Genome-wide association study of ear tip barrenness in waxy maize

Abstract

Ear tip-barrenness (ETB), which results from aborted kernels or infertile florets at the ear tip, is an undesirable factor affecting the yield and quality of waxy maize. To uncover the genetic basis of ETB, a genome-wide association study (GWAS) was conducted using the genotype with 27,354 SNPs and phenotype with three environments. Five SNPs that distributed on chromosomes 1, 3 and 6, were identified to be significantly associated with ETB based on the threshold of false discovery rate (FDR) at 0.05. Among these significant loci, three SNPs were clustered together and colocalized with genomic regions previously reported. The average length of ETB decreased almost linearly from the inbred lines containing no favorable alleles across the three loci (1.75 cm) to those with one (1.18 cm), two (0.94 cm) and three (0.65 cm) favorable alleles. Moreover, three important genes, Zm00001d030028, Zm00001d041510 and Zm00001d038676 were predicted for three significant QTLs, respectively. These results promote the understanding genetic basis for ETB and will be useful for breeding waxy maize varieties with high-quality and high-yield.

Introduction

Waxy maize (Zea mays L. var. ceratina Kulesh) was first reported in China, where a large number of germplasms with a wide range of genetic diversity has been discovered (Fan et al. 2008). Due to a natural genetic mutant of waxy genes, the endosperm starch of waxy maize comprises nearly 100% amylopectin, resulting in the stickiness of its kernels (Gu et al. 2020). Because of its excellent quality, rich nutrition, unique flavor, and sticky texture, waxy maize is widely consumed as a fresh food in some Asian countries (Hossain et al. 2019, Ramekar et al. 2020). Similar to sweet maize, ears of waxy maize was directly harvested and consumed at the milk stage, thereby good ear appearance, superior flavor, and high yield are the top priorities in breeding programs (Hossain et al. 2019, Wu et al. 2019).

Ear tip barrenness (ETB), which results from aborted kernels or infertile florets at the ear tip, negatively affects maize production (Ma et al. 2020). In general, ETB adversely affects the ear appearance, and consequently reduces the commercial quality of fresh waxy maize. Kernel abortion at the ear tip also leads to substantial yield losses, because it reduces the kernel number per ear that is an important component of maize yield (Li et al. 2017). A previous study reported that the kernel yield per plant decreases by 0.0869 g with every 1 cm of ETB in maize, and estimated that kernel abortion leads to more than 10% yield losses for maize in China (Meng et al. 2008). The factors that affect ETB includes genotype, planting density, fertilizer application, and abiotic and biotic stresses (Iremiren and Milbourn 1980, Ma et al. 2020). And, current practices in shortening ETB were mainly reported by some agronomic managements, while these methods are labor-consuming and high-investing process which is not profitable for the maximizing incoming for producers (Ma et al. 2020). Developing varieties with a low rate of ETB has become the favored approach for improving the yield and quality of waxy maize.

Marker-assisted selection (MAS) is a powerful tool in crop breeding programs because of its accuracy and efficiency (Ashraf and Foolad 2013, Xiao et al. 2017). The identification of quantitative trait loci (QTL) associated with ETB could provide valuable information for maize genetic improvement in breeding programs (Yang et al. 2010). Several studies have explored the associated loci of ETB for maize. The first study on QTL mapping of ETB detected three QTLs on chromosomes 2, 3 and 6, using an F_2:3 population and simple sequence repeat (SSR) markers (Ding et al. 2016). Another study failed to identify QTL related to ETB using a doubled-haploid population, because the phenotypic data did not conform to a normal distribution (Shi et al. 2018). Recently, GWAS and QTL mapping were jointly applied to identify genetic loci associated with ETB, which reported that only one stable QTL, qETB2-1, was co-detected by the two mapping methods (Li et al. 2020). Together, the results of those studies indicated that most QTLs were population-specific and none of the QTL overlapped or could be validated in different populations, which has severely restricted their application in MAS. ETB is normally considered as a quantitative trait determined by multiple genetic loci, and it is highly sensitive to environmental factors (Ding et al. 2016, Li et al. 2020). More studies need to be conducted to validate the QTLs detected in previous studies and to identify more loci underlying the mechanisms of ETB for MAS.

GWAS has been widely applied to detect QTLs/genes responsible for natural variations in crops. Numerous genes have been successfully detected by GWAS in maize, as reviewed by Xiao et al. (2017). As a special type of maize that is cooked and consumed as a traditional vegetable, kernel abortion at the tip of ear severely affected ear appearance and commercial value of waxy maize. However, there have been no previous reports on the use of GWAS to elucidate the genetic basis of ETB in waxy maize. In the present study, we conducted a GWAS in which the genotype of 158 waxy maize accessions was evaluated based on 27,354 SNPs and their phenotypes were determined when grown in three environments. These data, along with the best linear unbiased prediction (BLUP) values, were used to detect genomic regions associated with ETB and to identify candidate genes. The identification of additional loci/genes related to ETB will increase our understanding of the genetic basis of ETB, and will facilitate MAS to breed waxy maize cultivars with high-quality and high-yielding.

Materials and Methods

Plant materials and field design

An association panel consisting of 158 waxy maize accessions, including the backbone inbred lines from the south and north of China, landraces from different ecological regions of China, and tropical lines from Thailand, were grown at Nantong (Jiangsu Province, China, 31°58ʹN 120°53ʹ2E) in 2019 and 2020, and at Sanya (Hainan Province, China, 18°15ʹN 109°30ʹE) in 2019 (Supplemental Table 1). The planting dates were April 6, 2019 at Nantong (abbreviated as E1), November 11, 2019 at Sanya (abbreviated as E2), and April 10, 2020 at Nantong (abbreviated as E3). The field experiments followed a randomized complete block design with three replications. Each plot dimension was 3.6 m × 2.4 m with row space of 0.6 m. The crop was sown at a density of 62,500 plants ha^–1. Standard fertilization and irrigation management were applied throughout the whole maize growth period.

Phenotypic data collection and analysis

For each inbred line of waxy maize, five ears were collected from plants in the middle of each row during the milk stage (R3) for phenotype measurement. The length of ETB refers to the distance from ear top to the bottom of the first sunken kernel as determined using a vernier caliper (Supplemental Fig. 1). The mean value in each plot represented the phenotypic values for the following analyses.

Phenotypic descriptive statistics were generated using a one-way analysis of variance (ANOVA) fixed-effects model with SPSS 20.0.0 software. Multiple comparisons among different groups were conducted using the least significant difference (LSD) test. The broad-sense heritability (H²) for ETB was calculated using following formula:

  

$$H^{\mathit{2}}=\frac{{\mathrm{\sigma }}_{\mathrm{g}}^2}{({\mathrm{\sigma }}_{\mathrm{g}}^2+\frac{{\mathrm{\sigma }}_{\mathrm{gl}}^2}{\mathrm{n}}+\frac{{\mathrm{\sigma }}_{\mathrm{e}}^2}{\mathrm{nr}})}$$

where ${\mathrm{\sigma }}_{\mathrm{g}}^2$ is genetic variance, ${\mathrm{\sigma }}_{\mathrm{gl}}^2$ is the variance related to the interaction between genotype and environment, ${\mathrm{\sigma }}_{\mathrm{e}}^2$ is error variance, n is the number of environments, and r is the number of replicates. The BLUP values for ETB across different environments were calculated using the “lme4” package in the R environment as follows: lme4 = phenotype~(1|line) + (1|loc) + (1|yea) + (1|line × loc) + (1|line × yea), where line, loc and yea are genetic, location and year (Sun et al. 2021).

Population structure, relative kinship, and LD decay

The population structure was estimated using STRUCTURE 2.3.4 software, with setting parameters as described by Pritchard et al. (2000). The K value was determined by the log likelihood (LnP(D)) and an ad hoc statistic ΔK based on the rate of change of LnP(D) between successive K values (Evanno et al. 2005). Based on determined K, each accession was allocated to a subpopulation for the membership value (Q value) was >0.5 (Pritchard et al. 2000). The kinship matrix was calculated from identity-by-state (IBS) distances matrix in TASSEL 5.0 (Bradbury et al. 2007). The linkage disequilibrium (LD) was also predicted in TASSEL 5.0 by computing squared allele frequency correlations (r² values) from 27,354 high-quality SNPs.

GWAS and candidate gene annotation

The genotypes of the association panel were evaluated using an Affymetrix CGBM56K SNP Array, containing 56,000 single nucleotide polymorphisms (SNPs). These analyses were conducted at China Golden Marker (Beijing) Biotech Co., Beijing, China. A total of 27,354 high-quality SNPs with minor allele frequency (MAF) >5% and missing rate <20% were obtained from the whole panel for subsequent analyses.

The GWAS was conducted using TASSEL 5.0. To control false negatives and false positives, we used six statistical models to perform the GWAS: a native general linear model (GLM), a GLM with Q-matrix, (GLM (Q)), a GLM with PCA-matrix (the top three principal components, GLM (PCA)), a mixed linear model (MLM) with K-matrix (MLM (K)), an MLM with PCA-matrix and K-matrix (MLM (PCA + K)), and an MLM with Q-matrix and K-matrix (MLM (Q + K)). The significance of SNPs was identified on false discovery rate (FDR) at 5% (Benjamini and Hochberg 1995).

The potential candidate genes within LD decay distance (~100 kb) upstream and downstream of the significant SNPs were obtained from the maizeGDB B73 reference genome v4 (https://www.maizegdb.org/gbrowse). Candidate gene annotation was performed at the NCBI (https://www.ncbi.nlm.nih.gov/). Some candidate genes were also annotated on the basis of homologous genes in rice (RAP database) or Arabidopsis thaliana (TAIR database).

Results

Phenotypic variation analysis

Descriptive statistical parameters and the H² for ETB among the tested inbred lines are shown in Table 1 and Supplemental Fig. 2. The ETB varied widely among the 158 waxy maize inbred lines, ranging from 0 to 4.5 cm across three environments and BLUP value. The phenotypic values also exhibited large coefficients of variation (CV), with the mean CV ranging from 67.8 to 78.3%. The ETB showed small or moderate deviations among the three environments (r = 0.668–0.785) (Supplemental Fig. 3), suggesting that the collected data were sufficiently reliable for further analyses. There was no correlation between ETB and anthesis silking interval (ASI) (r = 0.162) (Supplemental Fig. 4). The ANOVA results indicated that ETB was significantly affected by genotype and the genotype × environment interaction. The H² of ETB was 0.733, suggesting that genetic factors were mainly responsible for phenotypic variations in this trait.

Table 1. Descriptive statistics, ANOVA results, and broad heritability of ear tip barrenness (ETB) in 158 maize accessions grown in three environments and BLUP value

Env.	Mean (cm)	Range (cm)	CVa (%)	Skewness	Kurtosis	${\mathrm{\sigma }}_{\mathrm{g}}^2$ b	${\mathrm{\sigma }}_{\mathrm{ge}}^2$ c	H2 (%)d
E1	1.09	0–4.5	73.8	1.01	1.21
E2	1.00	0–4.1	78.3	0.98	0.89
E3	0.93	0–4.1	67.8	0.94	0.83
BLUP	0.96	0–4.0	69.8	0.92	0.89	29.1**	19.8**	73.3

E1, E2, and E3 represent environments at NanTong (2019), SanYa (2019), and NanTong (2020), respectively; ^a coefficient of variation; ^b variance of genotype; ^c variance of genotype × environment; ^d broad-sense heritability; ** Significance level at 0.01.

Relative kinship, LD decay, and genetic diversity

The population structure analysis showed that the LnP(D) values contributed to increase as K varied from 1 to 10 (Supplemental Fig. 5A); however, peak delta K (ΔK) reached the maximum value at K = 3 (Supplemental Fig. 5B), indicating that the tested accessions could be separated into three subpopulations (Groups 1, 2, and 3) (Supplemental Fig. 5C, Supplemental Table 2). The germplasm in Group 1 (including 82 inbred lines) mainly derived from a core germplasm of ‘TX 5’ in Southeast China. The Group 2 included 39 inbred lines, most of which derived from another core germplasm of ‘HB 522’. And the Group 3 contained 24 lines, which related to Thailand landraces with subtropical and tropical genetic background. While the remaining 13 lines had membership probabilities lower than 0.5 in any given group, which not distributed in the three groups.

The distribution of r² between pair-wise markers physically linked on each chromosome, as well as across all chromosomes, is presented in Fig. 1A. The r² value showed a rapid decay over increasing physical distances. The LD decay differed among chromosomes; it was fastest for chromosome 5 and slowest for chromosome 8. Across all chromosomes, the average extent of genome-wide LD decay distance was approximately 100 kb, where the LD parameter (r²) dropped to half of its maximum value. Estimation of relative kinships showed that 82.42% of pairwise relative kinships ranged from 0 to 0.1 (Fig. 1B). This result showed that a weak relative kinship existed among the tested inbred lines.

Fig. 1.

Linkage disequilibrium across 10 chromosomes (A) and pairwise relative kinship for 158 waxy maize inbred lines (B).

GWAS and annotation of candidate genes

Considering the potential for spurious associations in GWAS, six statistical models were used to process the data. The quantile-quantile plots of estimated –log₁₀ p are shown in Fig. 2. The P values from the GLM models (native, PCA, and Q) greatly deviated from the expected P value, while those from the MLMs (native, Q + K and PCA + K) were close to the expected P value. This result suggested that the rate of false positives was lower with the MLM models than with the GLM models. Among the MLMs (native, Q + K and PCA + K), the PCA + K model produced the best results (the fewest false positives), so it was used for subsequent analyses.

Fig. 2.

Quartile-quartile plots of genome-wide association study results for ear tip barrenness using six statistical models.

Five SNPs associated with ETB were detected on chromosomes 1, 3 and 6, and explained 17.0 to 19.1% of the phenotypic variation in ETB (Fig. 3, Table 2). Of these SNPs, AX-86284275, AX-86304498 and AX-86305311 were co-localized in at least two environments. Among the five SNPs, three (AX-86238683, AX-86304498, AX-86305311) were located within a 100-kb region on chromosome 6. Based on the physical position of the lead SNPs in the B73 reference genome, five SNPs were allocated to three QTL regions. These QTLs were named with the prefix “qETB” followed by the chromosome bin identifier number.

Fig. 3.

Manhattan plots and quantile-quantile plots in E1, E2 and E3 and BLUP (Black line indicates the false discovery rate of 0.05 threshold converted to –log10 p value).

Table 2. Significant SNPs and their candidate genes associated with ear tip barrenness (ETB) in waxy maize

QTL	SNP	Bin	Position	P value					FDR				R2 a (%)
				E1	E2	E3	BLUP		E1	E2	E3	BLUP
qETB1.05	AX-91334885	1.05	100,746,296				1.11E-05					0.042	19.1
qETB3.04	AX-86284275	3.04	125,029,292	7.01E-06			1.16E-05		0.039			0.042	17.0–17.2
qETB6.07	AX-86238683	6.07	162,363,997				1.23E-05					0.042	17.1
	AX-86304498	6.07	162,373,068	7.57E-06		5.03E-06	3.80E-06		0.039		0.048	0.042	18.2–19.2
	AX-86305311	6.07	162,373,499	3.96E-06			7.85E-06		0.039			0.042	17.9

^a percentage of phenotypic variance explained.

Candidate genes within average LD decay distance between the upstream and downstream 100 kb of a given QTL were identified in Supplemental Table 3. Three promising candidate genes, Zm00001d030028, Zm00001d041510 and Zm00001d038676 were marked in bold. Zm00001d038676 related to phytohormones in response to the pathways of jasmonic acid (JA). Two genes, Zm00001d030028 and Zm00001d041510, involved in pollen tube growth.

Favorable allele mining

To evaluate allelic effects, the favorable alleles were mined and 8 haplotypes (each haplotype including at least two lines) were identified across three SNPs on chromosomes 1, 3 and 6, each represented by one QTL (Table 3). The 7 lines with favorable haplotypes across all three loci showed the shortest ETB length (0.65 cm). The average length of ETB decreased almost linearly from the inbred lines containing no favorable alleles across the three loci (1.75 cm) to those with one (1.18 cm) and two (0.94 cm) favorable alleles (Fig. 4), suggesting that the loci controlling ETB length were largely additive in gene action.

Table 3. Haplotypes identified using three SNPs and their phenotypic effects. Marked base stand for favorable alleles

Haplotype	AX-91334885	AX-86284275	AX-86304498	No. of lines	Sum of favorable alleles	Average of ETB (cm)
1	G	T	A	7	3	0.65
2	G	G	A	21	2	0.94
3	G	T	G	26
4	A	T	A	20
5	G	G	G	19	1	1.18
6	A	G	A	17
7	A	T	G	32
8	A	C	A	8	0	1.75

Fig. 4.

The length of ETB in three haplotype groups of the association panel. Different letters indicate significant difference at P ≤ 0.05 estimated by LSD test.

Discussion

Factors affecting association analyses

Improving crop yield and quality are the major objectives of modern breeders to meet the demands of producers and consumers. ETB is an undesirable factor affecting the yield and quality of waxy maize, while underlying genetic mechanisms of ETB was lagging in comparison with other yield components in which a considerable number of ear-related QTLs have been detected in earlier studies (Li et al. 2011, Xiao et al. 2017, Yang et al. 2020). In the present study, a GWAS was conducted to explore the genetic basis of ETB based on the genotyping of an association panel using 27,354 high-quality SNPs and phenotyping of the panel in three environments as well as BLUP values.

The H² of the association panel was 0.733, consistent with previous studies (Ding et al. 2016, Li et al. 2020), suggesting that genotype is the main factor responsible for the phenotypic variation of ETB. Population structure is one of the fundamental factors that determines the power of GWAS (Zhou et al. 2018b). As reported previously, population structure could lead to significant differences in allele frequencies in different subpopulations, resulting in spurious associations as a result of LD between alleles (Yan et al. 2011). In the present study, the tested inbred lines were clearly grouped into three subpopulations, which were agreement with our previous study in identifying genetic loci of southern corn rust by GWAS (Zhou et al. 2018b). Moreover, using the optimal statistical model is an important factor in eliminating false positive for GWAS (Pearson and Manolio 2008, Zhang et al. 2010, Zhou et al. 2018a). It has been reported that mixed models are superior to general models (Yang et al. 2010). In the present study, the unified MLM combined with PCA + K was the best model for reducing the rate of false positives for our data, as shown in the quantile-quantile plot (Fig. 3). This model enabled us to precisely detect SNPs significantly associated with ETB in waxy maize.

SNPs related to ETB

The identification of potential chromosomal regions harboring QTLs is expected to uncover the genetic mechanisms of the targeted trait. In the present study, we detected five SNPs significantly associated with ETB that were distributed on chromosomes 1, 3 and 6, and explained 17.0% to 19.1% of the phenotypic variation in ETB. Of these loci, three associated SNPs were detected on chromosome 6 and were co-identified in at least two environments, suggesting that those loci are stable QTLs for ETB. Comparison of significant loci with those identified in previous studies revealed that they are overlapped within a QTL region identified by QTL mapping using a F_2:3 population by Ding et al. (2016). These loci were stably detected in different environments and genetic backgrounds, suggesting that they are related to a major locus that should be further validated. Importantly, these SNPs were clustered together, making it easily to be pyramided together in the future breeding program. More importantly, the remaining SNPs are putatively novel loci for ETB of waxy maize. The newly identified loci associated with ETB in this study will be very useful for MAS to improve the ear appearance of waxy maize.

Application of favorable alleles

GWAS has been widely employed in exploring genetic loci associated with diverse traits in crops. The identification of favorable alleles and genes can improve the efficiency of molecular breeding (Xu et al. 2018, Zhou et al. 2018a). Transferring and pyramiding favorable alleles or genes by backcrossing and selection can create new materials with desirable target traits (Fiedler et al. 2014, Liu and Yan 2015). The average length of ETB decreased almost linearly from the inbred lines containing no favorable alleles across the three loci (1.75 cm) to those with one (1.18 cm), two (0.94 cm) and three (0.65 cm) favorable alleles. Although no significant difference of ETB length was observed in lines with 1, 2 and 3 favorable alleles, the mean length of ETB decreased from 1.18 to 0.65 cm when pyramiding favorable allele from 1 to 3. This result indicated that ETB length could be decreased by pyramiding more superior alleles.

Candidate genes for ETB

To identify candidate genes related to ETB, we further screened putative functional genes surrounding the five significant loci based on functional annotations from public databases. Three genes Zm00001d030028, Zm00001d041510 and Zm00001d038676 were identified as candidate genes for ETB.

Zm00001d038676, a candidate gene for qETB6.07, encodes xyloglucan 6-xylosyltransferase and xyloglucan glycosyltransferase associated with xyloglucan and biosynthesis. Xyloglucan is one of the important components for the pollen-tube wall to maintain cell growth. It was reported that its homolog Arabidopsis (ATCSLC12) is highly expressed in pollen grain, and ATCSLC12 mutant impaired pollen tube growth which is an indispensable step for double fertilization in flowering plants (Kim et al. 2020). Thus, Zm00001d038676 is an important candidate gene for ETB in waxy maize because of its potential effects on pollen tube growth.

Zm00001d041510, a candidate gene for qETB3.04, encodes a probable galactosyltransferase. Its homolog Arabidopsis (GAUT13) involved in the regulation of pollen tube growth (Wang et al. 2013). It was reported that a GAUT13 GAUT14 double mutant altered distribution of pectin in the pollen tube wall, which caused serious defects in pollen tube shape and growth in Arabidopsis (Wang et al. 2013). The growth of pollen tube is an indispensable step for double fertilization in flowering plants, and defects in this process may result in sterility. Thus, Zm00001d030790 might affect pollen tube growth in waxy maize, and disruptions in the fertilization process may lead to kernel abortion at ear tip.

Zm00001d030028, a candidate gene for qETB1.05, encodes a key transcription factor in response to the pathways of Jasmonic acid (JA) in plants. JAs is well-known phytohormones involving in various metabolic processes relating to seed development, fertility, and root growth. It was reported that deficiencies in biosynthesis of the phytohormone JA could degrade stamen development, disrupt male fertility, and abolish seed production in Arabidopsis thaliana (Qi et al. 2015, Song et al. 2017, Zhuo et al. 2020). Its homologous gene in Arabidopsis (ATMYC5) known as a IIIe basic helix-loop-helix (bHLH) transcription factor. The mutant of ATMYC5 decreased seed number, associated with defects in pollen grains of Arabidopsis thaliana (Chen et al. 2016, Qi et al. 2015). For this study, we therefore suggested that Zm00001d042378 potentially functioned in seed abortion and ear tip-barrenness in maize.

In summary, the current study is the first to conduct GWAS focusing on ETB in waxy maize using the Affymetrix CGBM56K SNP Array. We detected five SNPs significantly associated with ETB distributed on chromosomes 1, 3 and 6, explaining 17.0 to 19.1% of the phenotypic variation for ETB. Three important genes that may affect ETB were predicted as promising genes, all of which involved in fertility with pollen tube and pollen grain development. This study is beneficial for improving the yield and quality of waxy maize in breeding programs by MAS’ genetic basis of ETB.

Author Contribution Statement

H.Z. contributed to experimental design. D.H. supervised the study and drafted the manuscript. H.W., M.S., H.L., Y.M., G.C. and X.H. conducted the field experiments and collected the phenotypic data. Z.Z., H.Z. (Huiming Zhang) and L.X. analyzed the data. X.S. and G.Z. wrote the manuscript. All authors had read and approved the final version of the manuscript.

Acknowledgments

We thank Jennifer Smith, PhD, from Liwen Bianji (Edanz) (www.liwenbianji.cn/) for editing the English text of a draft of this manuscript.

This study was supported by the Open Competition Project of Seed Industry Revitalization of Jiangsu Province, China (JBGS [2021]054), the Scientific and Technological Project of Nantong City, China (JC12022088, JC2021153, MS12021080 and MS22021039), the Open Project of Shanghai Engineering Research Center of Specialty Maize, China (SHERCSM2022KF01), the Jiangsu Agriculture Science and Technology Innovation Fund, China [CX (20)1002].

Literature Cited

•  Ashraf,  M. and  M.R.  Foolad (2013) Crop breeding for salt tolerance in the era of molecular markers and marker-assisted selection. Plant Breed 132: 10–20.
•  Benjamini,  Y. and  Y.  Hochberg (1995) Controlling the false discovery rate: a practical and powerful approach for multiple testing. J R Stat Soc Series B Stat Methodol 57: 289–300.
•  Bradbury,  P.J.,  Z.  Zhang,  D.E.  Kroon,  T.M.  Casstevens,  Y.  Ramdoss and  E.S.  Buckler (2007) TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics 23: 2633–2635.
•  Chen,  X.,  H.  Huang and  T.  Qi (2016) New perspective of the bHLH-MYB complex in jasmonate-regulated plant fertility in arabidopsis. Plant Signal Behav 11: e1135280.
•  Ding,  J.Q.,  J.L.  Ma,  J.F.  Chen,  T.S.  Ai,  Z.M.  Li,  Z.Q.  Tian,  S.W.  Wu,  W.  Chen and  J.Y.  Wu (2016) QTL mapping for ear tip barrenness in maize. Span J Agric Res 14: e07SC01.
•  Evanno,  G.,  S.  Regnaut and  J.  Goudet (2005) Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol 14: 2611–2620.
•  Fan,  L.,  L.  Quan,  X.  Leng,  X.  Guo,  W.  Hu,  S.L.  Ruan,  H.  Ma and  M.  Zeng (2008) Molecular evidence for post-domestication selection in the Waxy gene of Chinese waxy maize. Mol Breed 22: 329–338.
•  Fiedler,  J.L.,  R.  Afidra,  G.  Mugambi,  J.  Tehinse,  G.  Kabaghe,  R.  Zulu,  K.  Lividini,  M.  Smitz,  V.  Jallier,  C.  Guyondet et al. (2014) Maize flour fortification in Africa: markets, feasibility, coverage, and costs. Ann N Y Acad Sci 1312: 26–39.
•  Gu,  W.,  D.  Yu,  Y.  Guan,  H.  Wang,  T.  Qin,  P.  Sun,  Y.  Hu,  J.  Wei and  H.  Zheng (2020) The dynamic transcriptome of waxy maize (Zea mays L. sinensis Kulesh) during seed development. Genes Genomics 42: 997–1010.
•  Hossain,  F.,  R.  Chhabra,  E.L.  Devi,  R.U.  Zunjare,  S.K.  Jaiswal and  V.  Muthusamy (2019) Molecular analysis of mutant granule-bound starch synthase-I (waxy1) gene in diverse waxy maize inbreds. 3 Biotech 9: 3.
•  Iremiren,  G.O. and  G.M.  Milbourn (1980) Effects of plant density on ear barrenness in maize. Exp Agric 16: 321–326.
•  Kim,  S.J.,  B.  Chandrasekar,  A.C.  Rea,  L.  Danhof,  S.  Zemelis-Durfee,  N.  Thrower,  Z.S.  Shepard,  M.  Pauly,  F.  Brandizzi and  K.  Keegstra (2020) The synthesis of xyloglucan, an abundant plant cell wall polysaccharide, requires CSLC function. Proc Natl Acad Sci USA 117: 20316–20324.
•  Li,  G.H.,  J.  Liu,  S.T.  Dong,  P.  Liu,  J.  Zhang,  B.  Zhao and  D.  Shi (2017) Effects of close planting and nitrogen application rates on grain yield and nitrogen utilization efficiency of different density-tolerance maize hybrids. Scientia Agricultura Sinica 50: 2247–2258 (in Chinese with English summary).
•  Li,  J.Z.,  Z.W.  Zhang,  Y.L.  Li,  Q.L.  Wang and  Y.G.  Zhou (2011) QTL consistency and meta-analysis for grain yield components in three generations in maize. Theor Appl Genet 122: 771–782.
•  Li,  Z.,  P.  Liu,  X.  Zhang,  Y.  Zhang,  L.  Ma,  M.  Liu and  Y.  Shen (2020) Genome-wide association studies and QTL mapping uncover the genetic architecture of ear tip-barrenness in maize. Physiol Plant 170: 27–39.
•  Liu,  N. and  J.  Yan (2015) The progress and perspective of vitamin A biofortification for maize. Chinese Bulletin of Life Sciences 27: 1028–1036 (in Chinese with English summary).
•  Ma,  D.,  S.  Li,  L.  Zhai,  X.F.  Yu,  R.Z.  Xie and  J.L.  Gao (2020) Response of maize barrenness to density and nitrogen increases in Chinese cultivars released from the 1950s to 2010s. Field Crops Res 250: 107766.
•  Meng,  S.D.,  F.J.  Zhang,  Z.H.  Ding,  Q.  Sun,  L.M.  Wang,  Q.F.  Guo and  H.G.  Wang (2008) Inheritance of ear tip-barrenness trait in maize. Scientia Agricultura Sinica 41: 280–285 (in Chinese with English summary).
•  Pearson,  T.A. and  T.A.  Manolio (2008) How to interpret a genome-wide association study. JAMA 299: 1335–1344.
•  Pritchard,  J.K.,  M.  Stephens and  P.  Donnelly (2000) Inference of population structure using multilocus genotype data. Genetics 155: 945–959.
•  Qi,  T.,  H.  Huang,  S.  Song and  D.  Xie (2015) Regulation of jasmonate-mediated stamen development and seed production by a bHLH-MYB complex in Arabidopsis. Plant Cell 27: 1620–1633.
•  Ramekar,  R.V.,  K.J.  Sa,  K.C.  Park,  J.Y.  Park,  K.J.  Park and  J.K.  Lee (2020) Genetic differentiation of Mutator insertion polymorphisms and association with agronomic traits in waxy and common maize. Genes Genomics 42: 631–638.
•  Shi,  Z.,  R.Y.  Zhang,  J.F.  Xing,  M.X.  Duan,  Y.D.  Wang,  A.  Su,  F.  Wang,  L.W.  Xu,  H.L.  Tian,  J.  Wang et al. (2018) QTL mapping of three ear traits using a doubled haploid population of maize. Plant Breed 137: 706–713.
•  Song,  S.,  H.  Huang,  J.  Wang,  B.  Liu,  T.  Qi and  D.  Xie (2017) MYC5 is involved in jasmonate-regulated plant growth, leaf senescence and defense responses. Plant Cell Physiol 58: 1752–1763.
•  Sun,  X.,  W.  Ren,  P.  Wang,  F.  Chen,  L.  Yuan,  Q.  Pan and  G.  Mi (2021) Evaluation of maize root growth and genome-wide association studies of root traits in response to low nitrogen supply at seedling emergence. Crop J 9: 794–804.
•  Wang,  L.,  W.  Wang,  Y.Q.  Wang,  Y.Y.  Liu,  J.X.  Wang,  X.Q.  Zhang,  D.  Ye and  L.Q.  Chen (2013) Arabidopsis galacturonosyltransferase (GAUT) 13 and GAUT14 have redundant functions in pollen tube growth. Mol Plant 6: 1131–1148.
•  Wu,  X.,  S.  Wu,  W.  Long,  D.  Chen,  G.Y.  Zhou,  J.  Du,  Q.  Cai and  X.Q.  Huang (2019) New Waxy allele wx-Reina found in Chinese waxy maize. Genet Resour Crop Evol 66: 885–895.
•  Xiao,  Y.,  H.  Liu,  L.  Wu,  M.  Warburton and  J.  Yan (2017) Genome-wide association studies in maize: praise and stargaze. Mol Plant 10: 359–374.
•  Xu,  C.,  H.  Zhang,  J.H.  Sun,  Z.  Guo,  C.  Zou,  W.  Li,  C.  Xie,  C.  Huang,  R.  Xu,  H.  Liao et al. (2018) Genome-wide association study dissects yield components associated with low-phosphorus stress tolerance in maize. Theor Appl Genet 131: 1699–1714.
•  Yan,  J.B.,  M.  Warburton and  J.  Crouch (2011) Association mapping for enhancing maize (Zea mays L.) genetic improvement. Crop Sci 51: 433–449.
•  Yang,  L.,  T.  Li,  X.K.  Tian,  B.P.  Yang,  Y.H.  Lao,  Y.H.  Wang,  X.H.  Zhang,  J.Q.  Xue and  S.T.  Xu (2020) Genome-wide association study (GWAS) reveals genetic basis of ear-related traits in maize. Euphytica 216: 1–13.
•  Yang,  X.,  J.  Yan,  T.  Shah,  M.L.  Warburton,  Q.  Li,  L.  Li,  Y.  Gao,  Y.  Chai,  Z.  Fu,  Y.  Zhou et al. (2010) Genetic analysis and characterization of a new maize association mapping panel for quantitative trait loci dissection. Theor Appl Genet 121: 417–431.
•  Zhang,  Z.W.,  E.  Ersoz,  C.Q.  Lai,  R.J.  Todhunter,  H.K.  Tiwari,  M.A.  Gore,  P.J.  Bradbury,  J.  Yu,  D.K.  Arnett,  J.M.  Ordovas et al. (2010) Mixed linear model approach adapted for genome-wide association studies. Nat Genet 42: 355–360.
•  Zhou,  G.,  D.  Hao,  L.  Xue,  G.  Chen,  H.  Lu,  Z.  Zhang,  M.  Shi,  X.  Huang and  Y.  Mao (2018a) Genome-wide association study of kernel moisture content at harvest stage in maize. Breed Sci 68: 622–628.
•  Zhou,  G.F.,  D.R.  Hao,  Y.X.  Mao,  Q.L.  Zhu,  G.Q.  Chen,  H.H.  Lu,  M.L.  Shi,  X.  Huang,  Z.L.  Zhang,  J.  Zhao et al. (2018b) Identification of genetic loci conferring partial resistance to southern corn rust through a genome-wide association study. Eur J Plant Pathol 150: 1083–1090.
•  Zhuo,  M.,  Y.  Sakuraba and  S.  Yanagisawa (2020) A jasmonate-activated MYC2–Dof2.1–MYC2 transcriptional loop promotes leaf senescence in Arabidopsis. Plant Cell 32: 242–262.