Weitao Li,Nian Liu,Li Huang,Yuning Chen,Jianbin Guo,Bolun Yu,Huaiyong Luo,Xiaojing Zhou,Dongxin Huai,Weigang Chen,Liying Yan,Xin Wang,Yong Lei,Boshou Liao,Huifang Jiang*

Key Laboratory of Biology and Genetic Improvement of Oil Crops,Ministry of Agriculture and Rural Affairs,Oil Crops Research Institute of the Chinese Academy of Agricultural Sciences,Wuhan 430062,Hubei,China

Keywords:Peanut Shelling percentage QTL mapping Genomic region Candidate genes InDel marker

ABSTRACT Peanut is a major oilseed and food legume.Shelling percentage(SP),closely associated with seed yield,is a trait whose improvement is a major goal of peanut breeding.In this study,a mapping population(Xuhua 13 × Zhonghua 6) was used to map quantitative trait loci (QTL) controlling SP in four environments.Two stable major QTL for SP were mapped on both SSR-and SNP-based genetic maps.qSPA07.1 on chromosome A07 explained up to 31.7% of phenotypic variation,and qSPA08.2 on chromosome A08 explained up to 10.8%.Favorable alleles of qSPA07.1 and qSPA08.2 were derived from the female and male parents,respectively.Eight recombinant inbred lines(RILs)carrying both favorable alleles showed superiority in SP over the two parents in all environmental trials.A combination of the two favorable alleles using the linked markers was verified to increase SP by～5% in the RIL population and by～3% SP in diverse peanut cultivars.qSPA07.1 and qSPA08.2 were delimited to respectively a 0.73-Mb interval harboring 96 genes and a 3.93-Mb interval harboring 238 genes.Respectively five and eight genes with high expression in pods,including enzymes and transcription factors,were assigned as candidate genes for qSPA07.1 and qSPA08.2.These consistent major QTL provide an opportunity for fine mapping of genes controlling SP,and the linked markers may be useful for genetic improvement of SP in peanut.

1.Introduction

Cultivated peanut (Arachis hypogaea L.) is an annual legume crop and a major source of plant oil,proteins essential vitamins and minerals beneficial for human health [1,2].It is an allotetraploid(AABB,2n=4x=40)grain crop that was formed by natural hybridization of its two diploid ancestors,A.duranensis (AA,2n=2x=20) and A.ipaensis (BB,2n=2x=20).Peanut originated in South America and is now grown in more than 100 countries[3].In 2019,the global planting area of peanut reached 29.60 million hectares,with an annual yield of 48.76 million tonnes [4].

The peanut pod has two parts:kernel (seed) and hull.The kernel is consumed as food [5] and the hull is a byproduct with few uses that may cause environmental pollution when discarded[6].Shelling percentage (SP) is a percentage ratio of the weight of kernels to that of pods.SP varies widely among peanut cultivars.The SP of a recombinant inbred line (RIL) population (Yuanza 9102 × Xuzhou 68-4) ranged from 65% to 83% in four environments [7].Thus,there is a potential to improve SP in peanut via marker-assisted breeding.

Identification of quantitative trait loci (QTL) and development of linked markers are prerequisites for marker-assisted breeding.Association analysis [8] identified one SSR allele with only 1.5%–3.0% phenotypic variation explained (PVE) linked with SP in three field trials.Huang et al.[9] identified three QTL for SP with 2.0%–11.8% PVE in an F2population,but in only one environment [9].Another study [10] identified two QTL with only 5.7%–7.0% PVE for SP under well-watered and water-stress conditions.However,minor phenotypic effects or lack of stability in these QTL impeded the development of linked markers for use in breeding.Recently,Luo et al.[7] identified one consistent major QTL cqSPA09 with 10.5%–17.0%PVE in four environments.Cost-effective Kompetitive Allele Specific PCR (KASP) markers were developed for cqSPA09[11].It is desirable to detect more stable and major QTL for SP and to develop tightly linked markers for peanut breeding.

The purpose of the present study was to identify stable major QTL controlling SP in a RIL population phenotyped in four environments,and to develop InDel (insertion or deletion) markers for marker-assisted selection (MAS).

2.Materials and methods

2.1.Plant materials and growing environments

A RIL population(F5–7) of 186 lines was generated from a cross between Xuhua 13 (XH13,female parent) and Zhonghua 6 (ZH6,male parent) by single-seed descent.The population and the two parents were planted in Wuhan,Hubei,China in 2014 (F5generation),2015 (F6generation) and 2016 (F7generation) as well as in Huanggang,Hubei,China (F7generation) in 2016 in a randomized complete block design (RCBD) with three replications.The four environments were designated as Wuhan 2014(WH2014),Wuhan 2015 (WH2015),Wuhan 2016 (WH2016) and Huanggang 2016(HG2016) in this study.Another set of 60 peanut cultivars was planted in Wuhan,China in 2016 and 2017.Field management followed standard agricultural practices.

2.2.Trait measurement

Eight representative plants in the middle of each row were selected for recording SP.Pods harvested from the eight plants were dried and weighed.SP was calculated as weight of kernels/weight of pods)× 100%,as previously described [7,9].The mean values of SP in each environment were used as phenotypic values in QTL mapping.

2.3.Data analysis and heritability estimation

Statistical analyses including correlation analysis,analysis of variance (ANOVA),and multiple comparisons (Fisher’s least significant difference) were calculated with IBM SPSS Statistics version 19 software (IBM SPSS,Chicago,IL,USA).The broadsense heritability of SP was estimated as described [12]:

where n and r represent respectively the number of environments and the number of replications in each environment.Parameters of(genotypic variance),(genotype × environment interaction variance) and(environmental variance) in the formula were calculated by SPSS Statistics using the linear mixed model method.

2.4.Genetic linkage maps and QTL analysis

The SSR-based genetic map with 1002 markers [13] and SNPbased genetic map with 2183 markers [14] were described in Table S1.QTL analysis was performed with Windows QTL Cartographer 2.5[15]using composite interval mapping(CIM).The standard CIM model (model six) and forward regression method were selected.The number of control markers,window size,and walk speed were 5,10,and 2 cM,respectively.The threshold of LOD for declaring the presence of a QTL was determined by 1000 permutation tests.When several QTL detected in different environments had overlapping 2-LOD support intervals,they were assigned as the same QTL.A number was added to the QTL name if more than one QTL was detected for the same linkage group for SP.For example,if two QTL for SP were detected on A07,they were named qSPA07.1 and qSPA07.2.Positive and negative additive effects mean that the favorable alleles were derived from parents ZH6 and XH13,respectively.

2.5.Cluster analysis of gene expression profile

TBtools [16] was used to perform hierarchical clustering of genes in the support intervals of the two major QTL that were identified.Gene expression information based on FPKM(fragments per kilobase of exon per million reads mapped) in 22 tissues was retrieved from a previous report [17].Genes with no expression(FPKM=0) in any tissues were filtered out,and the remaining genes in the QTL intervals were used to draw a heat map with TBtools [16].Z-score was calculated by zero-mean normalization.

2.6.Development of InDel markers and PCR amplification

The 500-bp upstream and 500-bp downstream sequences of target InDels were retrieved based on published annotation (data v1.0) of the genomic sequences of Arachis hypogaea cv.Tifrunner in PeanutBase (https://www.peanutbase.org/).The InDel markers were designed with Primer3Plus (https://www.bioinformatics.nl/cgi-bin/primer3plus/primer3plus.cgi) for qSPA07.1 and qSPA08.2.These two markers closely linked to qSPA07.1 and qSPA08.2 were named SP.InDel.A07 and SP.InDel.A08,respectively.The genomic DNAs of 60 peanut cultivars including XH13 and ZH6 were extracted by the CTAB method.The PCR template concentration of all samples was controlled between 20 and 40 ng μL-1.The 10 μL PCR system including PCR master mix (2X) (5 μL),forward primer (0.5 μmol L-1),reverse primer (0.5 μmol L-1),template DNA (1 μL),and nuclease-free water was prepared according to the instructions of Thermo Scientific PCR Master Mix (Thermo Fisher Scientific,Waltham,MA,USA).PCR amplification was performed under the following conditions:initial denaturation at 95 °C for 30 min;39 cycles of preincubation at 95 °C for 30 s,annealing at 59°C for 30 s,extension at 72°C for 30 s;final extension at 72 °C for 5 min.When reactions were finished,5 μL DNA loading buffer was added to each PCR mixture and the mixtures were separated on 8% polyacrylamide gels.

2.7.Annotation of genes in QTL intervals

To identify candidate genes,all SNPs and InDels in major QTL intervals were first filtered with GATK 4.0[18]with standard hard filtering parameters [19],and then polymorphic loci with missing or heterozygous genotypes as well as polymorphic loci without polymorphism between parents were filtered out.The remaining SNPs/Indels were consider as effective polymorphic loci.SnpEff 4.2 software [20] was then used to predict the functions of these effective polymorphic loci based on the published annotation(data v1.0) of the genomic sequences of Arachis hypogaea cv.Tifrunner[21].SnpEff can predict coding effects such as synonymous or non-synonymous amino acid replacement,start codon gains or losses,stop codon gains or losses,or frame shifts based on their genomic locations.If the impact of a polymorphic locus on one gene was high or moderate,the gene was selected as a candidate for SP.Candidate genes with low expression (FPKM ＜2) in fruit,pericarp,and seed were discarded.The function of candidate genes was predicted with Mapman 3.6.0R1[22].A protein dataset(arahy.Tifrunner.gnm1.ann1.CCJH.protein.faa)was retrieved from Peanut-Base and was used to identify transcription factors of A.hypogaea cv.Tifrunner according to PlantTFDB 5.0 [23,24].

3.Results

3.1.Phenotypic variation of SP

Large phenotypic variation in SP was observed in the RIL population in the four environments,showing continuous distributions with transgressive segregation (Table 1;Fig.1).For example,the highest SP in the RIL population was 83.0% and the lowest was 61.3% in WH2016 (Table 1;Fig.S1).In contrast,the SPs of XH13 and ZH6 were only 77.2%and 77.7%,respectively.The ANOVA indicated that genotype and environment significantly influenced SP.However,the effect of genotype × environment interaction was not significant (Table 2).The broad-sense heritability of SP was estimated to be 0.91,indicating that SP was strongly controlled by genetic factors (Table 2).

3.2.Detection of QTL for SP using SSR-or SNP-based genetic map

Using the SSR-based genetic map,15 QTL with 3.4%–31.7%PVE were detected in the four individual environments (Fig.S2;Table 3).Six,five,eight,and seven QTL were detected in WH2014 (3.6%–16.0% PVE),WH2015 (3.9%–31.7% PVE),WH2016(3.4%–24.2% PVE),and HG2016 (3.7%–14.3% PVE),respectively.These QTL were distributed on chromosomes A02,A07,A08,A09,B05,and B10 (Fig.S2).Among them,seven (qSPA07.1,qSPA07.2,qSPA07.3,qSPA08.1,qSPA08.2,qSPA09.1,and qSPA09.2) were detected in at least two environments(Fig.S2;Table 3).The major QTL qSPA07.1 was detected in four environments (14.3%–31.7%PVE).

Genome-wide QTL analysis was performed using the SNP-based genetic map.A total of 16 QTL with 3.3%–16.0%PVE were detected in the four environments(Fig.S3;Table 4).Respectively five,seven,eight,and eight QTL were detected in WH2014 (3.8%–16.0% PVE),WH2015 (3.3%–10.8% PVE),WH2016 (4.0%–13.3% PVE),and HG2016(4.4%–15.4%PVE).These QTL were distributed on chromosomes A02,A07,A08,A09,B07,and B10 (Fig.S3).Five QTL(qSPA02.2,qSPA07.1,qSPA08.1,qSPA08.2,and qSPA09.2) were detected in at least two environmental trials (Fig.S3;Table 4).Among these stable QTL,qSPA07.1 (7.9%–16.0% PVE) and qSPA08.2(5.2%–10.8% PVE) were major QTL with the largest phenotypic effect in four environments.The QTL qSPA07.1 and qSPA08.2 were located on both the SSR-and SNP-based genetic maps.The phenotypic effect ranges of the two QTL in different environmental trials were 0.92–2.42 and 0.73–1.14,respectively.

3.3.qSPA07.1 and qSPA08.2 were stable QTL with the largest phenotypic effect

To estimate the combined effect of major QTL qSPA07.1 and qSPA08.2,the SNP markers TIF.07:212015 and TIF.08:31495801,which were closest to the peaks of QTL,were selected to characterize the genotypes of RILs.The genotypes of TIF.07:212015 and TIF.08:31495801 derived from ZH6 were designated as ‘‘AA” and‘‘BB”,and those from XH13 were designated as ‘‘a(chǎn)a” and ‘‘bb”,respectively.RIL lines were classified into four groups according to the genotypes of these two markers.The mean SP across four environments was 78.0% ± 2.0% for the ‘‘a(chǎn)aBB” genotype,76.0% ±2.4% for the ‘‘AABB” genotype,76.9% ± 1.9% for the ‘‘a(chǎn)abb” genotype,and 72.8% ± 2.6% for the ‘‘AAbb” genotype.A multiplecomparison test indicated that the SP of RILs with genotype‘‘a(chǎn)aBB”was significantly higher than that of lines with genotype ‘‘AAbb”(Fig.2;Table S2).Lines with ‘‘AABB” and ‘‘a(chǎn)abb” showed significantly higher SP than those with‘‘AAbb”(Table S2).The SP of RILs with ‘‘a(chǎn)abb” genotype was higher than that of RILs with genotype‘‘AABB” but not significantly (Table S2).Similarly,the SP of RILs with ‘‘a(chǎn)aBB” genotype was higher than RILs with genotype ‘‘a(chǎn)abb”but not significantly(Table S2).Generally,the combined genotype(‘‘a(chǎn)aBB”)had a higher mean SP than the other three genotype combinations(Fig.2;Table S2).Based on the combination of favorable alleles of qSPA07.1 and qSPA08.2,we tested whether adding another stable QTL,qSPA09.2,would further increase SP in the RIL population.There was no significant difference in SP between RILs with two favorable alleles (qSPA07.1 and qSPA08.2) and those with three (qSPA07.1,qSPA08.2,and qSPA09.2) (Table S3).All eight elite lines with the highest SP carried the favorable alleles of qSPA07.1 and qSPA08.2,and six of the eight carried the favorable allele of qSPA09.2 (Table 5).

The primers used for amplification of SP.InDel.A07 and SP.InDel.A08 were described in Table S4.The SP of sixty peanut cultivars ranged from 61.7% to 81.6% in two environments (Table S5).The cultivars were classified into two groups according to the genotypes of the InDel markers (Table S5).The alleles of SP.InDel.A07 and SP.InDel.A08 from XH13 were designated as ‘‘+” and ‘‘-”,respectively,while those from ZH6 were designated as ‘‘-” and ‘‘+”.The means SP of cultivars with two favorable alleles in the two environments were 76.9% ± 2.9% and 74.3% ± 3.1%,values about 3% higher than those of cultivars without favorable alleles(Table S6).

3.4.Candidate genes of major QTL on chromosomes A07 and A08

The major stable QTL qSPA07.1 and qSPA08.2 were estimated to be located in a 0.73-Mb interval (0.12–0.85 Mb) on chromosome A07 and a 3.93-Mb interval (29.00–32.93 Mb) on chromosome A08,respectively(Fig.3).The 0.73-Mb interval contained 96 putative genes,including 74 annotated genes and 22 novel genes based on Mapman analysis (Table S7).Mapman pathways analysis assigned these candidate genes to 16 pathways including RNA biosynthesis,cell cycle organization,and cell wall organization.(Table S8).There were 167 SNPs between the two parents in the target region (Table S9).Of them,26 SNPs were located in the intergenic region and the remaining 141 SNPs were located in genic regions of 60 putative genes.According to the annotation of the 141 SNPs,nonsynonymous mutation (14),stop codon (1),frame shift (2),and codon deletion (1) may affect the biological function of 12 candidate genes (Fig.S4A;Table 6).

The 3.93-Mb interval contained 238 putative genes,including 179 genes with annotations and 59 novel genes based on Mapmananalysis(Table S7).Mapman pathway analysis assigned 130 genes to 22 pathways including RNA biosynthesis,cell cycle organization,and cell wall organization.(Table S8).The interval of qSPA08.2 contained 520 effective SNPs(Table S9),of which 256 were intergenic and the other 264 SNPs located in genic regions of 118 putative genes.According to their annotation,nonsynonymous mutation(14),stop codon (1),start lost (1),and start gained (1) may affect the biological function of 16 candidate genes (Fig.S4B;Table 6).

Table 1 Statistics of phenotypes of shelling percentage in the RIL population.

Table 2 Variance analysis for shelling percentage in the RIL population.

Fig.1.Phenotypic distribution of shelling percentage in the RIL population.Dotted and solid vertical lines indicate values for Xuhua 13 and Zhonghua 6,respectively.WH,Wuhan;HG,Huanggang.

3.5.Gene expression profile of genes within major QTL intervals on chromosomes A07 and A08

The transcriptome data for tetraploid peanut (Arachis hypogaea cv.Tifrunner) provided the expression profiles of 28 candidate genes in the qSPA07.1 and qSPA08.2 intervals.Given that the expression levels of four genes in 22 tissues were 0,these genes were unlikely to be associated with SP.The remaining 24 candidate genes were selected to profile expression patterns (Fig.4;Table S10).The candidate genes,including arahy.CDPA7L,arahy.QMJH5G,arahy.I6KC6I,arahy.NPC14S,arahy.R549UJ,arahy.NWI6KJ,and arahy.LFBK1H were expressed predominantly in seed.The three genes arahy.U5CYT6,arahy.K5EKT0,and arahy.20NM7L were expressed predominantly in fruit and pericarp.The transcripts of arahy.HC0VLG,arahy.6WW0WD,arahy.NF7AFG,arahy.N6PQGP,and arahy.JX1V6X showed higher expression levels in shoots than in other tissues.The expression levels of arahy.8L3EY1,arahy.VAAE0N,arahy.1Q05D2,and arahy.P2ZS9F were relatively high in floral organs including perianth,stamen,and pistil.The genes arahy.B473LU,arahy.X0LXK2,and arahy.32PB3W were expressed mainly in peg tip,leaf,and nodule,respectively.The genes arahy.MIP29L and arahy.82Q4AG had higher transcriptional activity in roots than in other tissues.

Table 3 Identification of QTL for shelling percentage across four environments based on SSR genetic map.

Table 4 Identification of QTL for shelling percentage across four environments based on SNP genetic map.

Table 5 Shelling percentage of eight elite lines and parents across four environments.

4.Discussion

Many QTL have been identified based on SSR/SNP genetic maps in peanut [9,25–34].However,few stable major QTL for SP have been reported.The parents showed similar mean SP values.This is a common feature in practical plant breeding,in which breeders tend to select excellent cultivars for crossing [35].However,as a fundamental rule of quantitative genetics,it does not necessarily limit segregation variance [36].The transgressive segregationobserved in the RILs suggests that the parents carry complementary alleles at different QTL that were newly combined in the progeny[37].The negative additive genetic effect of qSPA07.1 revealed the maternal parent XH13 as the source of alleles increasing SP.The positive additive genetic effects of qSPA08.2 and qSPA09.2 show that the alleles for increasing SP came from parent ZH6.The finding that eight RIL lines harboring favorable alleles of both qSPA07.1 and qSPA08.2 showed stably higher SP than the two parents in four environments indicates that XH13 and ZH6 carry different favorable alleles for SP.

Fig.2.Phenotypic effect of QTL(qSPA07.1 and qSPA08.2)in the RIL population.AA and BB represent RILs with alleles from Zhonghua 6,aa and bb represent RILs with alleles from Xuhua 13;different letters (a,b and c) in the plot represented values were significantly different at the 0.05 probability level.WH,Wuhan;HG,Huanggang.

The markers within QTL intervals were aligned to the cultivated peanut genomes of Arachis hypogaea cv.Tifrunner [21].qSPA09.2 was delimited to a 16.5-Mb interval (39.35–55.85 Mb on chromosome A09) according to the QTL mapping results of the SSR-based and SNP-based maps,in which (at 45.12–50.59 Mb on chromosome A09) there was a major QTL cqSPA09 controlling SP reported in previous studies[7,11],indicating the reliability of the QTL identified in this study.qSPA07.1 was detected in all four environments and its PVE was always the largest among all QTL,whether by SSRor SNP-based map analysis.Previous report [9] had described one QTL controlling SP on chromosome A07,with 11.78% PVE.However,the QTL was identified in a single environment and the physical intervals of the QTL ranged from 24.14 to 52.78 Mb,which was different from qSPA07.1.As for qSPA08.2,it was stably detected in at least two environments.The physical intervals of qSPA08.2 ranged from 29.00 to 32.93 Mb,and few QTL have been reported in this genome region.We infer that both qSPA07.1 and qSPA08.2 are novel QTL controlling SP.In order to estimate potential values of these two QTL in peanut breeding,InDel markers developed for both were used to profile 60 peanut cultivars.Those carrying the combination of favorable qSPA07.1 and qSPA08.2 alleles showed higher SP than those without (Table S6),indicating that these two markers would be effective in marker-assisted selection for high-SP breeding in peanut.

Table 6 Candidate genes for shelling percentage on chromosomes A07 and A08.

Table 6 (continued)

Fig.3.Co-localization of the major QTL identified by SSR-and SNP-based genetic maps for shelling percentage on chromosomes A07 and A08.A07/A08 (SSR) indicates the linkage map constructed based on SSR markers;A07/A08(SNP)indicates the reversed linkage map constructed based on SNP loci;A07/A08(Mb)indicates the estimated physical positions of SSR and SNP loci on the genomic sequence of chromosome A07/A08 of Arachis hypogaea cv.Tifrunner.The positions of QTL intervals are highlighted in pink.

Fig.4.Heat map showing the expression levels of 24 candidate genes in the intervals of qSPA07.1(eleven genes)and qSPA08.2(thirteen genes).Color scale represents Z-score.

According to the transcriptome data [17],respectively five and eight genes with high expression in pods were the most likely candidate genes for qSPA07.1 and qSPA08.2(Table S10).Among the five candidate genes on chromosome A07,arahy.LFBK1H,arahy.X0LXK2,and arahy.1Q05D2 showed high transcriptional activity in fruit,pericarp,or seed (Table S10).The gene arahy.LFBK1H encodes a HSP90 protein,which mediates correct folding and stability of many client proteins.A homolog of arahy.LFBK1H,AtHsp90.6,functions in early embryogenesis[38].The gene arahy.X0LXK2 encodes a red chlorophyll catabolite reductase (RCCR),a key enzyme of chlorophyll degradation [39].The gene arahy.1Q05D2 encodes a cellulose synthase-like protein E6.Cellulose is the main loadbearing polymer of the cell wall [40].In the qSPA08.2 interval,the genes arahy.NPC14S and arahy.QMJH5G code for glucuronoxylan glucuronosyltransferase (IRX7) and UDP-D-apiose/UPD-Dxylose synthetase,respectively,and both enzymes were reported[41,42] to act in cell wall formation.The gene arahy.B473LU encodes protein kinase (SD-1),and its homolog gene affected tomato stem diameter by controlling the size and number of secondary phloem cells [43].The gene arahy.U5CYT6 encodes ligandgated cation channel (GLR),which regulate plant development[44–47].The two genes arahy.CDPA7L and arahy.20NM7L encode NAC and YABBY transcription factors,respectively.In Arabidopsis thaliana,SND1,a NAC-domain transcription factor is involved in secondary wall biosynthesis in fibers.Expression pattern analysis showed that it was expressed specifically in interfascicular and xylary fibers in stems and might act as a negative regulator of secondary wall thickening in xylary fibers [48,49].The YABBY transcription factor family is involved in multiple biological functions in higher plants,including the formation and development of vegetative and reproductive organs [50].Recently [51],transgenic expression of VvYABBY4 in tomato was shown to confer reduced plant stature,dark green leaves,elongated pistils,and reduced size of fruit and seed.The gene arahy.HC0VLG might encode a sugar phosphate/phosphate translocator,which is involved in triose phosphate transport and metabolism in plants [52].

In summary,23 QTL for SP were identified based on SSR and SNP genetic maps,including two major stable QTL.RILs and cultivars carrying the favorable alleles of these two QTL showed SP values exceeding those of the two parents.The developed markers SP.InDel.A07 and SP.InDel.A08 linked to qSPA07.1 and qSPA08.2 could be employed to identify high-SP entries among diverse peanut accessions.The two QTL regions harbor genes encoding transcription factors and enzymes.These results will guide marker-assisted breeding for SP improvement and facilitate further fine mapping of genes controlling SP.

CRediT authorship contribution statement

Weitao Li:Investigation,Data curation,Writing–original draft.Nian Liu:Data curation,Writing–original draft.Li Huang:Writing–review&editing.Yuning Chen:Writing–review&editing.Jianbin Guo:Investigation.Bolun Yu:Investigation.Huaiyong Luo:Writing– review &editing.Xiaojing Zhou:Writing– review &editing.Dongxin Huai:Writing–review&editing.Weigang Chen:Investigation.Liying Yan:Writing– review &editing.Xin Wang:Writing– review &editing.Yong Lei:Writing– review &editing.Boshou Liao:Writing– review &editing.Huifang Jiang:Conceptualization,Funding acquisition,Project administration,Writing–original draft.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This work was supported by the National Natural Science Foundation of China(31870319,31871666,and 31801403),China Agriculture Research System (CARS-13),National Program for Crop Germplasm Protection of China (2020NWB033),National Crop Germplasm Resources Center (NCGRC-2020-036),and Central Public-interest Scientific Institution Basal Research Fund(Y2021CG05).

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2021.09.003.