant inbred line (RIL) population derived from the two inbred lines with contrasting seed-filling dynamics. We recorded seed dry weight at 14 time points after pollination, spanning the early filling phases to the late maturation stages. Fitting these data to a logistic model allowed for estimating 12 characteristic parameters that can be used to meaningfully describe the seed-filling process. Quantitative trait locus (QTL) mapping of these parameters identified a total of 90 nonredundant loci. Using bulked segregant RNA-sequencing (BSR-seq) analysis, we identified eight genes that showed differential gene expression patterns at multiple time points between the extreme pools, and these genes co-localize with the mapped QTL regions. Two of the eight genes, GRMZM2G391936 and GRMZM2G008263, are implicated in starch and sucrose metabolism, and biosynthesis of secondary metabolites that are well known for playing a vital role in seed filling. This study suggests that the logistic model-based approach can efficiently identify genetic loci that regulate dynamic developing traits.

Natural variation of ZmHKT1 affects root morphology in maize at the seedling stage.

MAIN CONCLUSION: Eight variants in ZmHKT1 promoter were significantly associated with root diameter, four haplotypes based on these significant variants were found, and Hap2 has the largest root diameter. Roots play an important role in uptake of water, nutrients and plant anchorage. Identification of gene and corresponding SNPs associated with root traits would enable develop maize lines with better root traits that might help to improve capacity for absorbing nutrients and water acquisition. The genomic sequences of a salt tolerance gene ZmHKT1 was resequenced in 349 maize inbred lines, and the association between nucleotide polymorphisms and seedling root traits was detected. A total of 269 variants in ZmHKT1 were identified, including 226 single nucleotide polymorphisms and 43 insertions and deletions. The gene displayed high level of nucleotide diversity, especially in non-genic regions. A total of 19 variations in untranslated region of ZmHKT1 were found to be associated with six seedling traits. Eight variants in promoter region were significantly associated with average root diameter (ARD), four haplotypes were found based on these significant variants, and Hap2 has the largest ARD. Two SNPs in high-linkage disequilibrium (SNP-415 and SNP 2169) with pleiotropic effects were significantly associated with plant height, root surface area, root volume, and shoot dry weight. This result revealed that ZmHKT1 was an important contributor to the phenotypic variations of seedling root traits in maize, these significant variants could use to develop functional markers to improve root traits.

Origin, evolution and functional characterization of the land plant glycoside hydrolase subfamily GH5_11.

Colonization of the land by plants was a critical event in the establishment of modern terrestrial ecosystems, and many characteristics of land plants originated during this process, including the emergence of rosette terminal cellulose-synthesizing complexes. Cellulases are non-homologous isofunctional enzymes, encoded by glycosyl hydrolase (GH) gene families. Although the plant GH5_11 gene subfamily is presumed to encode a cell-wall degrading enzyme, its evolutionary and functional characteristics remain unclear. In the present study, we report the evolution of the land plant GH5_11 subfamily, and the functions of its members in terms of cellulase activity, through comprehensive phylogenetic analyses and observation of Arabidopsis mutants. Phylogenetic and sequence similarity analyses reveal that the ancestor of land plants acquired the GH5_11 gene from fungi through a horizontal gene transfer (HGT) event. Subsequently, positive selection with massive gene duplication and loss events contributed to the evolution of this subfamily in land plants. In Arabidopsis and rice, expression of GH5_11 genes are regulated by multiple abiotic stresses, the duplicated genes showing different patterns of expression. The Arabidopsis mutants atgh5_11a and atgh5_11c display low levels of cellulase and endoglucanase activities, with correspondingly high levels of cellulose, implying that the encoded proteins may function as endoglucanases. However, atgh5_11a and atgh5_11c also display an enlarged rosette leaf phenotype, and atgh5_11c is late-flowering under short photoperiods. These observations suggest that plant GH5_11s possess more functions beyond being endonucleases. To summarize, we demonstrate that the ancestor of land plants has acquired GH5_11 gene through HGT, which extends the cellulose degradation complexity. Our investigations illuminate features of part of the molecular framework underlying the origin of land plants and provide a focus on the cellulose degradation pathway.

Combined linkage and association mapping reveal candidate loci for kernel size and weight in maize.

Yield improvement is a top priority for maize breeding. Kernel size and weight are important determinants of maize grain yield. In this study, a recombinant inbred line (RIL) population and an association panel were used to identify quantitative trait loci (QTLs) for four maize kernel-related traits: kernel length, width, thickness and 100-kernel weight. Twenty-seven QTLs were identified for kernel-related traits across three environments and the best linear unbiased predictions (BLUPs) of each trait by linkage analysis, and four QTLs were stably detected in more than two environments. Additionally, 29 single nucleotide polymorphisms (SNPs) were identified as significantly associated with the four kernel-related traits and BLUPs by genome-wide association study, and two loci could be stably detected in both environments. In total, four QTLs/SNPs were co-associated with various traits in both populations. Using combined-linkage analysis and association mapping, PZE-101066560 on chromosome 1, associated with kernel width and with 100-kernel weight in the association panel, was co-localized within the QTL interval of qKW1-3 for kernel width in the RILs. Two annotated genes in the candidate region were considered as potential candidate genes. The QTLs and candidate genes identified here will facilitate molecular breeding for grain yield improvement in maize.

Ancient horizontal transfer of transaldolase-like protein gene and its role in plant vascular development.

A major event in land plant evolution is the origin of vascular tissues, which ensure the long-distance transport of water, nutrients and organic compounds. However, the molecular basis for the origin and evolution of plant vascular tissues remains largely unknown. Here, we investigate the evolution of the land plant TAL-type transaldolase (TAL) gene and its potential function in rice (Oryza sativa) based on phylogenetic analyses and transgenic experiments, respectively. TAL genes are only present in land plants and bacteria. Phylogenetic analyses suggest that land plant TAL genes are derived from Actinobacteria through an ancient horizontal gene transfer (HGT) event. Further evidence reveals that land plant TAL genes have undergone positive selection and gained several introns following its acquisition by the most recent common ancestor of land plants. Transgenic plant experiments show that rice TAL is specifically expressed in vascular tissues and that knockdown of TAL expression leads to changes in both the number and pattern of vascular bundles. Our findings show that the ancient HGT of TAL from bacteria probably plays an important role in plant vascular development and adaptation to land environments.

Characterization of Rubisco activase genes in maize: an α-isoform gene functions alongside a ß-isoform gene.

Rubisco activase (RCA) catalyzes the activation of Rubisco in vivo and plays a crucial role in regulating plant growth. In maize (Zea mays), only ß-form RCA genes have been cloned and characterized. In this study, a genome-wide survey revealed the presence of an α-form RCA gene and a ß-form RCA gene in the maize genome, herein referred to as ZmRCAα and ZmRCAß, respectively. An analysis of genomic DNA and complementary DNA sequences suggested that alternative splicing of the ZmRCAß precursor mRNA (premRNA) at its 3' untranslated region could produce two distinctive ZmRCAß transcripts. Analyses by electrophoresis and matrix-assisted laser desorption/ionization-tandem time-of-flight mass spectrometry showed that ZmRCAα and ZmRCAß encode larger and smaller polypeptides of approximately 46 and 43 kD, respectively. Transcriptional analyses demonstrated that the expression levels of both ZmRCAα and ZmRCAß were higher in leaves and during grain filling and that expression followed a specific cyclic day/night pattern. In 123 maize inbred lines with extensive genetic diversity, the transcript abundance and protein expression levels of these two RCA genes were positively correlated with grain yield. Additionally, both genes demonstrated a similar correlation with grain yield compared with three C4 photosynthesis genes. Our data suggest that, in addition to the ß-form RCA-encoding gene, the α-form RCA-encoding gene also contributes to the synthesis of RCA in maize and support the hypothesis that RCA genes may play an important role in determining maize productivity.

Adaptive evolution and divergent expression of heat stress transcription factors in grasses.

BACKGROUND: Heat stress transcription factors (Hsfs) regulate gene expression in response to heat and many other environmental stresses in plants. Understanding the adaptive evolution of Hsf genes in the grass family will provide potentially useful information for the genetic improvement of modern crops to handle increasing global temperatures. RESULTS: In this work, we performed a genome-wide survey of Hsf genes in 5 grass species, including rice, maize, sorghum, Setaria, and Brachypodium, by describing their phylogenetic relationships, adaptive evolution, and expression patterns under abiotic stresses. The Hsf genes in grasses were divided into 24 orthologous gene clusters (OGCs) based on phylogeneitc relationship and synteny, suggesting that 24 Hsf genes were present in the ancestral grass genome. However, 9 duplication and 4 gene-loss events were identified in the tested genomes. A maximum-likelihood analysis revealed the effects of positive selection in the evolution of 11 OGCs and suggested that OGCs with duplicated or lost genes were more readily influenced by positive selection than other OGCs. Further investigation revealed that positive selection acted on only one of the duplicated genes in 8 of 9 paralogous pairs, suggesting that neofunctionalization contributed to the evolution of these duplicated pairs. We also investigated the expression patterns of rice and maize Hsf genes under heat, salt, drought, and cold stresses. The results revealed divergent expression patterns between the duplicated genes. CONCLUSIONS: This study demonstrates that neofunctionalization by changes in expression pattern and function following gene duplication has been an important factor in the maintenance and divergence of grass Hsf genes.

Analyses of sequence polymorphism and haplotype diversity of LEAFY genes revealed post-domestication selection in the Chinese elite maize inbred lines.

Post-domestication selection refers to the artificial selection on the loci controlling important agronomic traits during the process of genetic improvement in a population. The maize genes Zfl1 and Zfl2, duplicate orthologs of Arabidopsis LEAFY, are key regulators in plant branching, inflorescence and flower development, and reproduction. In this study, the full gene sequences of Zfl1 and Zfl2 from 62 Chinese elite inbred lines were amplified to evaluate their nucleotide polymorphisms and haplotype diversities. A total of 254 and 192 variants that included SNPs and indels were identified from the full sequences of Zfl1 and Zfl2, respectively. Although most of the variants were found to be located in the non-coding regions, the polymorphisms of CDS sequences classified Zfl1 into 16 haplotypes encoding 16 different proteins and Zfl2 into 18 haplotypes encoding eight different proteins. The population of Huangzaosi and its derived lines showed statistically significant signals of post-domestication selection on the Zfl1 CDS sequences, as well as lower nucleotide polymorphism and haplotype diversity than the whole set. However, the Zfl2 locus was only selected for in the heterotic group Reid. Further evidence revealed that at least 17 recombination events contributed to the genetic and haplotype diversities at the Zfl1 locus and 16 recombination events at the Zfl2 locus.

Genomic investigation of 18,421 lines reveals the genetic architecture of rice.

Understanding how numerous quantitative trait loci (QTL) shape phenotypic variation is an important question in genetics. To address this, we established a permanent population of 18,421 (18K) rice lines with reduced population structure. We generated reference-level genome assemblies of the founders and genotyped all 18K-rice lines through whole-genome sequencing. Through high-resolution mapping, 96 high-quality candidate genes contributing to variation in 16 traits were identified, including OsMADS22 and OsFTL1 verified as causal genes for panicle number and heading date, respectively. We identified epistatic QTL pairs and constructed a genetic interaction network with 19 genes serving as hubs. Overall, 170 masking epistasis pairs were characterized, serving as an important factor contributing to genetic background effects across diverse varieties. The work provides a basis to guide grain yield and quality improvements in rice.

Evolution of land plant genes encoding L-Ala-D/L-Glu epimerases (AEEs) via horizontal gene transfer and positive selection.

BACKGROUND: The L-Ala-D/L-Glu epimerases (AEEs), a subgroup of the enolase superfamily, catalyze the epimerization of L-Ala-D/L-Glu and other dipeptides in bacteria and contribute to the metabolism of the murein peptide of peptidoglycan. Although lacking in peptidoglycan, land plants possess AEE genes that show high similarity to those in bacteria. RESULTS: Similarity searches revealed that the AEE gene is ubiquitous in land plants, from bryophytas to angiosperms. However, other eukaryotes, including green and red algae, do not contain genes encoding proteins with an L-Ala-D/L-Glu_epimerase domain. Homologs of land plant AEE genes were found to only be present in prokaryotes, especially in bacteria. Phylogenetic analysis revealed that the land plant AEE genes formed a monophyletic group with some bacterial homologs. In addition, land plant AEE proteins showed the highest similarity with these bacterial homologs and shared motifs only conserved in land plant and these bacterial AEEs. Integrated information on the taxonomic distribution, phylogenetic relationships and sequence similarity of the AEE proteins revealed that the land plant AEE genes were acquired from bacteria through an ancient horizontal gene transfer (HGT) event. Further evidence revealed that land plant AEE genes had undergone positive selection and formed the main characteristics of exon/intron structures through gaining some introns during the initially evolutionary period in the ancestor of land plants. CONCLUSIONS: The results of this study clearly demonstrated that the ancestor of land plants acquired an AEE gene from bacteria via an ancient HGT event. Other findings illustrated that adaptive evolution through positive selection has contributed to the functional adaptation and fixation of this gene in land plants.

Genetic Diversification of Starch Branching Enzymes during Maize Domestication and Improvement.

Elucidating the genetic basis of starch pasting and gelatinization properties is crucial for enhancing the quality of maize and its utility as feed and industrial raw material. In maize, ZmSBE genes encode important starch branching enzymes in the starch biosynthesis pathway. In this study, we re-sequenced the genomic sequences of ZmSBEI, ZmSBEIIa, ZmSBEIIb, and ZmSBEIII in three lines called 335 inbred lines, 68 landrace lines, and 32 teosinte lines. Analyses of nucleotide polymorphisms and haplotype diversity revealed differences in the selection patterns of ZmSBEI, ZmSBEIIa, ZmSBEIIb, and ZmSBEIII during maize domestication and improvement. A marker-trait association analysis of inbred lines detected 22 significant loci, including 18 SNPs and 4 indels significantly associated with three maize starch physicochemical properties. The allele frequencies of two variants (SNP17249C and SNP5055G) were examined in three lines. The frequency of SNP17249C in ZmSBEIIb was highest in teosinte lines, followed by landrace lines, and inbred lines, whereas there were no significant differences in the frequency of SNP5055G in ZmSBEIII among the three lines. These results suggest that ZmSBE genes play an important role in the phenotypic variations in the starch physicochemical properties in maize. The genetic variants detected in this study may be used to develop functional markers for improving maize starch quality.

A de novo evolved gene contributes to rice grain shape difference between indica and japonica.

The role of de novo evolved genes from non-coding sequences in regulating morphological differentiation between species/subspecies remains largely unknown. Here, we show that a rice de novo gene GSE9 contributes to grain shape difference between indica/xian and japonica/geng varieties. GSE9 evolves from a previous non-coding region of wild rice Oryza rufipogon through the acquisition of start codon. This gene is inherited by most japonica varieties, while the original sequence (absence of start codon, gse9) is present in majority of indica varieties. Knockout of GSE9 in japonica varieties leads to slender grains, whereas introgression to indica background results in round grains. Population evolutionary analyses reveal that gse9 and GSE9 are derived from wild rice Or-I and Or-III groups, respectively. Our findings uncover that the de novo GSE9 gene contributes to the genetic and morphological divergence between indica and japonica subspecies, and provide a target for precise manipulation of rice grain shape.