Skip to main content

Genetic basis of maize kernel oil-related traits revealed by high-density SNP markers in a recombinant inbred line population



Maize (Zea mays ssp. mays) is the most abundantly cultivated and highly valued food commodity in the world. Oil from maize kernels is highly nutritious and important for the diet and health of humans, and it can be used as a source of bioenergy. A better understanding of genetic basis for maize kernel oil can help improve the oil content and quality when applied in breeding.


In this study, a KUI3/SC55 recombinant inbred line (RIL) population, consisting of 180 individuals was constructed from a cross between inbred lines KUI3 and SC55. We phenotyped 19 oil-related traits and subsequently dissected the genetic architecture of oil-related traits in maize kernels based on a high-density genetic map. In total, 62 quantitative trait loci (QTLs), with 2 to 5 QTLs per trait, were detected in the KUI3/SC55 RIL population. Each QTL accounted for 6.7% (qSTOL1) to 31.02% (qBELI6) of phenotypic variation and the total phenotypic variation explained (PVE) of all detected QTLs for each trait ranged from 12.5% (OIL) to 52.5% (C16:0/C16:1). Of all these identified QTLs, only 5 were major QTLs located in three genomic regions on chromosome 6 and 9. In addition, two pairs of epistatic QTLs with additive effects were detected and they explained 3.3 and 2.4% of the phenotypic variation, respectively. Colocalization with a previous GWAS on oil-related traits, identified 19 genes. Of these genes, two important candidate genes, GRMZM2G101515 and GRMZM2G022558, were further verified to be associated with C20:0/C22:0 and C18:0/C20:0, respectively, according to a gene-based association analysis. The first gene encodes a kinase-related protein with unknown function, while the second gene encodes fatty acid elongase 2 (fae2) and directly participates in the biosynthesis of very long chain fatty acids in Arabidopsis.


Our results provide insights on the genetic basis of oil-related traits and a theoretical basis for improving maize quality by marker-assisted selection.

Peer Review reports


Maize (Zea mays ssp. mays) is one of the most commonly cultivated cereal crops in the world, and a major source of human food, animal feed and bioenergy. Maize has been used as a model system for plant genetics and solved a lot of uncertain plant biological problems [1]. Maize kernels are composed of approximately 4% fat, 10% protein, and 72% starch and, supplies an energy density of 365 Kcal/100 g [2]. Oil is one of the three main components in maize kernels, whose energy is 2.25 times that of starch [3]. As a mixture, maize oil contains five fatty acids that account for more than 98% of the oil concentration including palmitic (C16:0), stearic (C18:0), oleic (C18:1), linoleic (C18:2) and linolenic (C18:3) acids [4]. Three kinds of unsaturated fatty acids account for 27.5, 51.5 and 1.4%, respectively. The high energy and polyunsaturated fatty acids in maize oil make it a highly-quality edible oil that is healthy for humans. Maize can also be used as biomass energy, which can bring considerable income to industrial production. Therefore, with the increase in oil content in maize kernels, the additional value of maize varieties will certainly increase.

Long-term artificial selection of high-oil maize populations has led to the creation of a series of genetic resources, including Illinois high-oil (IHO) population and Beijing high-oil population [1, 5], which were subsequently widely applied to dissect the genetic architecture of oil biosynthesis in maize kernels. In view of the limitations of analytical methods and molecular markers, the accuracy of QTL mapping based on the biparental population is low, and it is difficult to clone genes decades ago. Subsequently, with the development of sequencing technology [6,7,8], an increasing number of molecular markers have been applied to QTL mapping, which greatly improves the accuracy of QTL mapping. A large number of chromosomal regions and QTLs affecting oil concentration and fatty acid composition were identified using segregating populations in maize [9,10,11,12,13,14,15,16], and these studies indicated that the oil concentration and fatty acid composition were controlled by a few major genes and many minor genes with mainly additive effects. In addition, epistatic interactions also contribute to variations in oil content in specific populations [15, 16]. Similar results were obtained in two publicly available maize genetic resources, NAM (the nested association mapping population) [17] and AMP508 (association mapping population) [3] based on high-resolution and high-power QTL analysis.

The biological processes of oil synthesis and accumulation are complex in plant seeds and are well known in Arabidopsis, in which 120 enzymes and more than 600 genes are involved [18]. However, little is known about maize, and only a few genes related to oil content and fatty acid composition have been cloned [19,20,21,22]. For example, the DGAT1–2 gene, which was cloned by map-based cloning, encodes an acyl-CoA: diacylglycerol acyltransferase, and catalyzed the final step of oil synthesis, which can affect oil and oleic-acid contents [20]. Afterwards, a DGAT-based association analysis was carried out to identify the functional loci and develop two PCR-based functional markers [23]. Stearoyl-ACP desaturase (SAD) plays a key role in fatty acid biosynthesis, and has been identified in maize by gene-based association analysis [24]. These results suggested that gene-based association mapping was a suitable strategy for revealing the candidate genes underlying QTLs and shortening the time of gene cloning.

In the present study, a RIL population derived from two maize inbred lines, KUI3 and SC55, consisting of 180 individuals was used to: (1) dissect the genetic architecture of oil concentration and fatty acid composition, (2) estimate the number and effects of QTLs and epistatic interactions underlying oil-related traits, and (3) identify candidate genes that control oil related traits.


Genetic materials and field experiments

A RIL population consisting of 180 individuals was developed by a cross between two maize inbred lines KUI3 and SC55. The two parents originated from an previous reported association panel that contained 508 genetically diverse maize inbred lines (AM508) [25]. Hybrid F1 was self-pollinated 6 times to produce the F7 generation by single-seed descent. All lines and parents were planted in a randomized complete block with two replicates at Beijing and Hainan in 2013. Each family line was grown in a single-row plot (2.5-m rows, 0.67 m between rows), and the planting density was 45,000 plants/ha.

Measurement of fatty acids in maize kernels

Mature ears were harvested and shelled manually. Fifty kernels were randomly selected, dried for 60 h at 45 °C, ground into powder, and then stored in a desiccator for fatty acid measurement. Lipid was extracted as described by Fang et al. [16]. A HP7890A gas chromatogram (GC) (Agilent Technologies, USA) was employed to analyze the fatty acids compositions. An HP-INNOWAX polyethylene glycol capillary column (30 m × 320 μm × 0.5 μm, Agilent Technologies) was used to separate the samples at 250 °C. The GC was operated at a constant flow pressure of 140.9 kPa, the initial oven temperature of 220 °C, with a 16 min isothermal, and then the oven temperatures were increased by 20 °C / min to 240 °C, with a 5 min isothermal. The FID temperature was 250 °C, and the split ratio of nitrogen was 10:1.

Nine distinct fatty acids were measured, including palmitic (C16:0), palmitoleic (C16:1), stearic (C18:0), oleic (C18:1), linoleic (C18:2), linolenic (C18:3), arachidic (C20:0), behenic (C22:0), and lignoceric (C24:0) acids, and the oil content was calculated as the sum of the oils. Another 9 ratio traits were derived from 9 fatty acids: C16:0/C16:1, C16:1/C18:0, C18:0/C18:1, C18:1/C18:2, C18:2/C18:3, C18:0/C20:0, C20:0/C22:0, C22:0/C24:0, SFA/USFA (saturated fatty acid = C16:0 + C18:0 + C20:0 + C22:0 + C24:0; unsaturated fatty acid = C16:1 + C18:1 + C18:2 + C18:3). The detailed protocol was described in Fang et al. [16].

Analysis of phenotypic data

Phenotypic data processing was performed with R Version 3.6.1 ( Analysis of variance was performed with the “aov” function in R to evaluate the genotype, environment and replication effect. The model for ANOVA was y = μ + αg + βa + (αβ)ge + γar + εgar, where μ represents the grand mean (the total of all the data values from two environments divided by the total sample size), αg represents the effect of the gth line, βa represents the effect of the area, (αβ)ge represents the effect of the line × area interaction, γar represents the effect of the area × replicate interaction, and εgar represents the residual. The broad-sense heritability of each trait was calculated as H2 = αg2/(αg2 + αge2/r + αδ2/ar) [15, 16], where e is equal to 2 environments, r is equal to 2 replications in each environment, αg2 is the genetic variance, αge2 is the interaction of genotype with environment, and αδ2 is the residual error. The 90% confidence intervals of H2 were computed.

The function “lmer” in the lme4 package of R was used to fit to a linear mixed model to obtain the best linear unbiased prediction (BLUP) values for each trait of each individual line: yijk = μ + gi + ej + εijk, where yijk is the kth phenotypic value of individual i in jth environment, μ is grand mean of all environments, gi is the ith genetic effect, ej is the effect of different environments and εijk is the random error. μ was considered a fixed effect and gi and ej were considered random effects [26]. The BLUP values for each line were used for the phenotype description statistics, Pearson correlation coefficient analysis and QTL mapping.

Construction of genetic map

All family lines, together with their parents, were genotyped with the Illumina MaizeSNP50 BeadChip (Illumina Inc., San Diego, CA, USA), which contains 56,110 single nucleotide polymorphisms (SNPs) covering 19,540 maize genes [27]. The in-house Perl scripts were used to compare the genotypes between parents and the RILs. Missing data, heterozygosity and minor allele frequency for all SNPs and the missing data and heterozygosity for each line were calculated. After quality control, 180 inberd lines with missing data of < 15.0% and heterozygosity of < 8.0% were used for further analysis [28]. A total of 11,841 SNPs were polymorphic with 2372 genetic blocks captured. A modified physical order method as described in Pan et al. [28], was used to construct the genetic linkage map, with all lengths of 1974.8 cM (Fig. S1).

QTL mapping

Windows QTL Cartographer 2.5 was used to perform QTL mapping for all traits using composite interval mapping [29]. Model 6 of the Zmapqtl module was used to detect QTLs in the whole genome. The scanning interval between markers was set at 2.0 cM, with a 10 cM window size. Forward-backward stepwise regression with five controlling markers was used to control for background from flanking makers. After 1000 permutations, the threshold logarithm of odds (LOD) value to declare putative QTLs was determined at a significance level of P < 0.05. The confidence interval of QTL position was estimated with the one-LOD support interval method [30]. The R function ‘lm’ was performed to determine total phenotypic variation explained (PVE) by significant QTLs [4, 16].

Epistasis analysis

As described as shown in Wen et al. [16, 31], a two-way ANOVA was carried out to estimate the pairwise additive × additive epistatic interactions for all identified QTLs for each trait at P < 0.05. The proportion of variance explained by epistasis was evaluated by comparing the residual of the full model that contained all single-locus effects and two-locus interaction effects with that of the reduced model that excluded two-locus interaction effects. In addition, the peak bin markers were used in the epistatic interaction analysis and all the heterozygous genotypes were assigned as missing values for simplicity.

Gene-based association analysis

The SNPs located in the gene body and regions within 5 kb upstream and downstream of the coding region were extracted from the 1.25 million SNPs with MAF ≥ 0.05 in a panel of 508 maize lines [32]. The associations between all SNPs and oil-related traits were analyzed using a mixed linear model [33] considering the population structure [34] and relative kinship [35]. Bonferroni adjusted significance thresholds (P ≤ 0.01/n) of, P ≤ 1.6 × 10− 4 for GRMZM2G101515 and P ≤ 1.5 × 10− 4 for GRMZM2G022558, were used to identify significant associations. Linkage disequilibrium (LD) between two sites was calculated with TASSEL 5.0 [35].

Correlation between gene expression and traits

The analysis of the correlation between gene expression in developing kernels at 15 days after pollination (DAP) [36] and oil-related traits in mature kernels [4] was performed using the R function ‘cor.test’.

Annotation and gene expression analysis of 19 colocalized genes

Based on the information available in the MaizeGDB (, the function of each gene was inferred from orthologues in Arabidopsis or rice. The data of gene expression in developing embryo, endosperm and seeds was obtained in Chen et al. [37].


Phenotypic variation, correlation and heritability

Phenotypic variations of the 19 target traits for parents and RILs are shown in Table 1. The mean value of SC55 (5.27%) was higher than that of KUI3 (4.13%) for oil content. The KUI3/SC55 RIL population harbored abundant diversity for most of the investigated phenotypic traits, for which continuous and approximately normal distributions were observed (Fig. S2). The mean of the KUI3/SC55 RIL population based on the BLUP values was close to the mid-parent value for almost all measured traits with transgressive segregation (Table 1), suggesting that both parents harbored the alleles responsible for increasing the oil-related traits. The coefficient of variation (CV) values ranged from 7.39% (C16:0) to 52.10% (C18:0/C18:1), with an average of 33.04% (Table 1). Highly significant effects of genotype, environment and genotype × environment interactions were observed by the ANOVA analysis of all traits except C16:1 (Table 1), indicating that oil-related traits are sensitive to genotypes and environments. Pairwise Pearson’s correlation coefficients of 19 traits revealed that most of the traits showed a significant correlation with each other, with coefficients from 0.17 between C16:0/C18:0 and C18:0/C18:1 to 0.984 between C18:1 and C18:1/C18:2 in the KUI3/SC55 RIL population (Fig. S3). Broad-sense heritability (H2) was high for all traits, ranging from 0.56 to 0.89, indicating that most of the phenotypic variations were genetically controlled (Table 1).

Table 1 Descriptive statistics, broad-sense heritability and ANOVA for 19 oil-related traits in KSC population

Genetic architecture of the oil-related traits

Based on a linkage map of 1974.8 cM, QTLs for 19 oil-related traits were detected in the KUI3/SC55 RIL population. After 1000 permutations, the empirical threshold logarithm of odds (LOD) value for all traits (P < 0.05) was 3.2, and the values ranged from 3.2 to 3.6. In total, 62 single QTLs distributed in 38 genomic regions across all chromosomes were detected, with the QTL number per trait ranging from 2 to 5 in the KUI3/SC55 RIL population (Fig. 1a; Table S1). The 1-LOD QTL interval averaged for 9.9 Mb (5.9 cM), with a range from 0.2 to 68.9 Mb (1.2 to 13.8 cM). The phenotypic variation that could be explained by each QTL (PVE) ranged from 6.68% (qSTOL1) to 31.02% (qBELI6), with an average of 10.3% and the total PVE of all detected QTLs for each trait ranged from 12.5% (OIL) to 52.5% (C16:0/C16:1) (Fig. 1b). Of all these identified QTLs, only 5 had a large effect, with PVE ≥ 15% in three genomic regions on chromosome 6 and 9. The QTL with the largest effect, qBELI6, was C22:0/C24:0 on chromosome 6, which was flanked by markers SYN12691 and SYN24474, and accounted for 31.02% of the phenotypic variation. The QTL- qLIG6 for C24:0 with the second largest effect was on chromosome 6, and accounted for 24.46% of the phenotypic variation, with alleles from KUI3 being responsible for the increasing effect. Additionally, two parents, KUI3 and SC55, harbored similar numbers of favorable alleles at, 29 and 33, respectively (Fig. 1c), suggesting that many favorable alleles existed in regular maize lines with minor effects.

Fig. 1
figure 1

The distribution and effects of single QTL identified for 19 oil related traits in KUI3/SC55 RIL population. a Distribution of single QTL on chromosomes. QTL regions across the maize genome are represented by confidence intervals, and LOD values are scaled by color. b Total PVE of single (blue bars) and epistatic (red bars) QTLs for each trait. c Effect size (represented by PVE) and the origin of the increasing alleles of the identified QTLs. Blue and red bars indicate that increasing alleles come from KUI3 and SC55, respectively

In addition to single QTLs, two pairs of epistatic QTLs referring to 3 loci were detected for two traits, C18:2 and C18:1/C18:2 (Table S2). The two epistatic QTL pairs explained 3.3 and 2.4% of the phenotypic variation. Considering that the number and effect of epistatic QTLs were small, epistatic interactions between two QTLs with additive effects contributed less than additive effects to the genetic basis of oil-related traits in the KUI3/SC55 RIL population.

QTL clusters

Fourteen QTL clusters were observed in this study, of which 6 covered no less than 3 single QTLs (Fig. 1a; Table S1), and the others covered 2 single QTLs. Specifically, L25 contained 5 QTLs for 5 oil-related traits on chromosome 6: C18:2, C22:0, C24:0, C18:1/C18:2 and C20:0/C22:0. The PVE of these QTLs ranged from 8.9 to 24.46%, of which two were main-effect QTLs (Fig. S4a). On chromosome 1 and 9, there were two loci harboring 4 QTLs: L6 for C16:0, C20:0, C18:0/C18:1, and SFA/USFA and L34 for C24:0, C22:0, C20:0, and C18:0/C20:0, respectively. All these QTLs were minor-effect QTLs except qARA9, whose PVE ranged from 6.68 to 15.22%. The region of L34 spanned 25.8–100.4 Mb and was much larger than that of L6 (257.6–264.5 Mb), as a result of low-frequency recombination events occurring in the interval of L34. The other 3 loci, i.e., L12, L20 and L28, contained 3 QTLs and were located on chromosome 3, 4 and 6, respectively. L28 spanned a small region from 164.0 to 165.4 Mb, contained qBELI6, qBEH6–2 and qARBE6–2 and could explain 31.02, 10.82 and 15.36% of the phenotypic variation, respectively, which makes it a valuable target for further gene cloning. The spanning interval of L20 was 1.5 Mb (237.5–239.0 Mb), which was much smaller than that of L12, whose interval was more than 20 Mb. All the 6 QTLs for L12 and L20 had minor effects, with PVEs ranging from 6.75 to 11.91%.

Interestingly, QTL-gene colocalization identified 2 known genes falling within 2 loci for L25 and L19 (Fig. S4). The DGAT1–2 gene [20], which encodes an acyl-CoA: diacylglycerol acyltransferase and catalyzes the final step of oil synthesis, is located in the interval of L25, and might be the candidate gene (Fig. 4a). Additionally, the FAD2 gene encoding fatty acid dehydrogenase colocalized with L19, which covered qPAE4–3 and qPALE4–2 (responsible for C16:1 and C16:0/C16:1, respectively) (Fig. S4b).

Mining of candidate genes for oil-related traits by linkage and gene-based association analysis

Combined with a previous report about a GWAS for 21 oil-related traits [4], 19 (25.7%) of the 74 candidate genes were detected in this study based on physical position (Fig. 2; Table S1). These genes covered 10 loci that have the potential to affect oil biosynthesis and accumulation in maize kernels. Of the 19 genes, 4 encoded enzymes involved lipid metabolism reactions that directly regulated the lipid synthesis and metabolism including fatty acid desaturase 2 (FAD2), fatty acid elongase 2 (FAE2), diacylglycerol acyltransferase (DGAT1–2) and Myristoyl-acyl carrier protein thioesterase (Fig. S5; Table S4). four genes were annotated as enzymes involved in other metabolism reactions, such as aldehyde dehydrogenase, Ser/Thr protein phosphatase, acid phosphatase and alpha/beta-Hydrolases. One gene was annotated as transcription factor. The proteins encoded by remaining 10 genes were classified as chaperonin protein, ribosomal protein, zinc finger, G protein, cytochrome P450 and proteins with unknow function (Fig. S5). In addition, combined with the published RNA-seq data, we found that 94.7% (18/19) of these genes expressed in developing embryo, endosperm and seeds except GRMZM2G141999 (Fig. S6; Table S4). Eight genes were highly expressed in developing embryo at various stages, which indicated the potential roles in lipid synthesis and metabolism, because the embryo was the main site of oil accumulation.

Fig. 2
figure 2

Co-localization of QTLs and known genes for oil-related traits. All 10 maize chromosomes (Chr 1–10) were depicted to scale (Mb, million base pairs). Red region indicated the confidence interval region of QTLs and the black lines indicated the position of genes

Coincidentally, two of the 8 genes highly expressed in embryo were located in two QTL clusters in physical position, L28 on chromosome 6 and L34 on chromosome 9, which fell in the peak bin of the most colocalized QTLs (Fig. 3a, e), and were considered important candidate genes. L28 contained 3 QTLs, namely, qBELI6 for C22:0/C24:0, with the largest PVE of 31.02%, moderate effect QTL-qBEH6–2 for C22:0 and major QTL-qARBE6–2 for C20:0/C22:0, encompassing the GRMZM2G101515 gene, which can encode a protein with an unknow function (Fig. 3a). To further explore the association between the gene and oil-related traits, 62 SNPs were extracted in the gene body and region within 5 kb upstream and downstream of the GRMZM2G101515 gene from 1.25 million high-quality SNPs with MAF ≥ 0.05 in 508 maize inbred lines [32]. A marker-trait association analysis with these SNPs using a mixed linear model identified 4 significant loci associated with C22:0/C24:0 at P ≤ 1.6 × 10− 4 (Fig. 3b, Table S3). The most significant SNP, chr6.S_164986588 for alleles A and C, was located on exon 5 at P = 7.91 × 10− 6, which can give rise to a change in amino acids for glutamate (T) to threonine (P). The LD between chr6.S_164986588 and the other three SNPs ranged from 0.34 to 1 (Fig. 3c). The lines with allele C have a higher ratio of C22:0/C24:0 than those with allele A (Fig. 3d). Meanwhile, the lines harboring allele A expressed GRMZM2G101515 at slightly high levels at P = 0.025 (Fig. 4a), and the gene expression and the ratio of C22:0/C24:0 showed a weak correlation (Fig. 4b). These results showed that the expression difference of the GRMZM2G101515 gene might affect the ratio of C22:0/C24:0.

Fig. 3
figure 3

Candidate genes for L28 and L34. a LOD profiles of the identified QTL bins for L28 and GRMZM2G101515 were colocalized with a QTL cluster. The dashed lines show the physical positions of genes. b Candidate-gene association analysis for GRMZM2G101515. The most significant locus is shown in red. The intensity of gray shading indicates the extent of LD (r2) between the most significant locus and the other variants identified in this region. The gene shown on the x-axis. The black and light-gray shading represents exons and UTRs, respectively. c The linkage disequilibrium (LD) patterns of all identified variants (MAF ≥ 0.05) in genes GRMZM2G101515. d The effect of peak locus for GRMZM2G101515 in an association panel. e LOD profiles of the identified QTL bins for L34 and GRMZM2G022558 were colocalized with a QTL cluster. f Candidate-gene association analysis for GRMZM2G022558. g The linkage disequilibrium (LD) patterns of all identified variants (MAF ≥ 0.05) in genes GRMZM2G022558. h The effect of peak locus for GRMZM2G022558 in an association panel

Fig. 4
figure 4

GRMZM2G101515 and GRMZM2G022558 alter the ratio of fatty acids in maize kernel. a Comparison of the relative expression at peak locus alleles A and C for GRMZM2G101515. b Correlation of C22:0/C24:0 with the relative expression of GRMZM2G10151 in kernels at 15 DAP. c Comparison of the relative expression at peak locus allele A and G for GRMZM2G022558. d Correlation of C18:0/C20:0 with the relative expression of GRMZM2G022558 in kernels at 15 DAP

In addition, 3 of 4 QTLs contained in L34, namely, qBEH9 for C22:0, qARA9 for C20:0 and qSTAR9 for C18:0/C20:0, colocalized with the GRMZM2G022558 gene (Fig. 3e). This gene encodes fatty acid elongase 2 (fae2), which is involved in the biosynthesis of very long-chain fatty acids in Arabidopsis [38], and is incorporated into a variety of plant lipids. Similarly, from a 1.25 million SNP database of 508 maize inbred lines, we extracted 66 SNPs with MAF ≥ 0.05 spanning from 5-kb up- to downstream of fae2 coding region, and then 35 SNPs that associated with C18:0/C20:0 were identified by marker-trait association analysis, including 2 in the 5’UTR, 6 in exon1, 1 in the 3′ UTR and the rest in the region behind the 3’UTR. The peak signal, chr9.S_86864550 in exon 1, whose P-value reached 7.52 × 10− 11(Fig. 3f, Table S3), was in LD with almost all the other significant SNPs, which can lead to amino acid changes (Fig. 3g). Lines harboring allele A were significantly greater than those harboring G at the ratio of C18:0/C20:0 (Fig. 3h). Meanwhile, the expression of fae2 in lines with allele A was distinctly higher than that in lines with allele G (Fig. 4c) and the expression level was correlated with the ratio of C18:0/C20:0 at P = 0.002 (Fig. 4d). These results indicated that changes in the expression of fae2 could change phenotypes.


Genetic components of oil-related traits in maize kernels

Maize oil is a compound made up of different kinds of fatty acids. Previous studies have shown that oil-related traits are quantitative traits controlled by multiple genes [9,10,11,12,13,14,15,16,17], which was also revealed by the finding that all traits followed normal distributions and showed transgressive segregation in this work. Our study detected 62 QTLs at 38 loci, of which only 5 were major QTLs, with PVE ≥ 15%. Therefore, two contrasting genetic architectures were found for 19 oil-related traits. Three fatty acid traits, C20:0, C22:0, and C24:0 and two ratio traits C20:0/C22:0 and C22:0/C24:0, were controlled by a single major QTL plus some small-effect QTLs, while the others were controlled by many small-effect QTLs. It is worth noting that only two minor-effect QTLs were detected for oil content in the present study, which is in keeping with a recent report [16], while 6–16 QTLs were identified in high-oil populations [11, 13,14,15], 22 QTLs identified in the NAM population [17] and 26 loci associated with oil content throughout a GWAS analysis [4]. The application of different mapping populations gave rise to differences in QTL numbers and effects. In the present study, subtle variations in oil content between two parents, KUI3 (4.13%) and SC55 (5.27%), were observed. Two QTLs contributed to 12.5% of the phenotypic variation in oil content in the KUI3/SC55 RIL population. The PVE for oil content was more than 50% in high-oil populations [15, 39]. Favorable allele accumulation is a route for increasing oil concentration, and high-oil maize lines have more favorable alleles, some of which have main effect; therefore, an increasing number and effect of QTLs were identified in the biparental population constructed by high-oil lines. Nevertheless, favorable alleles also existed in regular maize lines according to the founding in this study. In addition, varying environments could also influence the number of detected QTLs.

Another remarkable finding in this work is that two pairs of epistatic interaction QTLs with additive effects were identified for two oil-related traits. However, they presented limited contributions to increasing the fatty acids composition. This result was consistent with a few previous reports [15, 16]. As an example, 2–7 pairs of epistatic QTLs were detected for oil content and 5 fatty acid compositions in high-oil maize [15]. The proportion of total phenotypic variance explained by all epistatic QTLs ranged from 5.2 to 12.6% for each trait. Similar results were obtained in rapeseed, rice, peanut and wheat [40,41,42,43], demonstrating that epistasis could make a substantial contribution to variation in complex quantitative traits in different crops. The magnitudes of individual QTLs with additive effects and the percentage of total phenotypic variation explained by individual QTLs were greater than those of epistatic QTLs, indicating that additive effects rather than epistatic effects played a crucial role in contributing to the genetic basis of oil-related traits in the KUI3/SC55 RIL population.

Colocalization of oil-related QTLs identified in this study with previous studies

The mining of oil-related QTLs is beneficial for a better understanding of oil biosynthesis and accumulation in maize kernels. In comparison with previous studies [11, 13,14,15,16,17], 75.8% (47/62) of all identified QTLs in this study were previously reported based on the B73 reference genome version 2, indicating the reliability and accuracy of the results. All 15 newly identified QTLs had moderate effects and were distributed on 8 chromosomes (except for chromosome 5 and 6), and the PVE ranged from 6.75% (qSTOL4) to 11.94% (qPAST8), which revealed the specificity of the genetic background from the two parents. Considering the physical position of all QTLs, 26.3% (10/38) of loci were freshly verified in the current study, including two QTL clusters, L20 and L33, located on chromosome 4 for C18:0, C18:0/C18:1, and SFA/USFA and chromosome 8 for C16:0/C18:0 and C18:0/C20:0, respectively. Pleiotropy and close linkage could cause trait correlations and lead to colocalization of QTLs, which means that a few QTLs controlling different traits were identified in the same genomic regions [44]. These two new loci supported by multiple QTLs are credible and have the potential to serve breeding. Moreover, the reasons for colocalization of the two loci require further study.

The two candidate genes, GRMZM2G101515 and GRMZM2G022558, were associated with C20:0/C22:0 and C18:0/C20:0, respectively. The former encodes a kinase-related protein with unknow function, and in Arabidopsis, it was annotated as an RNA polymerase II degradation factor-like protein. No reports have shown that this gene is related to lipid metabolism. Given that this gene is not in the lipid metabolism pathway, it probably regulates lipid metabolism in an indirect way. In addition, the most significant SNP, chr6.S_164986588, is not necessarily the functional site of the gene. In most cases, the change of a single amino acid is not enough to change the protein function. The expression of GRMZM2G101515 is related to the phenotype, which means that the phenotype is likely to be regulated by gene expression. The real functional site is possible to be some transposons or structural variations undiscovered by the next-generation sequencing technology, which are in LD with the significant SNPs, just like the way of ZmNAC111 and ZmVPP1 [45, 46] work. The second encodes fatty acid elongase 2 (fae2), which can elongate fatty acyl-CoAs to produce C20-C24 acyl-CoAs and then further produce long-chain fatty acids, and it directly participates in the biosynthesis of very long-chain fatty acids in Arabidopsis [18]. Similarly, although many significant SNPs were identified to be associated with C18:0/C20:0, it’s hard to determine the functional site of GRMZM2G022558 gene. The significant correlation between the phenotype and gene expression in the association panel indicated the possibility that expression regulated the phenotype.

QTL application in the improvement of maize oil

QTL mapping is a classical strategy to identify loci for complex quantitative traits of interest. The ultimate goal of QTL mapping is to clone the causal genes for further application in trait improvement. It usually takes a long time to obtain genes underlying QTLs by constructing near-isogenic lines in maize [47,48,49], rice [50,51,52], wheat [53, 54] and other plants [55, 56]. Combining linkage analysis and GWAS can greatly shorten the journey [57, 58], and gene-based association studies can help identify the favorable allele. These QTLs or genes have the potential to contribute crop improvement by marker-assisted selection. To date, a large number of QTLs for different traits in multiple species have been detected, and some of which have actually been applied to crop improvement [59,60,61]. However, little is applied in oil improvement for maize kernels apart from the DGAT1–2 gene [62]. In detail, Hao et al. [62] transferred the favorable allele of DGAT1–2 from the high-oil inbred line (By804) into two parents of Zhengdan958 using marker-assisted backcrossing and successfully increased the oil content of the improved Zhengdan958 without a change in grain weight. In the present study, 5 major QTLs were identified, and 3 of these QTLs were isolated by joint gene-based association analysis, and two candidate genes were verified. In addition, these QTLs were mainly additive in the KUI3/SC55 population, which may accelerate molecular breeding by pyramiding the favorable maize alleles of these detected QTLs or by genomic selection.


In the present study, QTL mapping for 19 oil-related traits was conducted with high-density SNP markers in the KUI3/SC55 RIL population. A large number of QTLs regulating oil content and fatty acid composition were identified, most of which were moderate effect QTLs. Two contrasting genetic architecture were revealed for 19 oil-related traits. Of these traits, only five harbor a major QTL, reflecting the complex nature of oil-related traits. In addition, additive effects rather than epistatic effects played a crucial role in contributing to the genetic basis of oil-related traits in the KUI3/SC55 RIL population. Two genes, GRMZM2G101515 and GRMZM2G022558, were further verified to be associated with C20:0/C22:0 and C18:0/C20:0, respectively, by gene-based association analysis. The first gene encodes a kinase-related protein with unknown function, which is likely to act as a regulator to influence the genes involved in oil biosynthesis and metabolism pathway. While the second gene encodes fatty acid elongase 2 (fae2) and directly participates in the biosynthesis of very long-chain fatty acid in Arabidopsis, so that it can regulate the ratio of C18:0/C20:0 by affecting the content of both fatty acids in a direct way. In total, these findings provide insights into the genetic architecture of oil-related traits and an opportunity to increase oil content and improve oil quality in maize kernels.

Availability of data and materials

The datasets supporting the conclusions of this article are included within the article and its supplementary files.



Recombinant inbred line


Phenotypic variation explained


Palmitic acids


Palmitoleic acids


Stearic acids


Oleic acids


Linoleic acids


Linolenic acids


Arachidic acids


Behenic acids


Lignoceric acids


Saturated fatty acids


Unsaturated fatty acids


Best linear unbiased prediction


Quantitative trait locus/loci


logarithm of odds


Linkage disequilibrium


Days after pollination


Coefficient of variation


Minor allele frequency


  1. Moose SP, Dudley JW, Rocheford TR. Maize selection passes the century mark: a unique resource for 21st century genomics. Trends Plant Sci. 2004;9(7):358–64.

    Article  CAS  PubMed  Google Scholar 

  2. Ranum P, Pena-Rosas JP, Garcia-Casal MN. Global maize production, utilization, and consumption. In: PenaRosas JP, GarciaCasal MN, Pachon H, editors. Technical considerations for maize flour and corn meal fortification in public health: consultation rationale and summary. New York: Ann. NY Acad. Sci; 2014. p. 1–7.

    Google Scholar 

  3. Lambert RJ, Alexander DE, Mejaya IK. Single kernel selection for increased grain oil in maize synthetics and high-oil hybrid development. Plant Breed Rev. 2004;1:153–76.

    Google Scholar 

  4. Li H, Peng ZY, Yang XH, Wang WD, Fu JJ, Wang JH, et al. Yan, genome-wide association study dissects the genetic architecture of oil biosynthesis in maize kernels. Nat Genet. 2013;45(1):43–50.

    Article  CAS  PubMed  Google Scholar 

  5. Song TM, Chen SJ. Long term selection for oil concentration in five maize populations. Maydica. 2004;49:9–14.

    Google Scholar 

  6. Schnable PS, Ware D, Fulton RS, Stein CJ, Wei FS, Pasternak S, et al. The B73 maize genome: complexity, diversity, and dynamics. Science. 2009;326(5956):1112–5.

    Article  CAS  PubMed  Google Scholar 

  7. Chia J, Song C, Bradbury P, Costich D, De Leon N, Doebley J, et al. Maize HapMap2 identifies extant variation from a genome in flux. Nat Genet. 2012;44(7):803–7.

    Article  CAS  PubMed  Google Scholar 

  8. Bukowski R, Guo X, Lu Y, Zou C, He B, Rong Z, et al. Construction of the third-generation Zea mays haplotype map. Gigascience. 2018;7(4):1–12.

    Article  PubMed  Google Scholar 

  9. Goldman IL, Rocheford TR, Dudley JW. Molecular markers associated with maize kernel oil concentration in an Illinois high protein × Illinois low protein cross. Crop Sci. 1994;34(4):908–15.

    Article  Google Scholar 

  10. Alrefai R, Berke TG, Rocheford TR. Quantitative trait locus analysis of fatty acid concentrations in maize. Genome. 1995;38(5):894–901.

    Article  CAS  PubMed  Google Scholar 

  11. Song XF, Song TM, Dai JR. QTL mapping of kernel oil concentration with high-oil maize by SSR markers. Maydica. 2004;49:41–8.

    Google Scholar 

  12. Clark D, Dudley JW, Rocheford TR, Ledeaux JR. Genetic analysis of corn kernel chemical composition in the random mated 10 generation of the cross of generations 70 of IHO × ILO. Crop Sci. 2006;46(2):807–19.

    Article  Google Scholar 

  13. Wassom JJ, Mikkelineni V, Bohn MO, Rocheford TR. QTL for fatty acid composition of maize kernel oil in Illinois high oil × B73 backcross-derived lines. Crop Sci. 2008;48(1):69–78.

    Article  CAS  Google Scholar 

  14. Wassom JJ, Wong JC, Martinez E, King JJ, Debaene J, Hotchkiss JR, et al. QTL associated with maize kernel oil, protein, and starch concentrations; kernel mass; and grain yield in Illinois high oil × B73 backcross-derived lines. Crop Sci. 2008;48(1):243–52.

    Article  Google Scholar 

  15. Yang XH, Guo YQ, Yan JB, Zhang J, Song TM, Rocheford T, et al. Major and minor QTL and epistasis contribute to fatty acid compositions and oil concentration in high-oil maize. Theor Appl Genet. 2010;120(3):665–78.

    Article  CAS  PubMed  Google Scholar 

  16. Fang H, Fu XY, Wang YB, Xu J, Feng HY, Li WY, et al. Genetic basis of kernel nutritional traits during maize domestication and improvement. Plant J. 2019;101:278–92.

    Article  Google Scholar 

  17. Cook JP, Mcmullen MD, Holland JB, Tian F, Bradbury P, Ross-Ibarra J, et al. Genetic architecture of maize kernel composition in the nested association mapping and inbred association panels. Plant Physiol. 2012;158(2):824–34.

    Article  CAS  PubMed  Google Scholar 

  18. Li-Beisson Y, Shorrosh B, Beisson F, Andersson MX, Arondel V, Bates PD, et al. Acyl-lipid metabolism. Arabidopsis Book. 2013;11:e0161.

    Article  PubMed  PubMed Central  Google Scholar 

  19. Belo A, Zheng P, Luck S, Shen B, Meyer DJ, Li B, et al. Whole genome scan detects an allelic variant of fad2 associated with increased oleic acid levels in maize. Mol Gen Genomics. 2008;279(1):1–10.

    Article  CAS  Google Scholar 

  20. Zheng P, Allen WB, Roesler K, Williams ME, Zhang S, Li J, et al. A phenylalanine in DGAT is a key determinant of oil content and composition in maize. Nat Genet. 2008;40(3):367–72.

    Article  CAS  PubMed  Google Scholar 

  21. Shen B, Allen WB, Zheng P, Li C, Glassman K, Ranch J, et al. Expression of ZmLEC1 and ZmWRI1 increases seed oil production in maize. Plant Physiol. 2010;153(3):980–7.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  22. Li L, Li H, Li Q, Yang XH, Zheng DB, Warburton M, et al. An 11-bp insertion in Zea mays fatb reduces the palmitic acid content of fatty acids in maize grain. PLoS One. 2011;6(9):e24699.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. Chai YC, Hao XM, Yang XH, Allen WB, Li JM, Yan JB, et al. Validation of DGAT1-2 polymorphisms associated with oil content and development of functional markers for molecular breeding of high-oil maize. Mol Breed. 2011;29:939–49.

    Article  Google Scholar 

  24. Han YJ, Xu G, Du HW, Hu JY, Liu ZJ, Li H, et al. Natural variations in stearoyl-acp desaturase genes affect the conversion of stearic to oleic acid in maize kernel. Theor Appl Genet. 2017;130(1):151–61.

    Article  CAS  PubMed  Google Scholar 

  25. Yang XH, Gao SB, Xu ST, Zhang ZX, Prasanna BM, Li L, et al. Characterization of a global germplasm collection and its potential utilization for analysis of complex quantitative traits in maize. Mol Breed. 2010;28:511–26.

    Article  Google Scholar 

  26. Wang Q, Li K, Hu XJ, Shi HM, Liu ZF, Wu YJ, et al. Genetic analysis and QTL mapping of stalk cell wall components and digestibility in maize recombinant inbred lines from B73 × By804. Crop J. 2020;8(1):132–9.

    Article  CAS  Google Scholar 

  27. Ganal MW, Durstewitz G, Polley A, Berard A, Buckler ES, Charcosset A, et al. A large maize (Zea mays L.) SNP genotyping array: development and germplasm genotyping, and genetic mapping to compare with the B73 reference genome. PLoS One. 2011;6(12):e28334.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  28. Pan QC, Li L, Yang XH, Tong H, Xu ST, Li ZG, et al. Genome-wide recombination dynamics are associated with phenotypic variation in maize. New Phytol. 2016;210(3):1083–94.

    Article  CAS  PubMed  Google Scholar 

  29. Wang S. Windows QTL cartographer 2.5. WWW document. 2007.

    Google Scholar 

  30. Lander ES, Botstein D. Mapping mendelian factors underlying quantitative traits using RFLP linkage maps. Genetics. 1989;121(1):185–99.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Wen WW, Liu HJ, Zhou Y, Jin M, YangN LD, et al. Combining quantitative genetics approaches with regulatory network analysis to dissect the complex metabolism of the maize kernel. Plant Physiol. 2016;170(1):136–46.

    Article  CAS  PubMed  Google Scholar 

  32. Liu HJ, Luo X, Niu LY, Xiao YJ, Chen J, Liu J, et al. Distant eQTLs and non-coding sequences play critical roles in regulating gene expression and quantitative trait variation in maize. Mol Plant. 2017;10(3):414–26.

    Article  CAS  PubMed  Google Scholar 

  33. Yu JM, Pressoir G, Briggs WH, Vroh Bi I, Yamasaki M, Doebley JF, et al. A unified mixed-model method for association mapping that accounts for multiple levels of relatedness. Nat Genet. 2006;38(2):203–8.

    Article  CAS  PubMed  Google Scholar 

  34. Alexander DH, Novembre J, Lange K. Fast model-based estimation of ancestry in unrelated individuals. Genome Res. 2009;19(9):1655–64.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Bradbury PJ, Zhang Z, Kroon DE, Casstevens TM, Ramdoss Y, Buckler ES. TASSEL: software for association mapping of complex traits in diverse samples. Bioinformatics. 2007;23(19):2633–5.

    Article  CAS  PubMed  Google Scholar 

  36. Fu JJ, Cheng YB, Linghu JJ, Yang XH, Kang L, Zhang ZX, et al. RNA sequencing reveals the complex regulatory network in the maize kernel. Nat Commun. 2013;4:2832.

    Article  Google Scholar 

  37. Chen J, Zeng B, Zhang M, Xie SJ, Wang GK, Hauck A, et al. Dynamic transcriptome landscape of maize embryo and endosperm development. Plant Physiol. 2014;166(1):252–64.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. Blacklock BJ, Jaworski JG. Substrate specificity of Arabidopsis 3-ketoacyl-CoA synthases. Biochem Biophys Res Commun. 2006;346(2):583–90.

    Article  CAS  PubMed  Google Scholar 

  39. Zhang J, Lu XQ, Song XF, Yan JB, Song TM, Dai JR, et al. Mapping quantitative trait loci for oil, starch, and protein concentrations in grain with high-oil maize by SSR markers. Euphytica. 2007;162:335–44.

    Article  Google Scholar 

  40. Wurschum T, Maurer HP, Dreyer F, Reif JC. Effect of inter- and intragenic epistasis on the heritability of oil content in rapeseed (Brassica napus L.). Theor Appl Genet. 2013;126:435–41.

    Article  Google Scholar 

  41. Rani PJ, Satyanarayana PV, Chamundeswari N, Ahamed ML, Rani MG. Research article studies on genetic control of quality traits in rice (oryza sativa L.) using six parameter model of generation mean analysis, electronic J. Plant Breed. 2015;6:658–62.

    Google Scholar 

  42. Wang L, Yang X, Cui S, Mu G, Sun X, Liu L, et al. QTL mapping and QTL × environment interaction analysis of multi-seed pod in cultivated peanut (Arachis hypogaea L.). Crop J. 2019;7:249–60.

    Article  CAS  Google Scholar 

  43. Santantonio N, Jannink JL, Sorrells M. A low resolution epistasis mapping approach to identify chromosome arm interactions in allohexaploid wheat. G3-genes Genom. Genet. 2019;9:675–84.

    CAS  Google Scholar 

  44. Chen YS, Lubberstedt T. Molecular basis of trait correlations. Trends Plant Sci. 2010;15(8):454–61.

    Article  CAS  PubMed  Google Scholar 

  45. Mao HD, Wang HW, Liu SX, Li ZG, Yang XH, Yan JB, et al. A transposable element in a NAC gene is associated with drought tolerance in maize seedlings. Nat Commun. 2015;6:1–13.

    CAS  Google Scholar 

  46. Wang XL, Wang HW, Liu SX, Ferjani A, Li JS, Yan JB, et al. Genetic variation in ZmVPP1 contributes to drought tolerance in maize seedlings. Nat Genet. 2016;48(10):1233–41.

    Article  CAS  PubMed  Google Scholar 

  47. Wang C, Yang Q, Wang WX, Li YP, Guo YL, Zhang DF, et al. A transposon-directed epigenetic change in ZmCCT underlies quantitative resistance to Gibberella stalk rot in maize. New Phytol. 2017;215(4):1503–15.

    Article  CAS  PubMed  Google Scholar 

  48. Huang C, Sun HY, Xu DY, Chen QY, Liang YM, Wang XF, et al. ZmCCT9 enhances maize adaptation to higher latitudes. Proc Natl Acad Sci U S A. 2018;115(2):E334–41.

    Article  CAS  PubMed  Google Scholar 

  49. Tian JG, Wang CL, Xia JL, Wu LS, Xu GH, Wu WH, et al. Teosinte ligule allele narrows plant architecture and enhances high-density maize yields. Science. 2019;365(6454):658–64.

    Article  CAS  PubMed  Google Scholar 

  50. Zhang ZY, Li JJ, Pan YH, Li JL, Zhou L, Shi HL, et al. Natural variation in CTB4a enhances rice adaptation to cold habitats. Nat Commun. 2017;8(1):14788.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  51. Ma Y, Dai XY, Xu YY, Luo W, Zheng XM, Zeng DL, et al. COLD1 confers chilling tolerance in rice. Cell. 2015;160(6):1209–21.

    Article  CAS  PubMed  Google Scholar 

  52. Hua L, Wang DR, Tan LB, Fu YC, Liu FX, Xiao LT, et al. LABA1, a domestication gene associated with long, barbed awns in wild rice. Plant Cell. 2015;27(7):1875–88.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  53. Gadaleta A, Colasuonno P, Giove SL, Blanco A, Giancaspro A. Map-based cloning of QFhb.mgb-2A identifies a WAK2 gene responsible for Fusarium Head Blight resistance in wheat. Sci Rep. 2019;9:6929.

    Article  Google Scholar 

  54. He H, Zhu S, Zhao R, Jiang Z, Ji Y, Ji J, et al. Pm21, encoding a typical CC-NBS-LRR protein, confers broad-spectrum resistance to wheat powdery mildew disease. Mol Plant, 2018;11(6):879–82.

  55. Zhang YP, Wang LQ, Zuo DY, Cheng HL, Liu K, Ashraf J, et al. Map-based cloning of a recessive gene v1 for virescent leaf expression in cotton (Gossypium spp.). J Cotton Res. 2018;1:1–9.

    Article  Google Scholar 

  56. Deng ZH, Li X, Wang ZZ, Jiang YF, Wan LL, Dong FM, et al. Map-based cloning reveals the complex organization of the BnRf locus and leads to the identification of BnRfb, a male sterility gene, in Brassica napus. Theor Appl Genet. 2016;129(1):53–64.

    Article  CAS  PubMed  Google Scholar 

  57. Liu J, Huang J, Guo H, Lan L, Wang HZ, Xu YX, et al. The conserved and unique genetic architecture of kernel size and weight in maize and rice. Plant Physiol. 2017;175(2):774–85.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  58. Pan QC, Xu YC, Li K, Peng Y, Zhan W, Li WQ, et al. The genetic basis of plant architecture in 10 maize recombinant inbred line populations. Plant Physiol. 2017;175(2):858–73.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  59. Liu R, Lu J, Zhou M, Zheng SG, Liu ZH, Zhang CH, et al. Developing stripe rust resistant wheat (Triticum aestivum L.) lines with gene pyramiding strategy and marker-assisted selection. Genet. Resour. Crop Ev. 2020;67(2):381–91.

    Article  CAS  Google Scholar 

  60. Zong G, Wang A, Wang L, Liang GH, Gu MH, Sang T, et al. A pyramid breeding of eight grain-yield related quantitative trait loci based on marker-assistant and phenotype selection in rice (Oryza sativa L.). J Genet Genomics. 2012;39:335–50.

    Article  CAS  Google Scholar 

  61. Yao D, Wang PW, Yan W, Zhang Y, Qu J, Zhang J. Marker assistant selection and soybean oil content by QTL location using inclusive composite interval mapping. Chin J of Oil Crop Sci. 2010;3:369–73.

  62. Hao XM, Li XW, Yang XH, Li JS. Transferring a major QTL for oil content using marker-assisted backcrossing into an elite hybrid to increase the oil content in maize. Mol Breed. 2014;34(2):739–48.

    Article  CAS  Google Scholar 

Download references


We greatly appreciate Dr. Xiaohong Yang at China Agricultural University for sharing the RIL population and giving valuable suggestions to improve this manuscript.


This work was supported by the National Key R&D Program of China (2017YFD0101104), the Social Livelihood Science and Technology Project of Nantong City, China (MS22020033), and the Nantong University Scientific Research Start-up project for Introducing Talents (135420609055).

Author information

Authors and Affiliations



HF and XYF carried out the experiments, analyzed the data and wrote the manuscript; HQG, AXZ and TYS assisted in data collection and analysis; and YDW, PL and BHW designed the study and helped write or revise the manuscript. All authors have read and approved this manuscript. All authors reviewed the manuscript.

Corresponding authors

Correspondence to Yuandong Wang, Ping Li or Baohua Wang.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1: Figure S1.

Recombination bin map of 180 lines in the KUI3/SC55 RIL population. Figure S2. Phenotypic distributions of 19 oil-related traits in the KUI3/SC55 RIL population. Figure S3. Pearson correlation coefficients (upper right) for the oil-related traits and -log10 (P-value) of the Pearson correlation (bottom left). Figure S4. LOD profiles for the QTL clusters that colocalized with previously cloned genes. Figure S5. Functional category annotations for 19 colocalized genes. Figure S6. Heat map of gene expression for 19 colocalized genes in developing embryo, endosperm, and seed at various developing stages.

Additional file 2: Table S1.

Single QTLs for 19 oil-related traits identified in this study.

Additional file 3: Table S2.

Epistasis interactions between pairs of QTLs with additive effects.

Additional file 4: Table S3.

Associations between GRMZM2G101515, GRMZM2G022558 polymorphisms and two oil-related traits in 508 maize inbred lines.

Additional file 5: Table S4.

The functional annotation and the RPKM values of gene expression for the 19 colocalized genes.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit The Creative Commons Public Domain Dedication waiver ( applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Fang, H., Fu, X., Ge, H. et al. Genetic basis of maize kernel oil-related traits revealed by high-density SNP markers in a recombinant inbred line population. BMC Plant Biol 21, 344 (2021).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: