Application of ZM00001D012005 gene in regulating starch content of maize kernels

Description

CROSS-REFERENCE TO RELATED APPLICATION

This application claims priority of Chinese Patent Application No. 202411157001.7, filed on Aug. 22, 2024, the entire contents of which are incorporated herein by reference.

TECHNICAL FIELD

The disclosure relates to the field of molecular marker-assisted breeding of maize, and specifically relates to an application of a Zm00001d012005 gene in regulating starch content of maize kernels.

REFERENCE TO A SEQUENCE LISTING SUBMITTED VIA EFS-WEB

The content of the xml file of the sequence listing named “HKIP-US-1-1353-23_sequence_listing” which is 8, 311 b in size was created on Apr. 25, 2025 and electronically submitted via EFS_Web herewith. These sequence listing is incorporated herein by reference in its entirety.

BACKGROUND

Maize is a food crop with the largest cultivated area in the world, and provides 60% of global caloric intake along with rice and wheat. In recent years, the rapid increase of population and deterioration of ecological environment threaten food safety to a certain extent, making improving maize production being one of the important means of ensuring global food safety. Currently, it is difficult to further enlarge the cultivated area of maize, therefore, improving the maize quality, especially improving the starch content of maize kernels becomes an important research direction in maize breeding worldwide. In addition, it is found in the research that multiple starch synthesis genes undergo convergence selection in the process of grain domestication. Starch accounts for 65%-75% of dry weight of maize kernels, which directly affects the kernel weight and size, and is a dominant factor in determining yield. Therefore, it is necessary to accelerate the exploration of candidate genes related to starch content of maize kernels.

Compared with temperate maize, tropical and subtropical maize germplasms have higher genetic diversity, which are important germplasm resources for maize breeding. In this study, six tropical and subtropical maize inbred lines with significant differences in kernel starch content are used as parents, to construct a multiparent population (MPP) with rich variation in kernel starch content. All the six materials used as parents are inbred lines which have important breeding values (Yin et al., 2022; Jiang et al., 2023). Exploring functional genes closely associated with the starch content of maize kernels from the MPP composed of the six parents provides a theoretical basis for molecular marker-assisted selection of maize with high starch content.

SUMMARY

In view of the shortcomings in the prior art, the disclosure provides an application of a Zm00001d012005 gene in regulating starch content of maize kernels. The functional gene Zm00001d012005, which is closely associated with the starch content of maize kernels, is identified, providing a theoretical basis for molecular marker-assisted selection of maize with high starch content.

To realize the above objective, the disclosure employs the following technical solutions.

In one aspect, the disclosure provides an application of a gene related to starch content of maize kernels in molecular mark-assisted breeding of maize, and a sequence of the Zm00001d012005 gene is as shown in SEQ ID NO: 1.

Further, a base at 2724 bp locus from 5′ terminal of the Zm00001d012005 gene shows G/A polymorphism.

In another aspect, the disclosure provides a kit, and the kit contains a reagent for detecting the Zm00001d012005 gene in maize, and the sequence of the Zm00001d012005 gene is as shown in SEQ ID NO: 1.

Further, in an implementation, an application of the kit in identifying or assisting in identifying the starch content of maize kernels is provided.

Further, the application is for detecting a genotype at the 2724 bp locus from the 5′ terminal of the sequence as shown in SEQ ID NO: 1 in maize, the genotype is AA or AG, and maize with a target trait of high kernel starch content is obtained.

In yet another aspect, the disclosure provides an application of a single nucleotide polymorphism (SNP) locus in gene-editing breeding to enhance the starch content of maize kernels, and the application involves mutating a base G at the 2724 bp locus from the 5′ terminal of the sequence as shown in SEQ ID NO: 1 to A, resulting in an AA or AG genotype.

Further, a gene-editing breeding tool is a CRISPR/Cas9 system.

In still another aspect, the disclosure provides a method for increasing the starch content of maize kernels, which involves mutating the base G at the 2724 bp locus from the 5′ terminal of the sequence as shown in SEQ ID NO: 1 to A, resulting in the AA or AG genotype.

The technical effects achieved by the disclosure are as follows:

In this study, the temperate maize inbred line Ye107 with relatively low kernel starch content is used as a common parent, which is crossed with five tropical and subtropical maize inbred lines which have relatively high kernel starch content, to construct a maize MPP with significant differences in kernel starch content. Genome-wide association study (GWAS) analysis and genetic linkage analysis are utilized to co-localize SNP_166371888 which is on chromosome 8 and significantly associated with kernel starch content, and a functional gene Zm00001d012005 that regulates kernel starch content is further identified. The gene Zm00001d012005 can explain 10.19% of phenotypic variation in kernel starch content. Haplotype analysis shows that in 521 recombinant inbred lines (RILs), Zm00001d012005 has two haplotypes, i.e., Hap1(G) and Hap2(A), with Hap2 having significantly higher kernel starch content than Hap1. Therefore, Hap2 of the Zm00001d012005 gene is a haplotype that significantly increases kernel starch content. The results of this study favor for further research on a regulatory mechanism of starch content of maize kernels, and also provide a theoretical basis for developing maize varieties with high starch content.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1a shows a phylogenetic tree for a population structure of 521 RILs in an embodiment of the disclosure;

FIG. 1b shows principal component analysis (PCA) for the population structure in the embodiment of the disclosure;

FIG. 1c shows a Bayesian clustering plot of 521 RILs when K=5 in the embodiment of the disclosure; and

FIG. 1d shows a linkage disequilibrium (LD) decay plot in the embodiment of the disclosure.

FIG. 2a is an analysis diagram showing significant quantitative trait loci (QTLs) which are related to starch content of maize kernels and are identified in pop 1 in a 22YS environment in the embodiment of the disclosure; and

FIG. 2b is an analysis diagram showing significant QTLs which are related to the starch content of maize kernels and are identified in pop 1 in a 23JH environment in the embodiment of the disclosure.

FIG. 3a shows an analysis diagram showing significant QTLs which are related to the starch content of maize kernels and are identified in pop 2 in a 21YS environment in the embodiment of the disclosure;

FIG. 3b shows an analysis diagram showing significant QTLs which are related to the starch content of maize kernels and are identified in pop 2 in a 22YS environment in the embodiment of the disclosure; and

FIG. 3c shows an analysis diagram showing significant QTLs which are related to the starch content of maize kernels and are identified in pop 2 in a 23JH environment in the embodiment of the disclosure.

FIG. 4a shows an analysis diagram showing significant QTLs which are related to the starch content of maize kernels and are identified in pop 3 in a 21YS environment in the embodiment of the disclosure; and

FIG. 4b shows an analysis diagram showing significant QTLs which are related to the starch content of maize kernels and are identified in pop 3 in a 23JH environment in the embodiment of the disclosure.

FIG. 5a shows an analysis diagram showing significant QTLs which are related to the starch content of maize kernels and are identified in pop 4 in a 21YS environment in the embodiment of the disclosure;

FIG. 5b shows an analysis diagram showing significant QTLs which are related to the starch content of maize kernels and are identified in pop 4 in a 22YS environment in the embodiment of the disclosure; and

FIG. 5c shows an analysis diagram showing significant QTLs which are related to the starch content of maize kernels and are identified in pop 4 in a 23JH environment in the embodiment of the disclosure.

FIG. 6a shows an analysis diagram showing significant QTLs which are related to the starch content of maize kernels and are identified in pop 5 in a 22YS environment in the embodiment of the disclosure; and

FIG. 6b shows an analysis diagram showing significant QTLs which are related to the starch content of maize kernels and are identified in pop 5 in a 23JH environment in the embodiment of the disclosure.

FIG. 7a shows GWAS results for starch content on the basis of a mean starch phenotype at Yanshan (YS) in 2021 in the embodiment of the disclosure;

FIG. 7b shows GWAS results for the starch content on the basis of a mean starch phenotype at YS in 2022 in the embodiment of the disclosure;

FIG. 7c shows GWAS results for the starch content on the basis of a mean starch phenotype at Jinghong (JH) in 2023 in the embodiment of the disclosure; and

FIG. 7d shows GWAS results for the starch content on the basis of a best linear unbiased prediction (BLUP) value for starch in the embodiment of the disclosure.

FIG. 8 shows a haplotype analysis diagram of starch candidate genes in the embodiment of the disclosure, showing the distribution of haplotypes of Zm00001d012005 in four subpopulations.

DETAILED DESCRIPTION

For clearer objective, technical solutions and advantages of the disclosure, the technical solutions of the embodiment in the disclosure will be described clearly and completely by reference to the accompanying drawings of the embodiment in the disclosure below. Obviously, the embodiment described is only some, rather than all embodiments of the disclosure. On the basis of the embodiment of the disclosure, all other embodiments obtained by those ordinary skilled in the art without creative efforts are included in the scope of protection of the disclosure.

Embodiment 1

1. Experiment

1.1 Plant Materials and Trial Design

Six excellent lines Ye107, CML384, CML395, YML46, YML32 and CML171 were taken as parents (Table 1), among which, Ye107 came from temperate regions, CML384 from subtropical regions, and CML395, YML46, YML32 and CML171 from tropical regions.

YS(23°19′-23°59′N, 103°35′-104°45′E) and JH(21°27′-22°36′N, 100°25′-101°31′E) of Yunnan province in China were selected as trial sites.

Ye107 served as a common parent in this study, which was crossed with five other lines, respectively. A single-seed descent method was employed from F1 to F9, to construct an MPP (pop1: Ye107×CML384, pop2: Ye107×CML395, pop3: Ye107×YML46, pop4: Ye107×YML32, and pop5: Ye107×CML171).

TABLE 1
Parental information

			Kernel
			starch
		Ecological	content
Parents	Pedigree	type	(%)
Ye107	Derived from US hybrid DeKalb	Temperature	73.8
	XL80
CML384	P502c1#-771-2-2-1-3-B-1-1-3-1(DH)	Subtropical	75.0
CML395	90323B-1-B-1-B*4-1-1-2-1(DH)	Tropical	74.3
YML46	SW1-1-1-2-1-2-1	Tropical	68.0
YML32	Suwan 1(S)C9-S8-346-2 (Kei 8902)-3-	Tropical	69.8
	4-4-6
CML171	G25QS4B-MH13-5-B-1-1-2-B-1-B-B-	Tropical	68.1
	B-1-1-6-1(DH)

A completely randomized block design was employed in the experiment, with three replicates at each site. A field trial plot was 3 meters long, with a row spacing of 0.70 meters, 14 plants per row, and two rows per plot. The trials were conducted at YS and JH in year 2022 and 2023.

1.2 Phenotypic Statistical Analysis

A near infrared reflectance spectroscopy (NIRS, No. S-14105 Kungens Kurva, Sweden) was employed to quantify kernel starch content of 601 RILs. 30 seeds were randomly selected from each line for three repeated measurements, and a mean of the three measurements is taken as a final value. Additionally, BLUP values for the phenotypic data from three replicates in two environments were calculated using a mixed linear model (MLM). A mean, a standard deviation, and a coefficient of variation for 521 RILs were calculated using Excel 2019. A Shapiro-Wilk test was performed on the phenotypic data to assess whether they followed a normal distribution. The correlation of phenotypic data in three environments was visualized at http-shiplot.com.cn/home/index.html.

1.3 Genotyping-by-Sequencing (GBS)

Firstly, genomic DNA was extracted from mature seeds using a PureLink DNA kit (Thermo Fisher Scientific, USA). The genomic DNA was fragmented using restriction enzymes PstI and MspI (New England BioLabs, Ipswich, USA), and adapters were ligated to terminals of the DNA fragments using T4 ligase (New England BioLabs, Ipswich, USA). Before the polymerase chain reaction (PCR) amplification, ligation products were pooled and purified using a QIAquick PCR purification kit (QIAGEN, Valencia, USA). Final PCR products were also purified using the QIAquick PCR purification kit, and a library concentration was measured using a Qubit 2.0 fluorometer and a Qubit dsDNA HS assay kit (Life Technologies). The library was sequenced on an Illumina NovaSeq 6000 platform (Illumina Inc., San Diego, USA) in a 150 bp paired-end sequencing mode. After sequencing, original data were filtered to remove the adapters and low-quality sequences.

1.4 SNP Identification, Filtration and Annotation

Clean reads were compared with the maize B73 v4 genome using a BWA v0.7.17 tool to generate a bam file. SNP was extracted using GATK v4.1.4.0 software, and the clean reads were compared with the maize B73 v4 genome. SNP was extracted using Plink v1.9 software to filter out loci with a missing rate higher than 20% and SNPs with a minor allele frequency (MAF) lower than 5%, with parameters set to -−geno 0.2 and -−maf 0.05. The SNPs were annotated using ANNOVAR v2021-7-16 software to determine the regions of variation loci and the types of mutations on the genome.

1.5 Population Structure Analysis and LD Analysis

TreeBeST v1.9.2 software was used for calculating a distance matrix, to construct a phylogenetic tree. Bootstrap values were obtained by means of 1000 calculations. Genome-wide complex trait analysis (GCTA) was used for performing PCA, and scatterplot3d was used to visualize the results. Admixture v1.3.0 was employed to preset K value, and population structure analysis was performed, and the results are visualized using ggplot2.

PopLDdecay software was used for calculating the degree of LD (r²) between any two makers, and script Plot_OnePop.p1 of the PopLDdecay software was used to plot an LD decay plot.

1.6 Linkage Mapping and QTL Location

Firstly, progeny genotyping was filtered on the basis of a completeness threshold of 0.8 and a segregation distortion threshold of 0.001 to obtain population markers. Subsequently, bins were created on the basis of the population markers (with bins created every 15 unlinked markers), and final population markers are obtained. Joinmap4.0 was used to order the bin markers for each population and to calculate the genetic distances between markers using a Kosambi function.

A logarithm of the odds (LOD) threshold was determined to be 2.5 through 1000 random permutation tests (P<0.05). The QTL locations of starch content were determined using a composite interval mapping (CIM) method. If a genetic distance between intervals exceeding a threshold line was less than 10 cM, they were considered as a single interval.

1.7 GWAS and Haplotype Analysis of Candidate Genes

On the basis of the mean of 521 starch content data in three environments and their BLUP values, in this study, GWAS was performed using the MLM in genome-wide efficient mixed model association (GEMMA). Population structure and genetic relationship were introduced as covariates to reduce errors, with the parameter set to −1 mm 1. SNP loci meeting or exceeding the significance threshold were extracted using bedtools v1.7. The results were visualized using CMplot v3.6.2.

With reference to the maize B73 v4 reference genome sequence in the MaizeGDB genome browser (www.maizegdb.org/), candidate genes were predicted within a 10 kb region upstream and downstream of the significant SNPs. Functional annotations of the candidate genes were obtained by browsing the MaizeGDB and NCBI (www.ncbi.nlm.nih.gov/) databases. Finally, haplotype analysis for the candidate genes was performed using Haploview v4.2 software.

2. Results

2.1 Analysis of Kernel Starch Content

The starch phenotype data of the five subpopulations are statistically analyzed, as shown in Table 2 below:

TABLE 2
Statistical analysis of kernel starch content

					Coefficient
				Range	of
Popu-	Environ-		Standard	of	variation	Heritability
lation	ment	Mean	deviation	variation	(%)	(%)
pop1	21YS	69.50	1.593	65.3-74.7	2.29
	22YS	69.22	1.811	64.4-76.5	2.62	50.33
	23JH	69.73	2.455	63.6-75.3	3.52
pop2	21YS	71.33	1.499	67.9-75.9	2.10
	22YS	71.76	1.717	67.9-77.0	2.39	47.96
	23JH	69.85	2.108	65.2-74.7	3.02
pop3	21YS	70.31	1.674	65.8-75.1	2.38
	22YS	70.12	2.107	64.9-75.5	3.00	67.91
	23JH	70.66	2.166	64.6-75.5	3.07
pop4	21YS	70.66	1.729	64.3-76.1	2.45
	22YS	70.49	2.139	65.1-76.2	3.03	55.66
pop5	23JH	71.37	2.222	65.1-76.7	3.11
	21YS	69.89	1.448	65.1-73.7	2.07	58.79
	22YS	69.22	1.986	61.4-74.2	2.87
	23JH	69.87	2.302	62.9-79.5	3.30

2.2 Population Structure Analysis and LD Analysis

Population structure analysis results are shown in FIG. 1a-FIG. 1d. Overall, a phylogenetic tree shows that the 521 RILs are clustered into five populations (FIG. 1a). The PCA of starch content is consistent with the MPP construction results in this study (FIG. 1b). For the population structure, when K=5, the 521 RILs are clearly divided into five populations (FIG. 1c).

When r² drops gradually, a genetic distance between loci is 10 kb, and the degree of association between loci tends to stabilize. These loci may contain genetic variations associated with a target trait, and therefore, in this study, the significant SNP and its 10 kb range upstream and downstream as the criteria for screening candidate genes (FIG. 1d).

2.3 QTL Location of Kernel Starch Content

In this study, on the basis of high-density genetic linkage maps of the five subpopulations, significant QTLs associated with starch content of maize kernel are screened. For pop1, three significant QTLs in the 22YS environment are detected, including two significant QTLs, qSC2-1 and qSC4-1 (FIG. 2a), and one significant QTL, qSC5-1, is detected in the 23JH environment (FIG. 2b). For pop2, two significant QTLs, qSC4-2 and qSC7-1 are detected in the 21YS environment (FIG. 3a), and two significant QTLs, qSC5-2 and qSC7-2, are detected in the 22YS environment (FIG. 3b), and five significant QTLs, qSC2-2, qSC3-1, qSC4-3, qSC4-4 and qSC7-3, are detected in the 23JH environment (FIG. 3c). For pop3, two significant QTLs are detected in the 21YS environment, and qSC4-5 and qSC7-4 (FIG. 4a), along with two significant QTLs, qSC1-1 and qSC8-1 (FIG. 4b), are detected in the 23JH environment. For pop4, three significant QTLs, qSC1-2, qSC2-3 and qSC7-5, are detected in the 21YS environment (FIG. 5a), three significant QTLs, qSC1-3, qSC1-4 and qSC2-4, are detected in the 22YS environment (FIG. 5b), and one significant QTL, qSC1-5, is detected in the 23JH environment (FIG. 5c). For pop5, two important QTLs, qSC1-6 and qSC9-1, are detected in the 22YS environment (FIG. 6a), and one important QTL, qSC1-7, is detected in the 23JH environment (FIG. 6b).

Many of the QTLs identified in this study, which are closely related to starch content of maize kernels, exhibit overlaps in different subpopulations. The overlapping QTLs are important for further investigation. It is found that QTL qSC1-2 identified in pop4 in the 21YS environment has the same interval with QTL qSC1-3 identified in the 22YS environment. Additionally, partial overlaps are observed between QTL intervals. QTL qSC2-1 identified in pop1 in the 22YS environment partially overlaps with the intervals of QTL qSC2-3 and qSC2-4 identified in pop4 in the 21YS and 22YS environments, respectively. QTL qSC4-1 identified in pop1 in the 22YS environment partially overlaps with the interval of QTL qSC4-3 identified in pop2 in the 23JH environment. QTL qSC7-1 identified in pop2 in the 21YS environment partially overlaps with the intervals of QTL qSC7-5 identified in pop4 in the 21YS environment and QTL qSC7-4 identified in pop3 in the 21YS environment. QTLs qSC1-2 and qSC1-3 identified in pop4 in the 21YS and 22YS environments, respectively, partially overlap with the interval of QTL qSC1-1 identified in pop3 in the 23JH environment. QTL qSC7-5 identified in pop4 in the 21YS environment completely overlaps with the interval of QTL qSC7-4 identified in pop3 in the 21YS environment. QTL qSC1-4 identified in pop4 in the 22YS environment partially overlaps with the interval of QTL qSC1-1 identified in pop3 in the 23JH environment. QTL qSC1-7 identified in pop5 in the 23JH environment partially overlaps with the intervals of QTLs qSC1-2 and qSC1-3 identified in pop4 in the 21YS and 22YS environments, respectively. QTL qSC1-6 identified in pop5 in the 22YS environment partially overlaps with the intervals of QTL qSC1-4 identified in pop4 in the 22YS environment and the QTL qSC1-1 identified in pop3 in the 23JH environment.

TABLE 3
Significant QTL of kernel starch content

						Phenotypic
					Interval	variation
					mapping	explained
Population	Environment	QTL	Chromosome	Threshold	(bp)	(PVE, %)

pop1	22YS	qSC	2	3.10	3600538	11.42
		2-1			2-95882685
	22YS	qSC	4	2.58	3091019	9.37
		4-1			4-38257069
	23JH	qSC	5	2.85	1766221	10.84
		5-1			31-189983598
	21YS	qSC	4	4.62	8234598	15.34
		4-2			3-93625301
	21YS	qSC	7	4.05	8627244	16.91
		7-1			8-166108588
pop2	22YS	qSC	5	2.92	1357856	10.49
		5-2			67-138001307
	22YS	qSC	7	2.61	8148643	10.00
		7-2			8-83974624
	23JH	qSC	2	3.42	1440342	10.83
		2-2			11-147190981
	23JH	qSC	3	3.31	2846109	10.45
		3-1			0-106275349
	23JH	qSC	4	2.78	3496220	10.33
		4-3			2-36255362
	23JH	qSC	4	3.14	1439315	14.45
		4-4			16-148426883
	23JH	qSC	7	2.84	1771284	9.51
		7-3			47-179180904
	21YS	qSC	4	3.27	2014982	14.08
		4-5			47-203568347
pop3	21YS	qSC	7	5.62	1338155	26.15
		7-4			48-168488752
	23JH	qSC	1	6.25	9173370	25.04
		1-1			6-176336688
	23JH	qSC	8	3.70	1503922	12.17
		8-1			18-181122637
pop4	21YS	qSC	1	4.72	1619624	17.61
		1-2			46-190891876
	21YS	qSC	2	3.18	3066481	11.39
		2-3			8-36423523
	21YS	qSC	7	2.74	1474807	9.67
		7-5			30-150363074
	22YS	qSC	1	3.83	1619624	14.44
		1-3			46-190891876
	22YS	qSC	1	3.44	8269510	12.36
		1-4			0-101538733
	22YS	qSC	2	2.98	5200902	10.94
		2-4			5-56430250
	23JH	qSC	1	2.65	2100700	9.92
		1-5			4-47132971
pop5	22YS	qSC	1	4.41	8249926	17.00
		1-6			3-92972017
	22YS	qSC	9	3.09	8697493	10.77
		9-1			2-92313888
	23JH	qSC	1	2.72	1781690	24.28
		1-7			65-179379188

2.4 GWAS of Kernel Starch Content

GWAS is conducted using 582663 high-quality SNPs in combination with the mean starch content value of 521 RILs of the MPP in three environments. Additionally, GWAS is performed using the BLUP values of starch content in all subpopulations. The MLM model in GEMMA is employed to identify loci associated with kernel starch content. In the GWAS, population structure and genetic relationship matrices are used as covariates to mitigate false positives. In the 21YS environment, two significant SNPs are identified on chromosomes 5 and 8, explaining 11.23% and 10.19% of phenotypic variance, respectively (FIG. 7a and Table 4). In the 22YS environment, seven significant SNPs are detected on chromosomes 5 and 8, explaining 5.72%-10.49% of phenotypic variation (FIG. 7b and Table 4). In the 23JH environment, two significant SNPs are identified on chromosomes 1 and 2, explaining 4.86% and 4.38% of phenotypic variation, respectively (FIG. 7c and Table 4). The GWAS based on BLUP values identifies one significant SNP on chromosome 6, explaining 8.89% of phenotypic variation (FIG. 7d and Table 4). Notably, the GWAS shows that the same significant SNP 5_97046470 is detected in both the 21YS and 22YS environments.

Given that LD decay analysis shows that the physical distance between loci decays at 10 kb, candidate genes within 10 kb upstream and downstream of the significant SNP are screened, ultimately identifying 14 candidate genes potentially associated with the starch content of maize kernels (Table 5).

TABLE 4
Significant SNP of kernel starch content

			Position		PVE
Environment	SNP	Chromosome	(bp)	Mutatuib	(%)	Theshold

21YS	5_97046470	5	97046470	G/C	11.23	5.33
	8_166371888	8	166371888	G/A	10.19	5.24
	5_96705777	5	96705777	A/G	10.41	5.23
	5_97026470	5	97026470	G/C	10.29	5.36
	5_98879482	5	98879482	T/A	10.49	5.08
22YS	5_129613503	5	129613503	C/A	8.23	5.20
	5_138562866	5	138562866	G/C	6.55	5.03
	5_1473351276	5	147335276	G/A	10.22	5.50
	8_178656036	8	178656036	T/A	5.72	5.33
23JH	1_54575694	1	54575694	G/A	4.86	5.30
	2_11478963	2	11478963	A/T	4.38	5.74
BLUP	6_137604184	6	137604184	C/T	8.89	5.48

TABLE 5
Candidate genes for kernel starch content on the basis of GWAS

SNP	Candidate genes	Chromosome	Start & Ebd	Functional annotation
5_97046470	Zm00001d015551	5	97049470-	/
			97050003
8_166371888	Zm00001d012005	8	166369165-	Histidine kinase
			166375273
5_96705777	Zm00001d015545	5	96698291-	Protein phosphatase 2C
			96698985
	Zm00001d015546	5	96701931-	/
			96707760
5_97046470	Zm00001d015551	5	97049470-	/
			97050003
	Zm00001d015571	5	98866461-	/
			98875915
5_98879482	Zm00001d015572	5	98876242-	/
			98877021
5_129613503	Zm00001d015891	5	129629958-	Protein LRKS7
			129632006
5_138562866	Zm00001d016000	5	138562425-	Myb-related protein 3R-1
			138590869
5_147335276	Zm00001d016152	5	147337518-	/
			147343809
	Zm00001d012685	8	178635302-	Mitochondrial import inner
			178641627	membrane translocase subunit
				TIM50
8_178656036	Zm00001d012686	8	178642508-	/
			178644235
	Zm00001d012687	8	178645606-	Triglyceride lipase
			178650028
1_54575694	Zm00001d029008	1	54577150-	O-fucosyltransferase family
			54580933	protein
2_11478963	Zm00001d002378	2	11470204	Cationic transporter HKT7
			11474115

2.5 Integration of QTL Location and GWAS to Reveal Candidate Genes

In this study, QTL location and GWAS analysis are employed to identify loci associated with kernel starch content. The comparison of the two analysis results shows that the candidate SNIP 8_166371888 located on chromosome 8 and identified by GWAS in the 21YS environment overlaps within the QTL interval of qSC8-1 mapped in pop3 in the 23JH environment (Table 6). Similarly, another important SNP 8_178636036 located on chromosome 8 and identified by GWAS in the 22YS environment falls within the QTL interval of qSC8-1 identified in pop3 in the 23JH environment (Table 6). On the basis of the co-localization analysis, four candidate genes (Zm00001d012005, Zm00001d012685, Zm00001d012686, and Zm0000d012687) are identified as potentially related to the starch content of maize kernels (Table 6). Zm00001d012005 is located on SNP 8_166371888. Zm00001d012685 is located on SNP 8_178636036. Zm00001d012686 and Zm00001d012687 are located nearby SNIP 8_178636036. Functional annotations of the candidate genes are performed using the NCBI and MaizeGDB databases, and the results show that Zm00001d012005 encodes the histidine kinase, Zm00001d012685 encodes the mitochondrial import inner membrane translocase subunit TIM50, and Zm00001d012687 encodes the triacylglycerol lipase.

TABLE 6
Candidate genes for kernel starch content co-localized by QTL and GWAS

					Functional
Candidate genes	Chromosome	QTL	SNP	Start & End	annotation
Zm00001d012005	8	qSC8-1	8_166371888	166369165-	Histidine kinase
				166375273
Zm00001d012685	8	qSC8-1	8_178656036	178635302-	Mitochondrial
				178641627	import inner
					membrane
					translocase
					subunit TIM50
Zm00001d012686				178642508-	/
				178644235
Zm00001d012687				178645606-	Histamine kinase
				178650028

2.6 Haplotype Analysis

Haplotype analysis shows that in 521 RILs, Zm00001d012005 (a gene sequence of Zm00001d012005 is as shown in SEQ ID NO: 1) has two haplotypes: Hap1(G) and Hap2(A). In 521 RILs, the distribution frequency of Hap1 is 254, and the distribution frequency of Hap2 is 29 (FIG. 8). It is to be noted that, in the five subpopulations, the starch content of Hap2 is the highest, and the starch content of Hap1 is the lowest. There is significant correlation between the haplotypes of Hap1 and Hap2 (FIG. 8). Therefore, it is concluded that Hap2 is the main haplotype that increases the kernel starch content, and Hap1 is the main haplotype that decreases the kernel starch content.

The embodiment described above is merely used for illustrating the technical solutions of the disclosure, rather than limiting the disclosure. Although the disclosure is described in detail by reference to the foregoing embodiment, it is to be understood by those ordinary skilled in the art that the technical solutions in each embodiment can still be modified or some technical features can be replaced equivalently, and those modifications or replacements cannot make the essence of the corresponding technical solutions out of the spirit and scope of the technical solutions in each embodiment of the disclosure.