Waxy corn (Zea mays L. sinensis Kulesh) is a subspecies of Zea mays L., whose ears can be eaten fresh, and whose seeds can also be used for amylopectin processing or livestock feed together with straw (Wang, 2007; Li et al., 2010). With the improvement of people’s living standards, people’s demand for food quality and nutrition is increasing, and the planting area of waxy corn is increasing year by year, which not only greatly promotes the development of waxy corn industry, but also promotes the development of waxy corn breeding. However, breeding work is inseparable from the necessary germplasm resources. Germplasm resources are the basis of breeding work, and the collection and sorting of germplasm resources is the premise of breeding work. Although China is the center of genetic diversity and origin of waxy corn, which has extremely rich local germplasm resources (Chen et al., 2013), affected by the breeding system and commercial interests, there is a lack of necessary germplasm exchange among breeding units, resulting in the narrow genetic basis of germplasm resources of waxy corn, which greatly reduces the breeding probability of breakthrough dominant varieties. Therefore, it is of great significance to actively collect and sort out germplasm resources of waxy corn in various regions of China and to expand, improve and innovate these germplasm resources. At present, there are a large number of waxy corn varieties bred in various parts of China, and the source of varieties is wide. However, there are relatively few studies on diversity analysis and variety group analysis. How to collect, analyze and utilize these variety resources is worthy of further research.

The trait difference of different crop varieties is the key basis for their selection and utilization, and accurate observation and discovery of this difference is the premise of production and utilization (Long et al., 2019). In addition to selecting high-yield varieties for waxy corn planting, the differences among multiple traits such as the growth period of the variety, plant traits, ear commerciality should also be considered, and other trait-related varieties should be selected according to actual production needs. From the analysis of the production benefit of fresh-eating corn, varieties with an earlier growth period should be selected. These varieties can be marketed earlier to increase the sales price, and the crop stubble can be better arranged, and the multiple cropping index should be increased (Wang et al., 2017); According to the commercial requirement of ear, the varieties with small ear barren tip should also be selected, because corn lodging has a significant correlation with the plant height and ear height (Ren et al., 2016; Li et al., 2019; Tang et al., 2020). From the analysis of production cost, corn varieties with lower plant height and ear height should be selected, which can effectively reduce corn lodging, management cost and yield loss. Because the stems and leaves of waxy corn remain green when harvested, it can be used as high-quality feed for livestock. Therefore, the varieties with high plants can also be selected under suitable planting conditions.

In this study, the main agronomic traits of 115 waxy corn varieties collected from 17 provinces and cities in China were carried out statistical analysis and diversity analysis, the genetic variation of the tested varieties was understood, the interaction between yield traits and other agronomic traits was analyzed through correlation analysis, regression analysis and path analysis, and the characteristics of various varieties were comprehensively studied through cluster analysis, which finally provided a reference basis for the selection and utilization of waxy corn varieties.

1 Results and Analysis

1.1 Summary analysis of main traits of 115 waxy corn varieties

The average values of the original data of 9 main agronomic traits of 115 tested varieties were statistically analyzed to obtain the max, min, average and standard deviation of 115 tested materials, and the coefficient of variation and genetic diversity were calculated (Table 1). The results showed that there were great differences in different traits of different materials, and the variation was rich. The ear barren tip length showed the largest coefficient of variation, which was 399.91%, while the growth period showed the smallest coefficient of variation, which was 2.56%. The coefficient of variation from large to small was as follows: ear barren tip length>ear height>rows per ear>ear yield > kernels per row > plant height>ear length>ear diameter>growth period. The genetic diversity analysis showed that the highest was plant height (2.069) and the lowest was ear barren tip length (0.621), from large to small was as follows: plant height>ear height>ear yield>kernels per row>growth period>ear length>ear diameter>rows per ear>ear barren tip length. The results showed that there was no consistency between the coefficient of variation and the genetic diversity. For example, the coefficient of variation of ear barren tip length was the largest, but the genetic diversity of that was the lowest. The coefficient of variation of plant height was low, but the genetic diversity of that was the highest.

1.2 Correlation analysis among traits

The growth period has a very significant positive correlation with plant height, ear height and kernels per row (Table 2), and has a significant positive correlation with rows per ear; The plant height has a very significant positive correlation with growth period, ear height and ear weight, and has a significant positive correlation with kernels per row; The ear height has a very significant positive correlation with growth period, plant height and rows per ear; The ear length has a significant positive correlation with ear diameter, and has a very significant positive correlation with kernels per row, ear barren tip length and ear yield; The ear diameter has a significant positive correlation with ear length and rows per ear, and has a very significant positive correlation with ear weight; The rows per ear has a significant positive correlation with growth period and ear diameter, and has a very significant positive correlation with ear height, and has a significant negative correlation with kernels per row; The kernels per row has a very significant positive correlation with growth period and ear length, and has a significant positive correlation with plant height and ear weight, and has a significant negative correlation with rows per ear; The ear barren tip length has a significant positive correlation with ear length, but not with other traits; The ear weight has a very significant positive correlation with plant height, ear length, ear diameter and kernels per row.

1.3 Effects of main agronomic traits on ear yield

Because the simple correlation analysis can only reflect the interaction between the two indexes, and can not reflect the effect of each trait index on ear yield, it is necessary to make further regression analysis. In the multiple linear regression analysis, the independent variables with no significant effect were gradually eliminated, so that the obtained multiple regression equation would be relatively simplified and could accurately analyze and predict the response of dependent variables (Tang and Feng, 2007). Taking ear yield (Y) as dependent variable and other characters as independent variables, the following optimal linear regression equation was obtained by backward stepwise regression with DPS software: Y=-180.8222+0.4585X₂+7.4503X₄+48.7657X₅.

Through regression analysis, when the determination coefficient R²=0.5253, it reached a very significant level, and when Durbin Watson statistics d=1.9406, close to 2, the partial correlation coefficients of each regression coefficient reached a very significant level, indicating that the regression equation had regression significance. The above linear regression equation showed that ear yield had a significant linear regression relationship with plant height, ear diameter and ear length, but had no significant regression relationship with growth period, ear height, rows per ear, kernels per row and ear barren tip length. In order to study the influence degree of each significant effect factor on yield, path analysis was carried out (Table 3). In the direct effect, ear diameter had the greatest direct effect on ear yield, followed by ear length and plant height. The indirect effect was much smaller than the direct effect, so no further analysis was needed.

1.4 Cluster analysis

Cluster analysis was carried out on 115 corn varieties by Euclidean distance and the furthest neighbor method (Tang and Feng, 2007) (Figure 1). Taking 55 as the boundary, 115 varieties can be divided into 8 groups, and the average parameters of each group can be obtained (Table 4).

The average yield of group Ⅰ varieties ranked fifth. Many varieties showed long ear barren tip, and other traits are mostly at the middle level of various groups. This group includes 24 varieties, namely Jinyunuo 9, Yuxiangnuo 6, Nongkeyu 368, Yuebainuo 6, Jianuo 26, Zhenuoyu 7, Hangnuoyu 21, Nongkenuo 336, Tiannuo 828, Xiantiannuo 868, Qianjiangnuo 3, Heitiannuo 168, Meiyu 27, Shenhongtiannuo 1, Jinnuo 628, Cuinuo 5, Jingkenuo 569, Wannuo 161, Huitiannuo 2, Caitiannuo 168, Jinnuo 1813, Yunongjingnuo, Guixiangnuo 258, Sukenuo 1601, accounting for 20.9% of the tested varieties.

The average yield of group Ⅱ varieties ranked third. The average growth period of theses varieties is short, and the plant height and ear height are low. From the analysis of comprehensive characters, this kind of variety is an ideal corn type. This group includes 9 varieties, namely Jingkenuo 2016, Futiannuo 2, Yunonghuacainuo 7, Xiangtiannuo 828, Jingheng 1, Wannongtiannuo 158, Lingtiannuo 100, Xintiannuo 88 and Jiahua 6, accounting for 7.8% of the tested varieties.

The average yield of group Ⅲ varieties ranked fourth. The average growth period of theses varieties is long, and the average plant height and ear height are high. This group includes 16 varieties, namely Huitiannuo 810, Tiannuo 182, Huitiannuo 1, 2008 Jingnuo, Tianguinuo 161, Wancainuo 3, Yunuo 930, Chunqiang 8, Zhenuoyu 10, Shennuo 10, Wancaitiannuo 118, Shengkenuo 4016, Wannuo 158, Caitiannuo 100, Shengkenuo 598 and Xidanuo 2, accounting for 13.9% of the tested varieties.

The average yield of group Ⅳ varieties ranked sixth. The plant height and ear height of these varieties are high, the rows per ear are also high, and other traits are at the middle level. This group includes 14 varieties, namely Minyunuo 3, Huhongnuo 1, Jinnuo 20, Suyunuo 639, Jinnuo 10, Jingcainuo, Ketiannuo 8, Meiyu 9, Zhenuoyu 23, Meiyu 16, Jiaodiannuo 517, Chengnuo 8, Sukenuo 1602 and Shengkenuo 2008, accounting for 12.2% of the tested varieties.

The average yield of group Ⅴ varieties ranked second in all groups. The ear diameter of these varieties is large, and other traits are at the middle level. This group includes 20 varieties, namely Jinnuo 1607, Kenuo 6, Zhencainuo 1, Hongyu 2, Shengyuan 231, Jinnuo 1904, Jiahua 8, Tianguinuo 932, Zhenuoyu 14, Cuitiannuo 608, Cainuo 10, Wannuo 2000, Yunongkenuo, Mihuatiannuo 12, Baili, Jinnuo 685, Jingzinuo 218, Caitiannuo 627, Chunqiangbaitiannuo 3 and Tiannuo 302, accounting for 17.4% of the tested varieties.

Group VI varieties are high-stem and large-ear varieties, with the highest plant height and ear height. The average yield of these varieties is the highest, the ear diameter is large and the growth period is the longest. This group includes 4 waxy corn varieties, namely Weinuo 6, Lingnuo 8, Zhenuoyu 16 and Huayunbaitiannuo 601, accounting for 3.5% of the tested varieties.

The indexes of each trait of group Ⅶ varieties are relatively at the lowest level. This group includes 9 varieties, namely Huiyinnuo 1, Shennuo 2, Nantiannuo 601, Zhenuoyu 18, Mintiannuo 707, Shenbaitiannuo 1, Huayunheinuo 501, Kenuo 167 and Kenuo 2, accounting for 7.8% of the tested varieties.

The average yield of group Ⅷ varieties ranked seventh in all groups. The yield of these varieties is low, the average growth period is long, the plant height is high, the ear height is high, and the ear is thin. This group includes 19 varieties, namely Xuenuo 6, Sukenuo 10, Shengkenuo 618, Meiyunuo 11, Suyunuo 11, Heitiannuo 632, Sukenuo 1505, Heitiannuo 639, Mitiannuo 1, Meiyujiatiannuo 7, Mixiangtiannuo 265, Sukenuo 12, Huziheinuo 1, Punuo 818, Yuebaitiannuo 7, Suyunuo 1502, Shennuo 8, Yunuo 17-1 and Heizhenzhu, accounting for 16.5% of the tested varieties.

2 Discussion

2.1 Selection of agronomic traits in breeding

The correlation analysis of agronomic traits found that the growth period has a very significant positive correlation with plant height, ear height and kernels per row, and has a significant positive correlation with rows per ear. In breeding, the materials with short growth period should be selected, while the varieties with too high plant height, ear height and too many kernels per row are unsuitable to choose. There was a significant positive correlation between ear barren tip and ear length. The materials with small ear barren tip should be selected, while the variety resources with too long ear are unsuitable to choose. The yield has a very significant positive correlation with plant height, ear length, ear diameter and kernels per row. The selection of these traits in breeding is an important measure to improve yield. Through regression analysis and path analysis between yield and main agronomic traits, it was found that ear diameter, ear length and plant height have a significant regression relationship with ear yield, and the degree of effect on yield from large to small was as follows: ear diameter > ear length > plant height. Although plant height has a positive effect on yield, plant height is unfavorable to lodging resistance. The selection of progeny lines centered on high yield can focus on the selection of ear diameter and ear length, or reduce the selection of ear height within an appropriate plant height range.

2.2 Comprehensive evaluation of various groups and varieties

Through cluster analysis, 115 waxy corn varieties were divided into 8 groups. The cluster grouping was not clustered according to regional location, which may be the result of artificial selection in the process of cultivation or breeding, which is basically consistent with the results of previous studies (Lin et al., 2016). The results of cluster analysis showed that the average yield of group I varieties is medium, many varieties show long ear barren tip, and other traits are mostly at the middle level of various groups; The group Ⅱ varieties have high yield, short growth period, low plant height and ear height, which are the varieties with the best comprehensive traits; The group III varieties have high yield, high plant height and ear height, and medium growth period; The group IV varieties have high plant height and ear height, medium yield and medium growth period; The group Ⅴ varieties have high yield and other traits are in the middle level; The group VI varieties have the highest yield, the longest growth period, the highest plant height and ear height. Besides, the ear diameter, ear length, rows per ear and kernels per row are at the highest level in all groups; The group VII varieties have low plant height and ear height and have the lowest yield, the shortest growth period among all groups. The group VIII varieties have long growth period, high plant height and ear height, thin ear and low yield. From the perspective of production and planting, the varieties with early growth period can better arrange stubble, and the varieties with early growth period and short plant height can better achieve higher production benefits through the measures of promoting early cultivation in greenhouse. It is suggested to adopt the group II and VII varieties, and combined with the yield advantage, it is suggested to adopt the group II varieties. For the rainy and windy areas and seasons, considering yield, it is also suitable to select the group II varieties with low plant height and ear height, which can effectively reduce the occurrence of lodging. Within the allowable range of growth period, or in the seasons or areas with low wind force, the group VI varieties with the highest yield can be planted. This kind of varieties generally has high biological yield and is suitable for both grain and feed. The planting of fresh waxy corn emphasizes the commercialization of ears. Generally, the planting density is lower than that of ordinary corn, and the size of ears can better reflect the overall yield potential of varieties. However, from the production requirements, the selection of varieties should find an optimal balance between various characters. In this study, it is considered that group II waxy corn varieties have many excellent traits such as high yield, good plant morphology and short growth period, and are the corn varieties with the best comprehensive traits.

2.3 Selection and utilization as variety resources

The large coefficient of variation of agronomic traits indicates that the genetic variation is richer (He et al., 2018). The diversity analysis showed that the collected 115 waxy corn varieties have a wide genetic basis and basically represent the integration of the latest and most dominant genes of waxy corn in China. The division of variety groups is helpful to understand the characteristics of these varieties, which can not only recommend suitable variety types for production, but also provide a basis for the selection of genetic breeding resources, for example, bicyclic selection line, haploid selection line or gene introduction line as other existing materials, or create germplasm heterosis group according to target traits, broaden germplasm genetic basis, or select target trait genes more effectively for molecular breeding. According to the selection requirements of target traits, different populations are selected for progeny lines, which has clear objectives and high selection efficiency. To select the progeny lines with large ear and high yield, we can focus on group VI; To select the progeny lines with short growth period, we can focus on group VII; To select the progeny lines with better comprehensive traits, we can focus on group II. At present, many scholars use SSR molecular markers to cluster plant germplasm resources (Zhang et al., 2016; Liu et al., 2018). This cluster analysis method well reflects the genotypic differences among various germplasm resources. Plant phenotype is the result of the interaction between genotype and environmental factors. When environmental factors have a great impact, clustering according to plant phenotype can not well reflect the genotypic differences among plant germplasm (Shi, 2001; Zhu et al., 2012), but describing and identifying plant phenotypic traits is still the most basic method in the classification of germplasm resources, This clustering method can well reflect the real differences between phenotypic traits of plant resources, and has important reference significance for the selection of comprehensive traits in crop breeding, especially in the breeding of fresh waxy corn.

3 Materials and Methods

3.1 Test materials and design

115 waxy corn varieties bred in 17 provinces and cities in China were collected (Table 5). In the spring of 2019, they were planted in Dongyang Chengdong Experimental Base of Dongyang Corn Research Institute in Zhejiang Province. The experimental sites are all for scientific research. The previous crop was rice, the land was flat and the fertility was even. The experiment was arranged in random blocks, with 6 rows, and the row length was 6.4 m, row spacing was 0.65 m, plant spacing was 0.28 m, and plot area was 25 m².

3.2 Determination of traits

The growth period and other agronomic traits were observed and recorded according to the national regional test standards. In the middle two rows of the plot, 10 representative corn plants were select to measure plant height and ear height, and 20 representative effective commercial ears were select to remove bracts to measure ear length, ear diameter, rows per ear, kernels per row, ear barren tip and ear yield, and the average value was taken.

3.3 Statistical analysis

The overall mean, standard deviation and coefficient of variation of each quantitative trait of the tested materials were calculated, and the quantitative traits were divided into 10 grades with 0.5s as the gradient (Table 2). Shannon-Wiener index (H') was used for genetic diversity analysis (Long et al., 2017; Dong et al., 2019). The calculation formula was as follows: H'=-∑P_ilnP_i, where P_i was the probability of occurrence of the trait at grade i. WPS2019 software was used to calculate data and draw charts, and DPS software was used for relevant statistical analysis and drawing analysis. The furthest neighbor method was used to cluster the tested materials by Euclidean distance.

Authors’ contributions

THP completed the experimental design, data analysis and writing the first draft of the manuscript; WGY and ZFC completed the collection of variety resources; BF and HHL participated in designing the experiments and analyzing the results; LXC guided the experimental design, data analysis, manuscript writing and revision. All authors read and approved the final manuscript.

Acknowledgments

This study was jointly funded by Special Project of National Modern Agricultural Industrial Technology System in China (CARS-02-069), National Key Research and Development Program in China (2018YFD0200706), Major Science and Technology Project of New Agricultural (Grain) Variety Breeding in Zhejiang Province (2016C02050-9-1).

References

Chen G.Q., Cheng, Y.J., Sun Q.X., Peng C.J., Chen X.H., and Xue L., 2013, Application of local germplasms in waxy corn breeding in China, Jiangxi Nongye Xuebao (Acta Agriculturae Jiangxi), 25(3): 7-11,14

Dong X., Li S.J., Yang H., Guan L., Fu Z.J., Qi Z.Y., Jin C., Yu X.Y., Yi H.H., Chen R., and Zhang P.H., 2019, Phenotypic diversity of maize landraces collected from chongqing, Zhiwu Yichuan Ziyuan Xuebao (Journal of Plant Genetic Resources), 20(4): 861-870

He L.Y., Yin C.J., Huang S.C., He Q.Y., and Shu Y.J., 2018, Correlation, clustering and principal component analysis of primary agronomic traits of vegetable soybean, Zhejiang Nongye Xuebao (Acta Agriculturae Zhejiangensis), 30(1): 50-57

Li H.C., Zhang C.L., Qu Y.Z., Cui L.Y., Mao X.L., and Lui Z.H., 2019, Study on the genetic characteristic of the traits related to lodging resistance in maize, Henan Nongye Daxue Xuebao (Journal of Henan Agricultural University), 53(1): 15-21

Li Y.J., Lui T.T., and Zhang Z.Z., 2010, The origin, research and development of waxy corn, Gengzuo Yu Zaipei (Tillage and Cultivation), (3): 52-54, 63  

Lin K.Q., Li Y., Zou B.J., Zhang T.Y., He X.Z., and Zhao D.G., 2016, Genetic Ddiversity and phylogenetic relationship of cultivated eucommia ulmoides olive in Guizhou by SSR, Jiyinzuxue Yu Yingyong Shengwuxue (Genomics and Applied Biology), 35(6): 1521-1529

Liu H.Z., Song W., Wang B.Q., Wang J.H., Zhang Q.G., ZhangD.M., Li X.H., Wei J.F., and Li R.G., 2018, Genetic diversity analysis of 120 European and American maize inbred lines, Zhiwu Yichuan Ziyuan Xuebao (Journal of Plant Genetic Resources), 19(4): 676-684

Long F., Lin Q., Zou Y., Zhao M., He H.W., Wu P., Li X.G., Wang Q.Y., Pan Y.J., and Lin G.M., 2017, Genetic diversity and correlation analysis of fruit quality characters of dwarf banana resources in Guangxi province, Jiyinzuxue Yu Yingyong Shengwuxue (Genomics and Applied Biology), 36(3): 1066-1071

Long W.J., Wu X.Y., ZHOU G.Y., Chen D., Wu S.Y., and Cai Q., 2019, Changes and diversity of quantitative traits of waxy maize landraces collected from different ecological regions of Yunnan province, Hubei Nongye Kexue (Hubei Agricultural Sciences), 58(21): 35-41

Ren B.Z., Li L.L., Dong S.T., Lui P., Zhao B., Yang J.S., Wang D.B., and Zhang J.W., 2016, Effects of plant density on stem traits and lodging resistance of summer maize hybrids with different plant heights, Zuowu Xuebao (Acta Agronomica Sinica), 42(12): 1864-1872
https://doi.org/10.3724/SP.J.1006.2016.01864

Shi G.R., 2001, Effect of different characters on classifying superior maize germplasms, Yumi Kexue (Journal of maize Sciences), 9(2): 38-40

Tang Q.Y., and Feng M.G., 2007, Dps data processing system, Science Press, Beijing, China, pp. 684-685

Tang Y.F., Chen P.P., Yi Z.X., and Luo H.B., 2020, Research progress on influencing factors and chemical regulation of maize lodging, Zuowu Yanjiu (Crop Research), 34(2): 183-189

Wang G.Y., Zhao F.C., Bao F., Han H.L., Lou X.C., and Tan H.P., 2017, Grain filling characteristics of waxy maize hybrids with different sowing dates, Fenzi Zhiwu Yuzhong (Molecular Plant Breeding), 15(9): 3811-3818

Wang Y.F., Wang L.M., Shen X.F., Zhang B., and Zheng H.J., 2007, Orlgin, classification, variety improvement and industry development of waxy corn, Hunan Nongye Daxue Xuebao (Journal of Hunan Agricultural University), 33(8): 97-102

Zhang W.W., Zhang H.W., Wang H., Zhou P., Li J., and Yu P.G., 2016, Genetic diversity analysis of 213 fresh-eaten maize inbred lines, Fenzi Zhiwu Yuzhong (Molecular Plant Breeding), 14(7): 1898-1905

Zhu H.Y., Zhong Z.Y., Li Y.X., Chen W.X., Li Y., Huang Y.K., Zheng H.L., and Wang Y.M., 2012, Correlation and cluster analyses of the traits of 102 maize inbred lines in southeastern Yunnan, Xinan Shifan Daxue Xuebao (Journal of Southwest China Normal University (Natural Science Edition), 37(4): 96-1010