Flag leaf phenotypic variation of the parents
There are large variations in leaf traits between XieqingzaoB (XB) and Restore line Zhonghui9308 (ZH9308). The phenotypic difference between the two parents was presented and listed in Fig. 1 and Table 1, respectively. The t-tests suggested that significant differences between XB and ZH9308 for FLL, FLW, FLR, and yield per plant (PY) in Hangzhou (HZ) and in Hainan (HN). Compared with XB, the FLL of ZH9308 is longer by 8.8 cm and 9.15 cm, FLW is narrower by 0.20 cm and 0.44 cm, FLR is more curl by 55.48% and 30.49%, and PY is higher by 13.75 g and 7.44 g in HN and HZ, respectively.
Table 1
Variations of phenotypes between parents in Hainan and Hangzhou.
Variety | FLL (cm) | FLW (cm) | FLR (%) | PY (g) |
XB-HN | 28.20 ± 3.19 | 1.66 ± 0.08 | 6.00 ± 0.30 | 12.81 ± 2.00 |
ZH9308-HN | 47.00 ± 4.24 *** | 1.46 ± 0.11 *** | 61.48 ± 16.34 *** | 26.56 ± 8.40 * |
XB-HZ | 30.71 ± 2.29 | 1.70 ± 0.12 | 0.00 ± 0.00 | 20.02 ± 3.47 |
ZH9308-HZ | 39.86 ± 3.62 *** | 1.26 ± 0.05 *** | 30.49 ± 13.01 *** | 27.46 ± 7.47 *** |
Mean ± SD (n = 16). *, ** and ***indicate the least significant difference at 0.05, 0.01 and 0.001 probability level compared with XB in Hangzhou or Hainan, respectively. FLL = flag leaf length; FLW = flag leaf width; FLR = flag leaf rolling; PY = per plant yield. |
Population structure, linkage disequilibrium pattern, and trait variation in the RILs
Linkage disequilibrium (LD) analysis has been proved in our previous research (Zhao et al. 2021) based on 139 RILs and two parents. Our result showed that these groups have no population structure and the LD decay rate was estimated at almost 2000 kb on the genome level.
The morphology of the flag leaf was investigated, and noteworthy differences were observed for FLL, FLW, and FLR in HZ and HN (Table 2 and Fig. 2). The flag leaf traits including length, width, and rolling were distributed continuously with large variations and transgressive segregation, showing a typical pattern of quantitative variation in this population at both HZ and HN. In HZ, the FLL of the population ranged from 23.5 cm to 59.07 cm with a mean of 36.93 cm. FLW varied from 1.24 cm to 2.27 cm with a mean of 1.66 cm. FLR was averaged at 10 ranging from 0 to 47%. In HN, FLL ranged from 27.04 cm to 55.93 cm with a mean of 39.86 cm. FLW varied from 1.14 cm to 2.36 cm with a mean of 1.76 cm. FLR was averaged at 21 ranging from 4–71%.
Table 2
Summary statistics of three flag leaf traits of RILs in two experiment locations.
Trait | Hangzhou | Hainan | Geno | Env | H2 (%) |
| Mean ± sd | Range | CV (%) | Mean ± sd | Range | CV (%) | MS | sig | MS | sig | |
FLL (cm) | 36.93 ± 6.25 | 23.5-59.07 | 16.92 | 39.86 ± 5.67 | 27.04–55.95 | 14.22 | 17.388 | *** | 4.052 | *** | 44 |
FLW (cm) | 1.66 ± 0.19 | 1.24–2.27 | 11.44 | 1.76 ± 0.20 | 1.14–2.36 | 11.36 | 0.02493 | *** | 0.00416 | *** | 58 |
FLR (%) | 10 ± 10 | 0–47 | 100 | 21 ± 13 | 4–71 | 61.9 | 7817 | *** | 6255 | *** | 40 |
FLL = flag leaf length; FLW = flag leaf width; FLR = flag leaf rolling; CV = coefficient of variation; H2 = heritability. *, ** and ***indicate the least significant difference at 0.05, 0.01 and 0.001 probability leve. |
We determined the genetic and environmental effects on leaf-related traits in the population by variance analysis. Genetic and environmental effects were significantly different for all traits. We calculated the additive genetic variance for RILs, environmental variance, and heritability of three traits assessed in our artificial genetic population. Among these three leaf traits, flag leaf width had the highest heritability (58%), and flag leaf rolling had the lowest (40%).
Correlation of leaf and yield traits
The genetic and phenotypic correlations between the four traits are presented in Table 3. The results revealed that high positive correlations were detected between FLW and PY (0.692). Furthermore, FLL was positively correlated with FLR (0.352) significantly. A similar trend was consistent with the Pearson correlation coefficients (Table 4). The results suggested a weak trend that relatively longer and wider flag leaves may exhibit more flag leaf roller and higher yield per plant in the RIL population.
Table 3
Genetic and Phenotype Correlation between three flag leaf traits and yield per plant.
Genetic & phenotype correlation | FLL (cm) | FLW (cm) | FLR (%) | PY (g) |
FLL (cm) | | -0.129 | 0.352 *** | 0.141 |
FLW (cm) | 0.071 | | 0.004 | 0.692 *** |
FLR (%) | 0.064 | 0.070 | | 0.040 |
PY (g) | 0.067 | 0.060 | 0.067 | |
The up-triangle are genetic correlation; the low-triangle are phenotype correlation. *, ** and ***indicate the least significant difference at 0.05, 0.01 and 0.001 probability level. |
Table 4
Pearson correlation coefficients in HZ&HN between three flag leaf traits and yield per plant.
Traits correlation in HZ&HN | FLL (cm) | FLW (cm) | FLR (%) | PY (g) |
FLL (cm) | | 0.138 | 0.256 ** | 0.167 |
FLW (cm) | -0.072 | | 0.068 | 0.311 *** |
FLR (%) | 0.255 ** | 0.121 | | 0.055 |
PY (g) | 0.081 | 0.358 *** | 0.080 | |
The up-triangle are phenotype correlation coefficients for Hangzhou; the low-triangle are phenotypic correlation coefficients for Hainan. *, **, *** indicate the significant level of 0.05, 0.01 and 0.001. |
Comparison of GWAS implemented with QTXNetwork and EMMAX
There were 21 significant SNPs (8 SNPs for FLL, 5 SNPs for FLW, and 8 SNPs for FLR) shown in Table 5 by QTXNetwork with accumulated explanatory heritability ranging from 53.29–89.25%. Two additive-environment interaction effects (ae) were detected for FLL, and 2 epistatic effects (i/ie) were distinguished for FLL and FLW, and 19 additive and additive-environment interaction effects (a & ae) identified for three flag leaf traits (Table 5 and Fig. 3).
Table 5
The estimated heritability and predicted genetic effects of all detected significant SNP loci for three leaf shape traits by QTXNetworker.
Trait | QTS | Chr | Allele | Effect type | Effect size | −log10(P) | h 2 | h2T |
FLL | seq1_6645948 | 1 | A/G | a | 1.328 | 5.66 | 1.93% | 89.25% |
| seq3_27773262 | 3 | T/C | e1 | 0.909 | 1.94 | 0.91% | |
| | | | e2 | -0.881 | 1.84 | 0.85% | |
| seq3_30110703 | 3 | T/C | a | 2.385 | 18.76 | 6.23% | |
| | | | ae1 | 3.017 | 15.27 | 9.97% | |
| | | | ae2 | -3.129 | 16.36 | 10.72% | |
| seq3_30131604 | 3 | T/C | a | -3.745 | 43.83 | 15.35% | |
| | | | ae1 | -3.039 | 15.17 | 10.11% | |
| | | | ae2 | 3.313 | 17.87 | 12.02% | |
| seq5_15012722 | 5 | A/C | a | 1.586 | 7.12 | 2.75% | |
| seq6_771947 | 6 | G/A | a | -1.248 | 4.93 | 1.71% | |
| seq6_5306701 | 6 | T/A | e1 | 0.825 | 1.65 | 0.74% | |
| | | | e2 | -0.828 | 1.66 | 0.75% | |
| seq6_10897434 | 6 | G/T | a | 1.311 | 6.05 | 1.88% | |
| seq3_30110703 | 3 | T/C | aa | -3.978 | 49.30 | 34.64% | |
| seq3_30131604 | 3 | T/C | | | | | |
FLW | seq5_23349984 | 5 | C/G | a | -0.091 | 23.50 | 12.22% | 83.33% |
| | | | ae1 | 0.038 | 2.70 | 2.19% | |
| | | | ae2 | -0.037 | 2.59 | 2.08% | |
| seq5_23494718 | 5 | A/G | a | -0.034 | 3.82 | 1.73% | |
| | | | ae1 | 0.048 | 3.87 | 3.44% | |
| | | | ae2 | -0.050 | 4.11 | 3.68% | |
| seq6_8749143 | 6 | A/C | a | 0.024 | 1.85 | 0.84% | |
| seq6_26960605 | 6 | T/C | a | 0.075 | 12.76 | 8.46% | |
| seq8_18325619 | 8 | A/G | a | -0.183 | 91.09 | 49.68% | |
| seq5_23349984 | 5 | C/G | aa | -0.040 | 3.99 | 4.70% | |
| seq6_26960605 | 6 | T/C | aa | | | | |
FLR | seq1_5924014 | 1 | C/T | a | -2.150 | 3.67 | 6.01% | 53.29% |
| seq3_4147055 | 3 | G/A | a | -1.440 | 2.29 | 2.70% | |
| seq3_8786910 | 3 | C/T | a | -1.990 | 3.31 | 5.14% | |
| seq3_12015603 | 3 | G/A | a | -3.160 | 5.71 | 12.90% | |
| | | | ae1 | 2.140 | 1.75 | 5.91% | |
| | | | ae2 | -2.250 | 1.90 | 6.57% | |
| seq4_6532007 | 4 | C/T | a | -1.850 | 2.98 | 4.41% | |
| seq5_15014180 | 5 | C/T | a | 1.790 | 2.91 | 4.14% | |
| seq8_1618449 | 8 | G/A | a | 2.340 | 4.76 | 7.10% | |
| seq8_23600813 | 8 | C/T | a | -1.890 | 3.71 | 4.65% | |
QTS = the detected significant SNPs associated with the traits; Chr.= chromosome; Allele = paternal allele/maternal allele; a = additive effect for paternal allele homozygotes, e = environmental effect for paternal allele homozygotes ,ae = additive by environmental interaction effect, aa = additive by additive epistasis effect; −log10(P) = inverse of the base 10 logarithm of p value; h2 (%) = heritability in percentage due to the genetic component effect; h2T (%) = total heritability equal to summation of heritability of all individual QTSs. |
For FLL, 8 significant SNPs positioned on 4 chromosomes, accumulative taking up 89.25% of the heritability. A pair of additive-by-additive epistasis effects (seq3_30110703/ seq3_30131604) totally contributed to 56.22% of the heritability and showed opposite ae effects model in the two different locations. Additionally, there were two additive-environment interaction effects detected, and both of them showed negative additive effects in HN and positive additive effects in HZ.
For the FLW, there were 5 significant SNPs detected, and the SNP located on chromosome 8 (seq8_18325619) exhibited the largest main additive effects (h2 = 49.68%). Two SNPs located on chromosome 5 (seq5_23349984 and seq5_23494718) have relatively stronger pairwise LD (r2 = 0.92), which indicates these two SNPs likely represent one identical causal locus influencing the FLW. In addition, there was a couple of positive a&ae effect (seq5_23349984 / seq6_26960605) discovered; and their individual main additive demonstrated differing genetic effects (seq5_23349984 was negative although seq6_26960605 was positive).
For the FLR, the total heritability (53.29%) mainly consisted of additive heritability (40.08%) from 8 SNPs that had quite large individual heritabilities, especially for seq3_12015603 (h2 = 12.90%). The largest SNP also exhibited the a&e effect which showed reverse genetic effects in HN and HZ. Most of the remaining additive effects were negative with individual contributions to the phenotype variation (2.07% − 6.01%).
To validate the results from QTXNetwork, EMMAX was also carried out to analyze the data (Fig. 4). In total, 16 significant association peaks (2 SNPs for FLL, 11 SNPs for FLW, 3 SNPs for FLR) were identidied for three leaf phenotypes in HZ and HN by EMMAX (Table 6). For the FLL, the lead SNP, that p is 4.1×10− 6, recognized based on phenotype from HZ is seq3_30142073, and this SNP was close to the SNP (seq3_30131604) which was the largest additive heritability of FLL from QTXNetwork. For the FLW, a total of 11 associated peaks were detected on chromosomes 3, 4, 5, 6, 11, and 12 in both two environments. The lead SNP identified based on data from HN was seq6_8637474 with a P-value of 8.12×10− 9, and this SNP is close to SNP seq6_8749143 from QTXNetwork. For the FLR, the lead SNP identified based on data from HN was seq3_12363195 with a P-value of 1.99×10− 9, and this SNP was close to SNP seq3_12015603 which was the largest additive heritability of FLR from QTXNetwork. The results showed a consistency between the two analytical tools.
Table 6
The 16 QTLs detected by EMMAX for three leaf shape traits in HZ and HN.
QTL | Env | Chr | Alleles | Maf | Peak | −log10(P) | Effect | Candidate genes |
qFLL1 | HZ | 3 | G/A | 0.48 | seq3_30142073 | 5.39 | 11.664 | LOC_Os03g52540 |
qFLL2 | HN | 7 | G/A | 0.49 | seq7_16075487 | 5.63 | 2.495 | LOC_Os07g27510 |
qFLW1.1 | HZ | 4 | A/T | 0.47 | seq4_7067615 | 5.42 | 0.283 | LOC_Os04g12770 |
qFLW1.2 | HZ | 6 | G/A | 0.46 | seq6_27447966 | 5.46 | 0.083 | LOC_Os06g45390 |
qFLW2.1 | HN | 6 | G/A | 0.48 | seq6_8637474 | 9.09 | -0.289 | LOC_Os06g15240 |
qFLW2.2 | HN | 3 | C/T | 0.47 | seq3_20104630 | 6.30 | 0.360 | LOC_Os03g36239 |
qFLW2.3 | HN | 5 | A/G | 0.48 | seq5_24521657 | 6.20 | 0.590 | LOC_Os05g41870 |
qFLW2.4 | HN | 5 | G/A | 0.47 | seq5_26517897 | 6.18 | 0.311 | LOC_Os05g45779 |
qFLW2.5 | HN | 6 | T/A | 0.48 | seq6_14778432 | 6.20 | 0.590 | LOC_Os06g25270 |
qFLW2.6 | HN | 6 | C/T | 0.48 | seq6_20544872 | 6.20 | 0.590 | LOC_Os06g35220 |
qFLW2.7 | HN | 11 | T/C | 0.48 | seq11_13601041 | 6.25 | 0.502 | LOC_Os11g24000 |
qFLW2.8 | HN | 11 | T/C | 0.48 | seq11_26462918 | 5.92 | 0.371 | LOC_Os11g43800 |
qFLW2.9 | HN | 12 | A/G | 0.48 | seq12_23987153 | 6.20 | 0.590 | LOC_Os12g39000 |
qFLR1.1 | HZ | 5 | A/G | 0.49 | seq5_29686446 | 6.63 | 37.15 | LOC_Os05g51750 |
qFLR1.2 | HZ | 7 | T/G | 0.46 | seq7_362429 | 5.44 | 5.093 | LOC_Os07g01580 |
qFLR2 | HN | 3 | G/A | 0.46 | seq3_12363195 | 8.70 | -8.138 | LOC_Os03g21660 |
MAF = Minor allele frequency; Allele = paternal allele/maternal allele; −log10(P) = inverse of the base 10 logarithm of p value; Effect = Allele effect with respect to the minor allele. |
Bioinformatics analysis for candidate genes
According to the latest version of the MSU rice genome annotation project, 38 SNPs associated with the FLL, FLW, and FLR, 14 SNPs are located within annotated genes as presented in Table 7.
Table 7
The annotation information of genes harboring the significant SNPs of for three leaf shape traits.
Marker | DPE | Chr | Allele | Trait | Candidate genes | Gene Annotation |
seq1_5924014 | ZH9308 | 1 | C/T | FLR | LOC_Os01g11110 | 14-3-3/GF14 protein |
seq1_6645948 | XB | 1 | A/G | FLL | LOC_Os01g12180 | expressed protein |
seq3_4147055 | ZH9308 | 3 | G/A | FLR | LOC_Os03g08120 | retrotransposon protein, putative, unclassified, expressed |
seq3_8786910 | ZH9308 | 3 | C/T | FLR | LOC_Os03g15900 | SH3 domain containing protein, expressed |
seq3_12015603 | ZH9308 | 3 | G/A | FLR | LOC_Os03g21100 | transposon protein, putative, CACTA, En/Spm sub-class, expressed |
seq3_12363195 | ZH9308 | 3 | G/A | FLR | LOC_Os03g21660 | transposon protein, putative, unclassified, expressed |
seq3_20104630 | XB | 3 | C/T | FLW | LOC_Os03g36239 | retrotransposon protein, putative, Ty3-gypsy subclass, expressed |
seq3_27773262 | XB | 3 | T/C | FLL | LOC_Os03g48740 | auxin glucosyltransferase; IAA-glucose synthase; IAA-conjugating enzyme; auxin UDP-glycosyltransferase (UGT) gene |
seq3_30110703 | XB | 3 | T/C | FLL | LOC_Os03g52470 | WD domain, G-beta repeat domain containing protein, expressed |
seq3_30131604 | XB | 3 | T/C | FLL | LOC_Os03g52510 | retrotransposon protein, putative, unclassified, expressed |
seq3_30142073 | XB | 3 | G/A | FLL | LOC_Os03g52540 | retrotransposon protein, putative, Ty3-gypsy subclass, expressed |
seq4_6532007 | XB | 4 | C/T | FLR | LOC_Os04g11910 | transposon protein, putative, CACTA, En/Spm sub-class, expressed |
seq4_7067615 | XB | 4 | A/T | FLW | LOC_Os04g12770 | retrotransposon protein, putative, unclassified, expressed |
seq5_15012722 | XB | 5 | A/C | FLL | LOC_Os05g25800 | transposon protein, putative, unclassified, expressed |
seq5_15014180 | XB | 5 | C/T | FLR | LOC_Os05g25800 | transposon protein, putative, unclassified, expressed |
seq5_23349984 | XB | 5 | C/G | FLW | LOC_Os05g39750 | expressed protein |
seq5_23494718 | XB | 5 | A/G | FLW | LOC_Os05g39990 | expansin gene |
seq5_24521657 | XB | 5 | A/G | FLW | LOC_Os05g41870 | glycine-rich cell wall protein, putative, expressed |
seq5_26517897 | XB | 5 | G/A | FLW | LOC_Os05g45779 | expressed protein |
seq5_29686446 | ZH9308 | 5 | G/A | FLR | LOC_Os05g51750 | aspartyl protease family protein, putative, expressed |
seq6_771947 | XB | 6 | G/A | FLL | LOC_Os06g02320 | expressed protein |
seq6_5306701 | XB | 6 | T/A | FLL | LOC_Os06g10320 | expressed protein |
seq6_8637474 | ZH9308 | 6 | G/A | FLW | LOC_Os06g15240 | expressed protein |
seq6_8749143 | XB | 6 | A/C | FLW | LOC_Os06g15400 | expressed protein |
seq6_10897434 | ZH9308 | 6 | G/T | FLL | LOC_Os06g19150 | retrotransposon protein, putative, unclassified, expressed |
seq6_14778432 | XB | 6 | T/A | FLW | LOC_Os06g25270 | retrotransposon protein, putative, Ty1-copia subclass, expressed |
seq6_20544872 | ZH9308 | 6 | C/T | FLW | LOC_Os06g35220 | retrotransposon protein, putative, Ty3-gypsy subclass, expressed |
seq6_26960605 | XB | 6 | T/C | FLW | LOC_Os06g44660 | fasciclin-like arabinogalactan precuseqor protein, putative, expressed |
seq6_27447966 | XB | 6 | G/A | FLW | LOC_Os06g45390 | expressed protein |
seq7_362429 | XB | 7 | T/G | FLW | LOC_Os07g27510 | retrotransposon protein, putative, Ty1-copia subclass, expressed |
seq7_16075487 | XB | 7 | G/A | FLL | LOC_Os07g01580 | expressed protein |
seq8_1618449 | XB | 8 | G/A | FLR | LOC_Os08g03450 | ribosomal protein L37, putative, expressed |
seq8_18325619 | XB | 8 | A/G | FLW | LOC_Os08g29809 | resistance protein LR10, putative, expressed |
seq8_23600813 | XB | 8 | C/T | FLR | LOC_Os08g37350 | ubiquitin carboxyl-terminal hydrolase family protein, expressed |
seq11_13601041 | XB | 11 | T/C | FLW | LOC_Os11g24000 | retrotransposon protein, putative, Ty3-gypsy subclass, expressed |
seq11_26462918 | XB | 11 | T/C | FLW | LOC_Os11g43800 | PPR repeat domain containing protein, putative, expressed |
seq12_23987153 | XB | 12 | A/G | FLW | LOC_Os12g39000 | hypothetical protein |
DPE = Direction of phenotypic effect; Allele = paternal allele/maternal allele; Note: Gene annotation information comes from the database: MSU Rice Genome Annotation Project Release 7. |
Seven of these were classified as genes encoding expressed or hypothetical proteins. The other were found to encode particular domains, enzymes, retrotransposon proteins and transcription factors, which could play vital functional parts in organism. Such as, SNP seq3_27773262 has different effects on the length of the flag leaf in different environments, and it is 2 kb away from LOC_Os03g48740. In previous studies (Liu et al. 2019; Choi et al. 2012), LOC_Os03g48740 (indole-3-acetate beta-glucosyltransferase, OsIAGT) has the effect of regulating the auxin content of plants and crop morphology, and the encoded OsIAGT has the highest activity at 30°C. These results indicate that this gene may have a positive regulatory effect on the length of the flag leaf under high-temperature conditions. LOC_Os01g11110 encodes 14-3-3/GF14 proteins contained the seq1_5924014 which dectected in FLR. The protein function as chief regulators of leaf development, metabolism formation in a variety of plants (Lee et al. 2020; Chen et al. 2006).
Expression analysis of candidate genes
In order to verify whether the associated SNPs have an impact on leaf phenotype, we selected 13 candidate genes and 6 known marker genes (Wang et al. 2007; Nakamura et al. 2006; Xiang et al. 2012; Zhang et al. 2015; Huang et al. 2016b; Zhang et al. 2016; Ma et al. 2017; Shao et al. 2019; Liu et al. 2019; Zhao et al. 2016) that regulate leaf development to compare their expression levels between two parents. As shown in Fig. 5, a total of 18 genes had a significant difference between the two parents. The relative expression level of LOC_Os03g52510 that retrotransposon protein was 11000 times, the highest among these. And it contains seq3_30131604 which had the greatest additive effect on FLL. LOC_Os12g39000 is a hypothetical protein, and 2kb away from the site of seq12_23987153 is the lowest relative expression level (0.5) in ZH9308. This result indicates that this locus has a positive regulation of FLW. Among the 6 genes that have been reported to regulate rice leaf morphology, 5 genes were differentially expressed in the parents. Among them, LOC_Os03g19520, LOC_Os06g47150, and LOC_Os12g40900 were three auxin-responsive genes, which can regulate rice leaf rolling and length (Wang et al. 2007; Nakamura et al. 2006; Huang et al. 2012; Liu et al. 2019), the expression of these genes increases with the increase of IAA concentration, which in turn regulates leaf morphology. This result implies that more IAA in ZH9308 caused the difference in leaf morphology between the two parents. LOC_Os07g01240 is related to cell wall development and controls the width and rolling of rice flag leaves (Xiang et al. 2012), it has a high expression level in ZH9308 and is curlier, which is consistent with previous reports. LOC_Os03g21090 is an auxin-independent growth promoter protein which was 3kb away seq3_12015603 and 1kb away LOC_Os03g21100, its expression level is 2.88 times that of XB in ZH9308. Compared with XB, the expressions of 11 candidate genes increased significantly, and 2 decreased significantly in ZH9308. LOC_Os01g12180, LOC_Os03g52470, LOC_Os03g52510, LOC_Os05g25800, LOC_Os06g02320, and LOC_Os07g01580 were identified for FLL. LOC_Os03g52470 was a candidate gene for the main QTL, and its expression in the ZH9308 was highly significantly greater than that in XB (Table 7). In addition, the additive effect of this position is 2.385, which may positively regulate FLL. LOC_Os03g36239, LOC_Os06g45390, LOC_Os08g29809, and LOC_Os12g39000 were candidate genes of FLW. The seq8_18325619 was a major QTL for FLW located in the LOC_Os08g29809, has a function of response to stress (Jiang et al. 2020). The effect of this QTL was − 0.183, and the expression level in ZH9308 was higher, suggesting that the gene negatively regulates FLW. LOC_Os01g11110, LOC_Os03g21660, and LOC_Os08g37350 were found for FLR. The candidate gene of major QTL for FLR was LOC_Os03g21660, which is a transposon protein. Its allele comes from ZH9308, and its expression in the ZH9308 is significantly lower than that in XB. And the site had a negative effect, demonstrating that the site may be negatively regulating FLR. The experiment of gene expressions confirmed that there were differential expressions between two parents in the identified variants.