Allelic combinations of Hd1, Hd16, and Ghd7 exhibit pleiotropic effects on agronomic traits in rice

Abstract Heading date is a critical agronomic trait that significantly affects grain yield and quality in rice. As early heading is typically associated with reduced yield due to shorter growth duration, it is essential to harness optimum heading date genes and their allelic combinations to promote heading while minimizing yield penalties. In this study, we identified quantitative trait loci (QTLs) for heading date and other major agronomic traits in a recombinant inbred line (RIL) population derived from a cross between Koshihikari and Baegilmi. Analyses on 3 major QTLs for heading date and their underlying genes (Hd1, Hd16, and Ghd7) revealed their pleiotropic effects on culm length, panicle length, and head rice percentage. Additionally, Ghd7 exhibited pleiotropic effects on panicle number and grain size. Among 8 different types of allelic combinations of the 3 heading date genes, RILs carrying a single nonfunctional hd16 or ghd7 under the functional background of the other 2 genes (Hd1hd16Ghd7 and Hd1Hd16ghd7) showed potential for maintaining yield and quality-related traits while accelerating heading. These results provide valuable insights for fine-tuning heading dates in rice breeding programs.


Introduction
Flowering time, also known as the heading date, is a critical agronomic trait determining crop productivity (Izawa 2007a;Cho et al. 2017).Environmental factors such as day length and temperature exert a significant influence on the heading date, interacting with inherent characteristics such as photoperiod sensitivity, temperature sensitivity, and vegetative growth duration (Shang et al. 2012;Shrestha et al. 2014;Itoh et al. 2018).Either extremely early or overly late flowering can result in reduced grain yield attributed to insufficient sink strength or the susceptibility to adverse weather conditions during grain filling (Jung and Müller 2009;Gao et al. 2014;Li et al. 2018).Therefore, to optimize grain yield, it is essential to develop varieties with heading dates tailored for specific target environments.This optimized timing enables the maximization of available light and temperature resources within their respective cultivating regions, ensuring successful production.
To finely tune heading dates, it is crucial to understand and utilize the genetic mechanisms underlying heading date variations.In rice, numerous quantitative trait loci (QTLs) associated with heading date have been identified through genetic approaches (Yamamoto et al. 2012;Hori et al. 2016;Yano et al. 2016;Matsubara and Yano 2018).Furthermore, molecular investigations into the genes influencing heading date have unveiled their pleiotropic effects on various agronomic traits.For instance, Heading date 1 (Hd1) (Yano et al. 2000) is a component of OsGI-Hd1-Hd3a (rice GIGANTEA, Heading date 1, and Heading date 3a) signaling pathway.This pathway exhibits evolutionary conservation analogous to the GI-CO-FT (GIGANTEA, CONSTANS, and FLOWERING LOCUS T) pathway of Arabidopsis.Within this regulatory network, the upregulation of Hd1, mediated by OsGI, triggers the activation of Hd3a expression, thus inducing rice heading under both short-day and long-day conditions.The combination of Hd1 and Early heading date 1 (Ehd1) regulates panicle development, potentially impacting crop yield (Endo-Higashi and Izawa 2011).Additionally, delayed heading under long-day conditions influenced by Hd1 and Grain number, plant height and heading date 7 (Ghd7) interaction, contributes to enhanced plant height and higher grain yield (Yano et al. 2001;Hayama et al. 2003;Searle and Coupland 2004;Nemoto et al. 2016;Zhang et al. 2017).The other pathway involves the rice-specific gene Ghd7, which encodes CO, CO-LIKE, and TOC1 (CCT) domain proteins and has been identified as a flowering repressor under long-day conditions due to its inhibitory effect on Hd3a (Xue et al. 2008).Moreover, Ghd7 hinders the expression of Ehd1, an upstream regulator of Hd3a (Doi et al. 2004) and plays a pivotal role in regulating plant height, heading date, and yield (Xue et al. 2008).The Heading date 16 (Hd16) encodes casein kinase I, a key regulator in the photoperiodic flowering pathway of rice.Hd16 exerts its influence by initiating the activation of Ghd7 through phosphorylation.Within this framework, a missense mutation occurring in exon 10 leads to a reduction in kinase activity, resulting in the acceleration of heading under long-day conditions (Hori et al. 2013).
Previously, we performed QTL analysis on rice heading date using recombinant inbred lines (RILs) derived from 2 rice cultivars, Koshihikari and Baegilmi, and revealed that Hd1, Hd16, and Ghd7 underlie the 3 major heading date QTLs (Mo et al. 2020).In the present study, we used the same RIL population to conduct QTL analysis on major agronomic traits and characterize the https://doi.org/10.1093/g3journal/jkad300Advance Access Publication Date: 3 January 2024

Plant Genetics and Genomics
pleiotropic effects and interactions of Hd1, Hd16, and Ghd7.By analyzing the RILs with different allelic combinations of these 3 genes, we also defined potentially advantageous allelic combinations that could accelerate heading while maintaining grain yield and quality-related traits, providing valuable insights for rice breeding programs.

Plant materials and field test conditions
A total of 339 RILs derived from a cross between 2 japonica rice varieties, Koshihikari and Baegilmi, were used in this study.For QTL analysis, a subset (n = 142) of the RILs was used while the whole population (n = 339) were utilized to investigate the effects of 3 heading date genes, Hd1, Hd16, and Ghd7, on agronomic traits (Supplementary Tables 1-3).All plant materials were cultivated in the experimental field of the National Institute of Crop Science (NICS), Wanju, South Korea (35°84′N 127°05′E).Field tests (FTs) were conducted from May to October in 2016 (F 6 ; FT 1), 2017 (F 7 ; FT 2), and 2018 (F 8 ; FT 3).Climate conditions during the FTs are shown in Supplementary Table 4.The seedlings were transplanted into the paddy field 4 weeks after sowing under the following conditions: 1 RIL per row, 30 plants per row, and 1 plant per hill, with a spacing of 15 cm between plants in a row and 30 cm between adjacent rows.Fertilizers were applied uniformly across all FTs, with a dosage of 9 kg N/10a, 4.5 kg P 2 O 5 /10a, and 4.5 kg K 2 O/10a.The plants were grown and managed according to the standard rice cultivation methods of NICS, Rural Development Administration (RDA) (RDA 2012).

Phenotype evaluation
Days to heading (DTH) of each RIL were determined by counting the number of days from sowing to heading when the panicles of 40% of plants emerged in a row.Agronomic traits were collected from 10 randomly chosen plants in each line across 3 FTs.Culm length (CL) was measured as the length from the ground level to the panicle node of the main culm.Panicle length (PL) was determined by measuring the length from the panicle node to the end of the main panicle.The panicle number per plant (PN) was calculated by counting the total number of panicles per plant.The grain characteristics of brown rice were assessed only during FT2.Each RIL was harvested 50-55 days after heading upon maturity, dried, and maintained at 15% moisture content in a storage room at 10°C.After the entire RIL population was harvested, the samples were dehulled to measure the grain characteristics of brown rice.The head rice percentage (HRP) was determined using RN300 Rice Quality Analyzer (Kett, Japan) with the "Approval" mode of Quality Scan software for brown rice, which classified the grain based on 3 criteria: even, cracked, and others (immature, dead, and discolored).The percentage of grains classified as "even" over total grains was calculated to generate HRP.Thousand-grain weight (TGW) was measured by determining the weight of 1,000 dehulled grains.Grain length (GL) and grain width (GW) of each RIL were measured using 10 randomly selected fully filled grains with a Vernier caliper.

QTL analysis and allele determination of heading date genes
Genomic DNA was extracted from the young leaves of each RIL using the modified cetyltrimethylammonium bromide method (Murray and Thompson 1980).The concentration and quality of each DNA sample were assessed using the DeNovix DS-11 spectrophotometer (DeNovix, USA).A total of 128 high-quality polymorphic SNPs from the 142 RILs (missing rate < 10%) previously described in Mo et al. (2020) were used for linkage map construction.The QTL analysis was conducted using QTL ICIMapping 4.2 (Meng et al. 2015).The Kosambi mapping function was used to calculate recombination distance (Kosambi 1944).The inclusive composite interval mapping of additive (ICIM-ADD) method was used to detect QTLs with the default mapping parameters (step cM as 1.0 and PIN as 0.001).The logarithm of the odds (LOD) threshold for declaring significant QTLs ranged from 2.6 to 3.1 according to 1,000 permutation tests at P = 0.05.Phenotypic variance explained (PVE) and additive effect (ADD) estimates of each QTL position were obtained from the ICIM-ADD analysis.
Molecular markers used for genotyping Hd1, Hd16, and Ghd7 in the 339 RILs are described in Mo et al. (2020).Each PCR reaction comprised 20 ng of template DNA, 5 pmol of each primer, 10 mM of dNTPs, 10× PCR buffer, and 0.5 U of Taq DNA polymerase (Inclone).For Hd1 and Ghd7 genotyping, DNA was initially incubated for 5 min at 94°C followed by 35 cycles of amplification: denaturation at 94°C for 30 s, annealing at the proper melting temperature (58°C for Hd1 and 60°C for Ghd7) for 30 s, and extension at 72°C for the relevant duration (30 s for Hd1 and 240 s for Ghd7) followed by final extension at 72°C for 10 min.For Hd16, PCR was conducted by a combined procedure of annealing and extension steps at 68°C for 60 min, followed by the NheI digestion of the PCR product.Detailed primer sequences and expected band patterns are shown in Supplementary Table 5.

Statistical analysis
All statistical analyses were performed using R version 4.2.1 R Core Team (2022).The frequency distribution and correlation plots of the agronomic traits among different FTs were analyzed using "ggpubr" package.To examine the interactions and allelic combination effects among Hd1, Hd16, and Ghd7, we utilized the mean values of agronomic traits across the 3 FTs.Boxplots visualizing the agronomic traits and grain characteristics of different allelic combinations were created using "ggplot2", "gridExtra", "gapminder", and "dplyr" packages.The mean comparison of traits among different allelic combinations was conducted by Duncan's multiple range test using "agricolae" package (Mendiburu and Yaseen 2020).To verify the existence of highorder genetic interactions, a 3-way factorial ANOVA was performed to determine the main effects of 3 heading date genes using "rstatix", "lme4", "lmertest", and "car" packages.To visualize the simple effects of the heading date genes, 2-way and 3-way interaction plots were generated using "dplyr" package.Correlation analyses and their visualization were conducted using Pearson's correlation coefficient (P ≤ 0.05) via the "corrplot" package.

Variation of major agronomic traits in the Koshihikari/Baegilmi RIL population
The phenotypic frequency distribution of 8 agronomic traits among the 339 RILs derived from a cross between Koshihikari and Baegilmi are shown in Fig. 1a and b.The traits CL, PL, PN, HRP, TGW, GL, and GW were normally distributed, whereas DTH showed a rightly skewed distribution.In the 3 FTs, the DTH of Koshihikari ranged from 86 to 91 days, whereas those of Baegilmi ranged from 67 to 78 days.The DTH values of the 339 RILs varied from 64 to 113 days across the 3 FTs.In addition, the RIL population showed large phenotypic variation in other agronomic traits, with ranges of 51.3-114 cm for CL, 16.5-27.3cm for PL, 4.5-20.3for PN, 21.6-89.7%for HRP, 14.5-25.0g for TGW, 4.7-5.6 mm for GL, and 2.6-3.1 mm for GW.Transgressive segregations were observed in all traits, as a number of RILs displayed agronomic trait values higher or lower than both parents.Strong positive correlations were observed among the DTH values of the RILs for all 3 FTs, with correlation coefficients of 0.86-0.95(Fig. 1c).Additionally, CL and PL showed significant correlations among the FTs, whereas PN exhibited significant correlations only between FT1 and FT3.

QTL analysis of the agronomic traits of RILs
A total of 128 SNPs from a subset (n = 142) of the RILs were used to construct a linkage map encompassing 1,293 cM composed of an average of 10.7 markers per chromosome, with an average genetic distance of 11.1 cM between adjacent markers (Mo et al. 2020).Three significant QTLs were detected for DTH in 3 FTs (Fig. 2 and Table 1).These 3 stable QTLs harbored previously identified heading date genes, Hd1 (qDTH6), Hd16 (qDTH3), and Ghd7 (qDTH7) (Mo et al. 2020).qDTH6 colocalized with qCL6 in 2 FTs and qHRP6 in 1 FT.qDTH7 colocalized with qCL7 in 3 FTs, qPL7 in 1 FT, and qHRP7 in 1 FT.These results confirmed the previously well-known pleiotropic effects of Hd1 and Ghd7 on multiple agronomic traits (Xue et al. 2008;Zhang et al. 2012;Nemoto et al. 2016).

Effects of Hd1, Hd16, and Ghd7 on major agronomic traits
Previous studies confirmed that Koshihikari carries a G-to-A (alanine-to-threonine) mutation of Hd16 responsible for early heading under long day (hereafter hd16) and functional alleles of Hd1 and Ghd7, while Baegilmi carries nonfunctional Hd1 due to a 43-bp deletion in exon 2 (hereafter hd1) and a 1901-bp retrotransposon insertion in the promoter region of Ghd7 (hereafter ghd7) with Nipponbare-type functional Hd16 (Yano et al. 2000;Hori et al. 2013;Fujino and Yamanouchi 2020;Mo et al. 2020).Using a larger number (n = 339) of RILs, we characterized the main effects of the 3 heading date genes (Hd1, Hd16, and Ghd7) and their 2-way and 3-way interactions on major agronomic traits (Table 2, Fig. 3).

Days to heading
The main effects of Hd1, Hd16, and Ghd7 were highly significant, explaining 23.9, 17.9, and 26.0% of the DTH variation, respectively (Table 1).On average, RILs carrying ghd7 (Baegilmi allele) headed 9.0 days earlier than those carrying Ghd7 (Koshihikari allele).Similarly, hd1 (Baegilmi allele) and hd16 (Koshihikari allele) accelerated heading by 8.3 and 8.2 days, respectively.The 2-way and 3-way interactions among Hd1, Hd16, and Ghd7 were highly significant (P < 0.001) except for the Hd16×Ghd7 interaction (Fig. 3a).In both the Hd1×Hd16 and the Hd1×Ghd7 interactions, the acceleration of DTH by the nonfunctional allele of 1 gene exhibited a greater effect when coexisting with the functional allele of another gene.For example, hd1 accelerated heading by 13.0 days and 3.6 days under the background of Hd16 and hd16, respectively.Similarly, hd1 allele accelerated heading by 11.3 days and 5.5 days when combined with Ghd7 and ghd7, respectively.The 3-way (Hd1×Hd16×Ghd7) interaction indicated that the effect of the

Culm length
Significant differences in CL (P < 0.001) were observed for all 3 heading date genes, with the nonfunctional alleles being associated with shorter CL.Ghd7 explained the largest proportion (29.9%) of CL variation, exhibiting a 10.3 cm difference between the RILs carrying the functional and nonfunctional alleles, followed by Hd1 (22.7%, 8.3 cm) and Hd16 (6.7%, 5.6 cm).However, only the Hd1×Hd16 interaction was significant (Fig. 3b), with hd1 decreasing CL by 12.6 and 3.8 cm under the background of Hd16 and hd16, respectively.

Panicle length
The main effects of the 3 heading date genes and their interaction effects were highly significant (P < 0.001), accounting for 24.7% of the total PL variation.Similar to CL, the nonfunctional allele of each gene was associated with shorter PL, with Ghd7 being the largest contributor (9.3%, 0.6 cm), followed by Hd16 (4.5%, 0.5 cm) and Hd1 (3.1%, 0.2 cm).Interestingly, the patterns of the 2-way and 3-way interactions on PL were different from those on DTH and CL.For example, the Hd1×Hd16 interaction showed that hd1 decreased PL by 0.5 cm under the hd16 background while exhibiting no significant effect under the Hd16 background (Fig. 3c).Similarly, the Hd1×Ghd7 and the Hd16×Ghd7 interactions also indicated that the effect of the nonfunctional allele of 1 gene on decreasing PL is significant only under the presence of the nonfunctional allele of another gene.The 3-way (Hd1×Hd16×Ghd7) interaction showed that hd1 increased PL under the presence of both Hd16 and Ghd7, while decreasing PL under the presence of either hd16 or ghd7.

Panicle number
Only the main effect of Ghd7 was marginally significant (P < 0.05), explaining 1.6% of the PN variation.There was no significant 2-way or 3-way interaction of the 3 heading date genes on PN.Fig. 2. QTL mapping for agronomic traits from the Koshihikari/Baegilimi RILs.The horizontal dashed line denotes the LOD score of 3.0 (LOD thresholds ranged from 2.6 to 3.1 according to 1,000 permutation tests at P = 0.05).Shaded regions indicate the locations of previously detected heading date genes, i.e.Hd16, Hd1, and Ghd7.

Head rice percentage
The main effects and interactions of the 3 heading date genes accounted for 36.2% of the HRP variation, with the nonfunctional allele of each gene associated with lower HRP.On average, the RILs carrying hd1, hd16, and ghd7 exhibited 8.6, 5.9, and 8.7% lower HRP compared to those carrying Hd1, Hd16, and Ghd7, respectively.Notably, all 2-way and 3-way interactions except the Hd1×Hd16 interaction were significant (Fig. 3d).The hd1 allele decreased HRP by 6.7 and 13.1% under the background of Ghd7 and ghd7, respectively.Similarly, ghd7 allele decreased HRP by 5.6 and 11.6% when combined with Hd16 and hd16, respectively.Interestingly, the 3-way interaction revealed that the effect of the nonfunctional allele of 1 gene decreasing HRP can be mitigated under the presence of the functional alleles of the other 2 genes.

Thousand-grain weight and grain length
Only the main effect of Ghd7 was significant, with the nonfunctional ghd7 allele associated with reduced TGW and GL.There were no significant 2-way or 3-way interactions of the 3 heading date genes on TGW and GL.

Grain width
The 3 heading date genes and their interaction exhibited minimal effects on GW, with only the 3-way interaction being marginally significant (P < 0.05) explaining 1.5% of the GW variation (Fig. 3e).

Effects of the allelic combinations of Hd1, Hd16, and Ghd7 on major agronomic traits
To evaluate the effects of pyramiding nonfunctional hd1, hd16, and ghd7 alleles on major agronomic traits, we compared the agronomic traits of the RILs carrying 8 different allelic combinations of the 3 genes (Fig. 4a).Compared to the RILs carrying Hd1Hd16Ghd7 (FFF; F refers to a functional allele), those with a single nonfunctional allele, i.e. hd1Hd16Ghd7 (NFF; N refers to a nonfunctional allele), Hd1hd16Ghd7 (FNF), and Hd1Hd16ghd7 (FFN), exhibited significantly earlier DTH and shorter CL.The RILs with the NFF combination exhibited greater reduction in DTH and CL compared to those with the FNF and FFN combinations, indicating a stronger effect of hd1 compared to hd16 and ghd7.However, there was no significant difference in PL, PN, TGW, GL, and GW among the RILs carrying the NFF, FNF, and FFN combinations.When 2 nonfunctional alleles were pyramided, i.e. hd1hd16Ghd7 (NNF), hd1Hd16ghd7 (NFN), and Hd1hd16ghd7 (FNN), substantial reductions were observed in DTH, CL, PL, and HRP compared to the impact of a single nonfunctional allele.Furthermore, pyramiding the 3 nonfunctional alleles (NNN) showed the most significant decreases in the aforementioned traits.
Next, we focused on PL and HRP to test the possibility of identifying optimal allele combinations with reduced growth duration while concurrently preserving agronomic traits affecting yield and grain quality (Fig. 4b).PL increased with an increase in DTH, but reached a limit at 88 DTH.Notably, the allele combinations of NFF, FNF, and FFN exhibited potential in maintaining PL with earlier DTH compared to the FFF combination.Similar to PL, HRP reached a plateau at approximately 88 DTH, suggesting The number after the "S" indicates the chromosome number and physical position of SNP according to the IRGSP-1.0 reference genome.
c Logarithm of the odds.LOD thresholds ranged from 2.6 to 3.1 according to 1,000 permutation tests at P = 0.05.Pleiotropic effects of rice heading date genes | 5 that the FNF and FFN allele combinations may be suitable for breeding early heading rice varieties without sacrificing grain appearance.Taken together, our analyses indicate that the strategic incorporation of hd16 (FNF) or ghd7 (FFN) allele for breeding early maturing rice varieties may effectively reduce the growth duration without adversely affecting yield and grain quality traits such as PL and HRP.

Pleiotropic effects of heading date genes on grain yield and quality
Understanding the pleiotropic effects of heading date genes on major agronomic traits is crucial for breeding rice varieties with optimal maturity and productivity for target environments.In rice, several key genes controlling heading date have been discovered to exert significant influence on yield-related traits.Notably, Hd1, Hd16.and Ghd7 have been shown to exhibit pleiotropic effects on heading date, plant height, panicle development, and yield potential (Xue et al. 2008;Hori et al. 2012;Zhang et al. 2012;Hori et al. 2013;Weng et al. 2014;Chen et al. 2023).In this study, we detected pleiotropic QTLs colocalized with Hd1, Hd16, and Ghd7 (Fig. 2 and Table 1), and further validated that the 3 loci exhibit significant main effects on DTH, CL, PL, and HRP (Table 2).Previous literature suggests that the pleiotropic effects of the heading date genes on CL and PL are likely genetic, given their functional involvement in internode elongation or panicle development (Endo-Higashi and Izawa 2011).However, the effects of these genes on HRP is less clear with the following possibilities to be further investigated: (1) The influence of heading date genes on HRP can potentially be attributed to "environmentally induced pleiotropy" (Auge et al. 2019).In temperate regions, earlier heading exposes rice plants to hot summer weather during grain filling, which typically results in deteriorated HRP (Chun et al. 2009;Kim et al. 2011;Tanamachi et al. 2016).
(2) Alternatively, heading date genes might exhibit genetic pleiotropy on HRP as well as CL and PL, potentially through their involvement in starch accumulation during grain filling.(3) Additionally, it is plausible that there are other genes affecting HRP (and other agronomic traits) in close proximity to the heading date genes, which might have been missed due to tight linkage or the limited number of molecular markers and RILs used in our present study.Further exploration is necessary to comprehensively assess the genetic mechanisms underlying HRP and its correlation with DTH.

Varying effects of three-gene combinations in regulating agronomic traits
Previous studies have been conducted to elucidate the genetic effects of different heading date genes and their combinations on major agronomic traits: Hd1/Ghd7/OsPRR37 (Fujino et al. 2019), Table 2. Three-way ANOVA of Hd1, Hd16, and Ghd7 on the agronomic traits.Hd1/Ghd7 (Zhang, Zhu, Wang, et al. 2019), Ghd7/Ghd8/OsPRR37/ Hd1 (Zhang, Liu, Qi, et al. 2019), Hd1/Ghd7/DTH8 (Zong et al. 2021;Sun et al. 2022), Hd1/DTH8 (Fujino 2020), and Ghd7/Ghd8/Hd1 (Zhang et al. 2015).In the present study, we focused on the main and combined effects of Hd1, Hd16, and Ghd7, the major photoperiod-sensitive genes that played significant roles in rice adaptation to diverse environments (Yano et al. 2000;Xue et al. 2008;Hori et al. 2013).Our study conducted under the natural long-day environments, revealed that nonfunctional hd1 (NFF), hd16 (FNF), and ghd7 (FFN) could independently promote heading while concurrently decreasing CL, PL, and HRP with varying degrees.Furthermore, pyramiding 2 or more alleles for early heading (NNF, NFN, FNN, and NNN) could significantly affect DTH, CL, PL, and HRP.These results are in agreement with the previous findings that nonfunctional hd1, ghd7, and hd16 alleles accelerate heading under long days and play important roles in rice adaptation to high latitude regions (Yano et al. 2000;Yano et al. 2001;Xue et al. 2008;Hori et al. 2013;Zhang et al. 2015).The present work also emphasizes the importance of understanding the interaction and integration of different heading date genes and elucidating their impacts on key agronomic traits.This understanding holds paramount importance in breeding early maturing varieties, ensuring the concurrent maintenance of grain yield and quality.

Optimizing heading date to maximize yield
Heading date optimization represents a crucial objective in rice breeding as it aims to enhance land use efficiency by facilitating diverse double cropping systems involving rice and alternative crops.Additionally, this optimization is integral for ensuring consistent yield and superior grain quality in specific target environments (Takahashi et al. 2009;Fujino et al. 2010;Ebana et al. 2011).A particular focus lies in the development of early heading rice cultivars, as they play a pivotal role in regions with high latitudes, where early heading characteristics are necessary for rice plants to complete grain filling before the onset of cold winter conditions (Izawa 2007b;Shrestha et al. 2014).These early heading rice cultivars prove invaluable in confronting the challenges posed by climate change, allowing for better adaptability to extreme weather events and aiding in the reduction of methane emissions from rice paddies by minimizing the duration of rice cultivation (Lee et al. 2012;Korea 2020;Lee et al. 2023).
To emphasize, a longer growth duration due to delayed heading days does not always result in a higher yield (Zhang, Zhu, Wang, et al. 2019).When compared to the FFF group, which exhibited 22.1 cm PL, the NFF, FNF, and FFN groups with earlier heading exhibited PL of 22.7 cm, 22.5 cm, and 22.4 cm, respectively (Fig. 4).As for HRP, the FFF group showed 66.7%, while the FNF and FFN groups exhibited 65.2 and 67.8%, respectively.In the RIL population, these groups can be considered as the optimum allelic combinations maintaining yield and quality-related traits with reduced DTH.Our previous work showed that 111 out of 293 Korean commercial rice cultivars carry nonfunctional hd1 alleles (Mo et al. 2021).However, only 7 cultivars possess Koshihikari-type hd16, and only 1 cultivar carries Sorachi-type ghd7 (Mo et al. 2020).Since deploying a single allele of Koshihikari-type hd16 or Sorachi-type ghd7 presents a possibility of accelerating heading without adversely affecting PL and HRP, further development of early maturing cultivars in Korean rice breeding programs will facilitate the use of these alleles for finetuning flowering time without deteriorating yield and qualityrelated traits.Considering the varying allelic effects of heading date genes influenced by environmental conditions and genetic backgrounds (Zhang et al. 2017;Kim et al. 2018;Fujino et al. 2019;Zong et al. 2021), it is important to emphasize that optimal allelic combinations of heading date genes should be customized for distinct target environments and markets.

d
Phenotypic variation explained.ePositive additive effect (Add) indicates that Koshihikari allele contributes positively to the trait.

Fig. 4 .
Fig. 4. Effects of the allelic combinations of Hd1, Hd16, and Ghd7 on major agronomic traits.a) Combined boxplots of each agronomic trait according to the 8 allelic combinations.The different letters above the boxplots denote significance among allele combinations at P < 0.05 determined by Duncan's multiple range test.F and N indicate functional and nonfunctional allele, respectively.Bold letter indicates allele associated with early heading.b) The relationship among DTH, PL, and HRP.Error bar and dashed line represent standard deviation and LOESS (locally estimated scatterplot smoothing), respectively.
a Chromosome number.b