​​Phenotypic Identification and Diversity of Mung Bean Germplasm in Liaoning, China

Mung bean (Vigna radiata L.) is an important cultivated species in the Phaseoleae family of Leguminosae (Papilionaceae) (Long et al., 1989). It has been developed and domesticated as an important food crop for a long time. China is one of the original centers of mung bean, with a cultivation history of more than 2000 years (Zheng,1997; Cai and Wang, 2007). Mung bean is mainly distributed in Northeast China, North China Plain and the Huang and Huai River Basin. Mung bean has a short growth period, long suitable sowing period, drought resistance, barren tolerance, wide adaptability, nitrogen fixation and other characteristics. Alternatively, mung bean can be used as a grain and vegetable, food and medicine, with high economic value.
       
Germplasm resources research is the basis in any crop research. In recent years, especially after establishing the National Edible Bean Industry and Technology System, China had intensified research on edible bean germplasm resources. In terms of mung bean research, Chen et al., (2020) and Qiao et al., (2015) collected hundreds of mung bean resources from home and abroad, analyzed and evaluated their agronomic traits and genetic diversity and it is believed that mung bean has  high variability for most of the phenotypic traits and selected germplasm selections possess excellent characteristics and superior for most of the traits. A total of 7067 mung bean germplasm resources were collected, preserved and catalogued in the National Germplasm Bank of China and most of them were identified for main agronomic traits, which were included in the Catalog of Edible Bean Germplasm resources of China and the core germplasm of mung bean was constructed (Liu et al., 2008). Based on this, Wang et al., (2014) evaluated the phenotypic variation and ecological adaptability of mung bean core germplasm resources and believed that the growth habit and pod-setting habit of mung bean were significantly affected by the ecological environment. Simultaneously, Bisht et al., (1998) also made a comprehensive evaluation of mung bean resources in India and constructed the core germplasm population. With the progress of technology, molecular biology technology has been applied to genetic diversity analysis of mung bean resources, Chen et al., (2015) compared 157 cultivated and wild germplasms from East Asian regions by SSR markers and found that the polymorphism was low and hence there was a need to collect resources widely. Singh et al., (2009) compared 12 traits in mung bean resources in three ecological areas and  clustered 80 resources into 11 groups, indicating that mung bean resource panel has a rich diversity. Kumar et al., (2012) analyzed the diversity of 48 mung bean resources with different origins using ISSR, URP and SSR and perceived that URP technology was more applicable to the analysis of interspecific and intraspecific diversity analyses and it was necessary to expand the genetic basis of cultivated mung bean by using wild and closely related species. Zhao et al., (2022) constructed DNA fingerprinting based on 23 mung bean (Vigna radiata L.) germplasm, all resources could be distinguished by digital fingerprint. With the development of statistics, multiple statistical analysis methods have been comprehensively adopted to analyze the relationship between mung bean resource traits (Chen et al., 2020; Hou et al., 2020), especially the relationship between traits and yield (Zhu et al., 2015) and abundant achievements have been reported.
       
Liaoning is located in Northeast China and the eastern part is the remnant of Changbai Mountain, a mountainous and hilly region; the central plain is an important high-quality grain-producing area; the west borders on Inner Mongolia, arid areas for sand with barren soil. Mung beans are mainly distributed in the western part and eastern mountain areas. Mung beans produced in the western part of Liaoning Province have been widely accepted in the domestic and foreign markets for their excellent quality. In this study, 170 mung bean germplasms were comprehensively evaluated, with the objective to fully understand the basic situation of mung bean germplasm resources in Liaoning Province and to provide a theoretical basis for collection, preservation, innovative and usage of mung bean germplasm resources.

A total of 170 mung bean resources were selected, which are being preserved in the Germplasm bank of Liaoning Academy of Agricultural Sciences. The materials are from Liaoning (147), Inner Mongolia (18) and five materials from the Germplasm Bank of Beijing Agricultural University. Fig 1 showed the resources distribution in Liaoning.
 

       
The experiment was conducted at the experimental site of Industrial Crops Institute of Liaoning (Liaoyang City, Liaoning Province) from 2015 to 2019. The soil is sandy loam, with 1.97% organic matter, 0.08% total nitrogen, 73.4 mg/kg alkali-hydrolyzable nitrogen, 23.6 mg/kg available phosphorus and 247.5 mg/kg available potassium in the top soil. Following the order of obtaining resources and the conditions of the test site, all resources were planted in succession, ensuring that each material was planted for at least two years. During planting, the materials were arranged in sequence. Each material was planted in two rows with 5 m length, 50 cm row spacing and 8-10 cm plant spacing, without repetition. The sowing time is around June 25^th and harvested around October 5^th, three-compound fertilizer (N:P:K=15%:15%:15%) was applied as the base fertilizer disposably; ploughing and weeding were conducted thrice during the crop period.
       
Following the Descriptors and Data Standard for Mungbean [Vigna radiata (L.) Wilczek] compiled by Cheng et al., (2006), 10 plants with uniform growth were selected for each accession and 37 qualitative and 18 quantitative traits were investigated during the whole growth period. Twelve quantitative traits, including growing days, plant height, yield and 34 quality traits, including mature habit, color and growth habit, were selected for statistics. The mean value of 10 plants and inter-annual mean value were obtained.
       
EXCEL 2007 and DPS 9.5 were used for data collection and analysis. First, the diversity of qualitative and quantitative traits was analyzed and the Shannon diversity index was analyzed by referring to the method of Wang et al., (2021). Then, principal component analysis was performed based on quantitative trait indexes. After standardizing all quantitative 46 traits, the grey relational  analysis was conducted. Based on all the traits (including quality and quantitative traits), systematic cluster analysis was conducted through standardized transform-Euclidean distance-deviation square sum method and further K-mean cluster analysis was conducted on the basis of quantitative traits.

Diversity analysis of mung bean germplasm resources
 
Among the 34 tested qualitative traits, nine exhibited the same characteristics, i.e., 70-90 d growth period, green unearthed cotyledon, single green leaf, whole margin three-compound leaves, infinite podding, mature pod brown with cylindrical shape and brown pod fuzz. Twenty five qualitative traits exhibited different characteristics, of which 12 traits exhibited all the characteristics listed in criteria (Cheng et al., 2006) and 13 traits exhibited partial characteristics. Among the traits, grain color was the most complex, showing 4 of the 5 characteristics. Leaf shape, leaf color, plant fuzz growth conditions and color showed an obvious tendency, whereas growth habit, pod splitting and grain size were scattered.
       
For the 12 quantitative traits, the coefficient of variation (CV) of all resources in this experiment ranged from 2.44% to 60.10%. CV of plant yield (PY), pod number per plant (PNP) and main stem branch number (BRN) were over 30%, showing that the variation range of these three traits was large and the potential of genetic improvement was large, the same result was found by Kumar et al., (2020) in faba bean and Saidaiah et al., (2021) in cow bean. CV of growth period (GP) was the lowest, indicating that the natural climate conditions in this region limited the growing period of mung bean resources. Shannon diversity index (DI) ranged from 0.979 to 2.054. Plant height (PH), leaf size (LS), growth period (GP) and 100-grain weight (HSW) recorded the higher Shannon DI, showing that these four traits had the most abundant phenotypes and the largest range of parents to choose from. The DI of seed length (SL) was the smallest at only 0.979, while the CV was also small (15.19%); as mung bean grains were smaller, so the amplitude of variation was small and the selection range was limited. 
       
Comparison of CV and Shannon DI of mung bean resources, for 46 tested traits showed rich diversity, Win et al., (2020) got the same result while Saidaiah et al., (2021) got similar result on cowpea. GP (the number of growing days) and some traits showed strong adaptability through natural and artificial selection in the long-term evolution process and their potential for genetic improvement was negligible, which was in agreement with the earlier results of Gao et al., (2020). However, the traits showing remarkable differences and rich diversity are either directly related to yield or indirectly affecting yield on behalf of plant growth and development, which indicates that mung bean resources still have great potential for yield improvement.
 
Correlation analysis of quantitative traits
 
Except for GP and HSW, all other traits were positively correlated with plant yield (PY), of which the most significant correlated traits were PNP, followed by PH, BRN and main stem diameter (MSD). It can be seen that pod-bearing capacity and plant development state had great impaction on yield. Comparing the correlation coefficients of other essential traits, with an increase in GP, stem node number (SNN) increases, but the pod bearing capacity decreased, which notably led to a decrease in yield. The resources with vigorous growth and bigger pods and seeds have bigger HSW, while the increase of pods number per plant (PNP) inevitably led to smaller pods and seeds, but seeds number per pod (NSP) increased accordingly. There was a significant positive correlation between seed length (SL) and NSP, while a significant positive correlation between each index of vegetative growth of the plant, these results were in line with expectations.
 
Grey relational analysis of quantitative traits
 
After standardized processing , grey relational analysis was conducted in the measured values of 12 quantitative traits of mung bean resources, The results indicate that pod-bearing capacity and plant development ability were closely correlated with yield. Generally, the results of grey relational analysis and correlation analysis were consistent, i.e., the yield was mainly limited by the pod-bearing capacity of the plant and was directly related to the vegetative growth of the plant, whereas the impact of GP and seed size (SS) on the yield was limited.
 
Principal component analysis of quantitative traits
 
Five principal components with eigen values bigger than or close to 1 were selected. By comparing the eigen vectors of each principal component, it could be named as the HSW factor, GP factor, PL factor, PH factor and NSP factor. The first, third and fifth components were directly related to yield and more emphasis should be done in these traits while selecting high-yielding varieties. The fourth component represents plant development, which is indirectly related to yield should also be considered during selection. The second component represents the period of growth and can be utilized according to the specific situation.
       
Combining the results of the correlation studies and grey relational analysis, it can be concluded that the yield of mung beans is mainly determined by the pod-bearing capacity and growth status of the plant, The results of the principal component analysis also support this conclusion. This conclusion is consistent with the research findings of Hou et al., (2015) and Yang et al., (2015).
 
Cluster analysis and evaluation
 
Based on all the 46 traits with different performance (12 quantitative traits + 34 quality traits) of mung bean resources, this study adopted a standardized data transformation-Euclidean distance-deviation square sum method to conduct cluster analysis and all the accessions were divided into seven groups at Euclidean distance = 27.59 (Fig 2).
 

       
The first group consists of 36 samples, which were later-maturing, dwarf and low-yield. The second group consists of 49 samples, which recorded big-pod and big-seed size and strong-stem. The third group represented 11 accessions which were late-maturing, multi-nodule and small-seeded. The fourth group contained nine samples, which were dwarf, big-podded and low-yielding types. The fifth group contains 47 samples, which were late-maturing and with small-pods, small-seeds. The sixth group contains four samples, which were early-maturing accessions with big-grain and high-yield. Finally, the seventh group comprised 14 samples, with big-leaf, big-grain and high-yield.
       
All groups showed two kinds of compound leaves, i.e. triangular and oval shapes with relatively uniform distribution and two kinds of seed skin characteristics namely  bright and grey and mainly bright. The other traits exhibited differences among different groups. Group II showed the main characteristics of resources in this region. Groups III and VI exhibited more uncommon characteristics. The traits of group V showed abundant characteristics. Group VII is a special group, with certain traits performance in contrary to other groups.
 
K-mean clustering analysis
 
The theoretical distribution intervals and central values of the 12 quantitative traits were set  for the experimental material. The traits of germplasm studied in this experiment were mainly concentrated near the central value, which was also the main characteristic of mung bean germplasm resources in Liaoning Province. Based on the results, No. 130 resource was outstanding for yield and its yield potential is good. No. 149 resources combined high yield and earliness in maturity. Simultaneously, four large-grain resources, 15 early-maturing resources and multi-pods resources, high-stems resources, lobules resources, as well as other resources with outstanding usable characteristics were screened out, which could be referenced in breeding (Table 1).
 

In summary, mung bean resources in Liaoning have rich diversity in phenotype and there are close correlations between different traits. Based on the results of multivariate analyses, it was concluded that pod-bearing capacity and vegetative growth ability were closely related to yield.

Therefore, with respect to high yield breeding, germplasm resources with more pods per plant, bigger pod size, bigger seeds and vigorous growth should be selected for further utilization. Furthermore, resources with outstanding usable characteristics were screened out, include one high-yield resource from Beipiao and one early-maturing high-yield resource from Tieling of Liaoning Province, provides a good chance for the selection of parents for the improvement program.