A soybean seed vitality assessment method based on multi-index comprehensive evaluation

Abstract

The present invention relates to a soybean seed vitality evaluation method based on multi-index comprehensive evaluation, which belongs to the field of plant production technology. The artificial accelerated aging method was used to measure 11 seed vitality-related indicators of 126 soybean local varieties screened from a germplasm bank containing 23,587 soybean resources. Through principal component analysis, a soybean seed vitality comprehensive evaluation system was established, and on this basis, a concise and efficient seed vitality evaluation model was constructed: V=‑ 0.026+0.625×RSL+0.485×RGI. In this model, relative seedling length (RSL) and relative germination index (RGI) have a significant positive effect on seed vigor. The model provides excellent vigor assessment performance and 87.30% vigor classification accuracy. The present invention provides a new method for the scientific evaluation of soybean seed vigor, which can effectively evaluate the vigor of soybean seeds before sowing and monitor the changes in soybean seed vigor, thereby improving soybean planting efficiency and yield, and at the same time provides a valuable method for seed quality control and screening.

Description

Technical Field

[0001] The invention relates to a soybean seed vitality assessment method based on multi-index comprehensive evaluation, and belongs to the technical field of plant production. Background Art

[0002] High-vigor soybean seeds have significant growth advantages and high-yield potential, and are the key to achieving efficient production. Identification and evaluation of high-vigor germplasm is crucial to improving soybean production capacity, but there is currently no unified comprehensive evaluation system. High-quality seeds have high germination efficiency and adaptability to harsh field environments, which is a prerequisite for high and stable crop yields. Seed vigor is an important indicator for evaluating seed quality. It represents the comprehensive potential of seeds for rapid and uniform germination and normal growth of seedlings under various field conditions. High-vigor seeds are more resistant to adverse environments during germination, can germinate quickly and grow strong seedlings, thereby obtaining higher yields. Seed vigor testing is essential for evaluating seed vigor and screening germplasm, including direct and indirect measurements of current seed physiological quality, germination, seedling emergence and growth, storage tolerance and stress resistance (Marcos Filho, 2015). Germination test is a traditional method to evaluate seed vigor. Germination rate measures seed vigor, while seedling growth is evaluated based on growth-related traits, such as seedling establishment, root and shoot length, root and shoot weight, and vigor index (Wang et al., 2010; He et al., 2019; Zhang et al., 2019). Seed vigor also depends on the tolerance to long-term storage and aging damage. The aging process reduces seed vigor, so aging is a key factor causing differences in seed vigor and is the basis of seed vigor testing (Ventura et al., 2012; Finch-Savage et al., 2016). Accelerated aging tests accelerate the loss of seed vigor by exposing seeds to high temperature and humidity for a short period of time. High-vigorous seeds show stronger resistance, less loss of vigor, and can still maintain a good germination rate after aging (Mcdonald, 1999; Tekrony, 2005). At present, there are many indicators for evaluating seed vigor, mainly focusing on germination, seedling formation, stress resistance, and physiological and biochemical traits. Therefore, seed vigor is a complex comprehensive indicator.

[0003] Soybean ( Glycine max[L.] Merr.) is one of the most important oil crops in the world, accounting for more than half of global oilseed production and providing more than a quarter of the plant protein needed by humans and animals worldwide. Soybeans are high in oil and protein and are prone to aging and deterioration, especially when harvested or stored for long periods of time under high temperature and humidity conditions. Rapid loss of vigor can lead to a decrease in germination rate, stunted seedling growth, and in some cases, complete failure to germinate (Parrish et al., 1978; Wei et al., 2020). Soybean plants developed from high-vigor seeds consistently show higher yields, while low-vigor seeds often result in large yield differences between plants (Caverzan et al., 2018). Currently, accelerated aging tests and electrical conductivity methods are the main methods recommended by the International Seed Testing Association (ISTA) for soybean seed vigor testing, but a unified soybean seed vigor evaluation index has not yet been established. At present, the evaluation indexes mainly focus on germination-related traits, such as germination potential (GP), germination rate (GR), normal seedling rate (SL), germination index (GI), imbibition rate (IR), etc.; growth-related traits, such as normal seedling fresh weight (FW), normal seedling length, root length, vitality index, etc.; and related indicators after accelerated aging, such as relative germination rate (rGR), relative normal seedling length (rSL), relative normal seedling fresh weight (rFW), etc. (Zhang et al., 2019; Ma et al., 2024; Jiamtae et al., 2022; Wang et al., 2021). In addition, the conductivity of soybean seed soaking solution is also a valuable indicator for evaluating seed vitality (Colete et al., 2004). In summary, there are many indicators for evaluating soybean seed vitality, but there is still a lack of a complete evaluation system.

[0004] Principal component analysis (PCA) is a general statistical method that simplifies the original variables into several main components by reducing the dimension, thereby maximizing the explanation of the variance of all variables (Greenacre et al., 2022). PCA combined with multivariate analysis methods has been used to evaluate stress resistance traits and screen soybean germplasm. Wang et al. (2020) comprehensively evaluated the drought resistance of 62 soybean germplasms under PEG-simulated drought stress at the seedling stage by analyzing their wilting index and root morphological characteristics, and selected the drought-resistant excellent variety Tiefeng 31 using PCA. Zhou et al. (2023) conducted a comprehensive evaluation of the salt tolerance of soybeans during the germination period, and used PCA and other analytical methods to grade the salt tolerance of 20 soybean varieties, determine the theoretical identification concentration, and identify excellent salt-tolerant varieties such as Dongnong 254. However, no research on the evaluation of soybean seed vigor has been reported.

[0005] At present, there are many indicators for evaluating soybean seed vitality. A single indicator is difficult to accurately and comprehensively reflect the vitality differences between different germplasms. Existing methods often require complex indicators and cumbersome experimental steps, which makes them difficult to promote and apply on a large scale. In the evaluation and assessment of soybean seed vitality, there is still a lack of a simple, efficient and accurate comprehensive evaluation method based on multiple indicators. Summary of the invention

[0006] Technical issues

[0007] The purpose of the present invention is to provide a soybean seed vitality evaluation method which is simple, efficient, accurate and applicable to large-scale seed screening.

[0008] Technical Solution

[0009] The present invention provides a soybean seed vitality evaluation method based on multi-index comprehensive evaluation, comprising:

[0010] 1) Artificial accelerated aging of seeds

[0011] The soybean seed samples to be tested were subjected to artificial accelerated aging treatment for 5 days, and the treated seeds were placed at room temperature of 20℃~28℃ for 1 day before a standard germination test was conducted;

[0012] 2) Standard germination test

[0013] After sowing, the germination box was placed in an artificial climate chamber for 7 days for seed germination test. The temperature of the climate chamber was 23℃~28℃, the light intensity was 60000 Lux~80000 Lux, the photoperiod was 16 hours during the day and 8 hours at night, and the relative humidity was 50%~70%;

[0014] 3) Determination of seed vigor-related traits

[0015] The number of germinated seeds Gt was recorded every day from the second day to the seventh day after sowing, and the seedling length SL was recorded and measured on the seventh day; the germination index GI trait was calculated using the following formula: Germination index (GI) = ∑Gt / Dt, where Gt is the number of germinated seeds within t days, and Dt is the corresponding germination days; the relative value RSL of the seedling length trait SL and the relative value RGI of the germination index trait GI were calculated using the following formula: Relative value = trait value after accelerated aging treatment / trait value of the control × 100%.

[0016] 4) Soybean seed vitality assessment model

[0017] Model evaluation using seed vigor V=- 0.026+0.625×RSL+0.485×RGI calculates the estimated value of soybean seed vitality to be tested V, in this model: RSL is the relative value of the seedling length trait SL, RGI is the relative value of the germination index trait GI;

[0018] 5) Soybean seed vitality assessment results

[0019] According to the evaluation value of soybean seed vitality measured in step 4) V , determine the type of soybean seed vitality to be tested according to the following standards:

[0020] High-vigor soybean seeds: Grade I, 0.848 ≤ V ≤1;

[0021] Medium and high vigor soybean seeds: Grade II, 0.662 ≤ V <0.848;

[0022] Medium vigor soybean seeds: Grade III, 0.388 ≤ V <0.662;

[0023] Medium and low vigor soybean seeds: Grade IV, 0.195 ≤ V <0.388;

[0024] Low-vigor soybean seeds: Grade V, 0≤ V <0.195.

[0025] The parameters of the artificial accelerated aging treatment of seeds in step 1 of the method are 41°C ± 1°C and 95% ± 1% relative humidity. The standard germination test in step 2 of the method is carried out in accordance with GB / T 3543.4-1995 "Germination Test of Crop Seed Inspection Procedure". The standard germination test in step 2 of the method adopts the sand culture method, and the particle size of the river sand used is less than 0.5 mm; after sterilization, 160 mL of double distilled water is added to each kilogram of dry sand and fully mixed; before the germination test, the seeds are soaked in 75% ethanol solution for 1 minute for disinfection, and then rinsed three times with sterile water; in the germination test, 40 plump seeds with intact seed coats are selected for each sample and evenly sown in a germination box containing 1 kg of sand, and each germplasm is sown 3 times. Beneficial Effects

[0026] In the prior art, there are many indicators for evaluating soybean seed vitality, and a single indicator is difficult to fully and accurately reflect the vitality differences between different germplasms. In addition, the existing methods usually involve complex indicator systems and cumbersome experimental steps, which limits their feasibility in large-scale applications. At present, the field of soybean seed vitality evaluation and assessment is in urgent need of a simple, efficient and accurate comprehensive evaluation method. The present invention conducts artificial accelerated aging tests on 126 soybean local varieties screened from a soybean germplasm bank containing 23,587 copies, uses principal component analysis and membership functions to conduct in-depth analysis of various indicators, establishes a comprehensive evaluation system for soybean seed vitality, and constructs an efficient seed vitality evaluation model on this basis. This method can accurately and quickly monitor changes in seed vitality, which is of great significance to the promotion of high-vigor soybean varieties and the improvement of soybean production efficiency. Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0027] 1. High evaluation accuracy:

[0028] The present invention provides a new method for scientific evaluation of soybean seed vitality. The present invention is based on a multi-index comprehensive evaluation to ensure the comprehensiveness and accuracy of seed vitality. The key influencing factors are screened out through stepwise regression analysis, and the indicators of the fitting effect of the established regression model are R adj 2 The value reaches 0.916, which is close to 1, and can well evaluate the changes in seed vigor, and the accuracy of vigor classification evaluation reaches 87.30%, all of which indicate that the model has high accuracy.

[0029] 2. Simplify the operation process:

[0030] The present invention can evaluate seed vitality by measuring only two indicators, namely, relative seedling length and relative germination index, thus simplifying the experimental process, reducing experimental time and cost, and having higher simplicity and practicality.

[0031] 3. Improve sowing efficiency:

[0032] The present invention can accurately and quickly monitor the change of seed vitality and effectively evaluate the seed vitality before soybean sowing, thereby improving the soybean sowing efficiency.

[0033] 4. Promote high-vigor soybean breeding:

[0034] The evaluation method of the present invention has the advantages of high efficiency and simple operation, is convenient for seed vigor screening and evaluation of large-scale germplasm resources, helps to promote the breeding work and yield of high-vigor soybeans, and ultimately improves soybean production capacity.

[0035] 5. Applicable to other crops:

[0036] The establishment of the evaluation method of the present invention provides a new idea. The artificial accelerated aging method is used to measure 11 seed vitality-related traits of 126 soybean local varieties, and principal component analysis is performed. Then, the membership function standardization method is used to calculate the comprehensive evaluation value of seed vitality ( V The present invention is not only applicable to the vitality evaluation of soybean germplasm resources, but also helps to promote and establish the seed vitality evaluation method of other crops, and has a wide application prospect. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 . Comprehensive evaluation of 126 soybean local varieties V Value cluster analysis.

[0038] Figure 2 . Evaluate V Value and comprehensive evaluation V The linear relationship of the values. DETAILED DESCRIPTION

[0039] 1. Materials and methods

[0040] 1. Plant Materials

[0041] This study used 126 soybean landraces selected from a micro-core germplasm bank containing 23,587 soybean accessions. These varieties came from 23 provinces in China and four soybean ecoregions: the Northeast Ecoregion, the North Ecoregion, the Huang-Huai-Hai Ecoregion, and the South Ecoregion, with broad representativeness and genetic diversity (Table 1) (see Li X, et al. Genome-wide association study of four yield-related traits at the R6 stage insoybean [J]. BMC Genetics, 2019). The 126 soybean landraces are all publicly known and used, provided by the National Soybean Germplasm Bank of the Chinese Academy of Agricultural Sciences, and the China Crop Germplasm Resources Information Network (https: / / www.cgris.net / ). The unified number "ZDDXXXX" is the unified identification code of China's soybean germplasm resources, which is used to identify and manage soybean germplasm resources. In 2021 and 2022, the above 126 soybean local varieties were planted at the Baima Experimental Station (31°36′N, 119°9′E) and Yongning Experimental Station (32°13′N, 118°57′E) in Nanjing, Jiangsu Province. The experiment adopted a completely randomized block design and was repeated three times. The row length of each plot was 2 meters, the row spacing was 0.5 meters, and the plant spacing within the row was 0.1 meters. The seeds were harvested when fully mature, dried to a moisture content of about 10%, and then used for seed vigor trait determination.

[0042] 2. Artificially accelerated aging

[0043] In order to ensure that the vitality difference between samples is obvious, the fully mature seed samples are placed in small nylon mesh bags and neatly arranged in an aging box for accelerated aging at 41℃±1℃ and 95%±1% relative humidity for 5 days. The treated seeds are placed at room temperature of 20℃~28℃ for 1 day to allow moisture balance before conducting a standard germination test.

[0044] 3. Standard germination test

[0045] The standard germination test was conducted in accordance with the national standard GB / T 3543.4-1995 "Germination Test of Crop Seed Inspection Procedure". The standard germination test adopted the sand culture method. The particle size of the river sand used was less than 0.5 mm. After sterilization, 160 mL of double distilled water (equivalent to 80% of the maximum water holding capacity of the sand) was added to each kg of dry sand and mixed thoroughly. Before the germination test, the seeds were soaked in 75% ethanol solution for 1 minute for disinfection, and then rinsed three times with sterile water. 40 plump seeds with intact seed coats were selected from each germplasm and evenly sown in a germination box (19×13×12 cm) containing 1 kg of sand, and each germplasm was sown 3 times. After sowing, the germination box was placed in an artificial climate chamber for 7 days for seed germination test. The temperature of the climate chamber was 25℃, the light intensity was 70000 Lux, the photoperiod was 16 hours during the day and 8 hours at night, and the relative humidity was about 60%. The number of germinated seeds was recorded every day starting from the second day, and other morphological indices were measured on the seventh day.

[0046] 4. Determination of seed vigor-related traits

[0047] In order to comprehensively evaluate seed vigor, 11 related traits were measured. Germination-related traits were calculated using the following formulas: germination rate (GR) = number of germinated seeds / total number of seeds × 100%; germination potential (GP) = number of germinated seeds on the fourth day / total number of seeds × 100%; germination index (GI) = ∑Gt / Dt, where Gt is the number of germinated seeds within t days and Dt is the corresponding number of germination days. The taproot length (RL), seedling length (SL) and hypocotyl length (HL) were measured with a ruler in centimeters. After absorbing the surface moisture with absorbent paper, the seedling fresh weight (SFW) and root fresh weight (RFW) were measured and weighed with an electronic balance in grams. The seedling dry weight (SDW) and root dry weight (RDW) were measured by first sterilizing the fresh weight sample in an oven at 120°C for 15 minutes, then drying it at 80°C to constant weight, and finally weighing it with an electronic balance in grams.

[0048] For conductivity determination, 50 full seeds with intact seed coats were selected and weighed using an electronic balance (W). The seeds were placed in a clean 300 mL plastic cup, 250 ± 1 mL of double distilled water was added, and the lid was closed to prevent contamination. In addition, a blank control was prepared (the plastic cup contained only 250 ± 1 mL of double distilled water). After all samples were placed at 25 ± 1 ° C for 24 hours, the conductivity was measured using a conductivity meter. The conductivity reading was recorded as X and the blank control was recorded as X0. Each material was tested three times. The conductivity formula is: EC (µS·cm -1 ·g -1 )=(X-X0) / W. The relative values ​​of the 11 traits were calculated using the following formula: relative value=trait value after accelerated aging treatment / trait value of control × 100%.

[0049]

[0050] Note: The unified number "ZDDXXXX" is the unified identification code for China's soybean germplasm resources.

[0051] 2. Research Results

[0052] 1. Analysis of soybean vitality evaluation indicators

[0053] To evaluate the degree of artificial accelerated aging of different varieties, the ratio of the trait value after treatment to that before treatment was used as the vitality evaluation index, and the average value of the two environments was used for analysis. The results showed that the coefficients of variation of relative germination potential, relative germination index, relative germination rate, relative root fresh weight and relative root dry weight were large, ranging from 54.70% to 74.01%. In contrast, the coefficients of variation of relative conductivity, relative root length, relative seedling length, relative hypocotyl length, relative seedling fresh weight and relative seedling dry weight were small, ranging from 22.53% to 34.29% (Table 2). This shows that the vitality evaluation indexes show different degrees of variation, and the variation of all indicators is greater than 20%, indicating that a single evaluation index has certain limitations in evaluating seed vitality.

[0054]

[0055] 2. Principal component analysis of soybean vitality evaluation indexes

[0056] Refer to Wang G, et al. Wilting index and root morphologicalcharacteristics used as drought-tolerance variety selection at the seedlingstage in soybean ( Glycine maxL.). Plant Growth Regulation, 2020. The principal component analysis was used to comprehensively evaluate the seed vigor of different soybean varieties. The 11 seed vigor evaluation indicators that are easy to measure, including relative germination potential (RGP); relative germination rate (RGR); relative germination index (RGI); relative electrical conductivity (REC); relative taproot length (RRL); relative seedling length (RSL); relative hypocotyl length (RHL); relative seedling fresh weight (RSFW); relative root fresh weight (RRFW); relative seedling dry weight (RSDW); and relative root dry weight (RRDW), were selected for principal component analysis to reduce the dimension of all seed vigor indicators. After Kaiser normalization and variance maximum rotation, the principal components were extracted according to the standard of eigenvalue greater than 1, and finally two principal component factors were extracted, namely the first principal component factor F1 and the second principal component factor F2, with contribution rates of 43.08% and 34.72%, respectively. The cumulative contribution rate of the two principal component factors was 77.80%, which is quite representative (Table 3). The component matrix showed that the first principal component factor contributed the most, among which the relative germination potential (0.927), relative germination rate (0.913), relative germination index (0.924), relative root fresh weight (0.734) and relative root dry weight (0.758) had large loads, reflecting the germination situation and underground biomass, so it was named "germination and root growth". The second principal component factor had large loads of relative seedling length (0.767), relative hypocotyl length (0.821), relative seedling fresh weight (0.816) and relative seedling dry weight (0.798), mainly reflecting seedling growth and biomass, so it was named "seedling growth". In summary, the 11 vitality evaluation indicators were transformed into two principal component factors F1 and F2 for evaluating soybean seed vitality.

[0057]

[0058] 3. Establishment of a comprehensive evaluation method for soybean seed vitality

[0059] According to the principal component factor eigenvalue and component matrix, F1 and F2 are further analyzed, and the weight calculation formula of each principal component factor is: ; In the formula, ω p is the weight of the extracted Pth principal component factor; p is the eigenvalue corresponding to the extracted principal component factor, and the weights ω of F1 and F2 are determined according to the eigenvalue 1、 ω2 were 55.37% and 44.63% respectively.

[0060] Using the formula: ; Calculate the weight of each vitality evaluation index in each principal component, that is, the score coefficient (Table 4), where W ipis the weight of the ith evaluation index on the Pth principal component; ip is the contribution rate of the i-th evaluation index on the P-th principal component; p is the eigenvalue corresponding to the extracted Pth principal component.

[0061] According to the score coefficients of each vitality evaluation index in the two principal component factors, the scores of F1 and F2 are calculated as follows: F1=0.426×RGP+0.419×RGR+0.424×RGI-0.126×REC+0.254×RRL+0.255×RSL+0.178×RHL+0.224×RSFW+0.337×RRFW+0.012×RSDW+0.348×RRDW; F2=0.125×RGP+0.125×RGR+0.136×RGI-0.256×REC+0.297×RRL+0.392×RSL+0.420×RHL+0.418×RSFW+0.254×RRFW+0.408×RSDW+0.245×RRDW.

[0062] Using the membership function method, using the formula: ; Standardize the scores of F1 and F2, where μF i represents the standardized value of germplasm i on a principal component, Fi represents the score of germplasm i on the principal component, and F min and F max They represent the minimum and maximum values ​​of all germplasms on this principal component, respectively (Sun et al., 2022), and the standardized scores μF1 and μF2 are calculated.

[0063] Finally, according to the weights ω of F1 and F2 1、 ω2 and the standardized μF1 and μF2, using the formula: ; Calculate the comprehensive evaluation value of soybean germplasm i seed vitality, referred to as V Value. V The value measures the vitality level of soybean varieties. V The higher the value, the stronger the ability to resist artificial accelerated aging and the higher the vitality. Comprehensive evaluation of soybean seed vitality of 126 varieties V The value range is between 0 and 1.

[0064]

[0065] 4. Comprehensive evaluation of soybean local varieties V Value Cluster Analysis

[0066] In order to classify soybean varieties into different vitality levels and reduce the subjectivity of artificial grouping, this study used the "hclust" function in the R language software (V4.4.1) to comprehensively evaluate different varieties using the complete connection method. V Hierarchical cluster analysis was performed on the values ​​( Figure 1 ), used to evaluate the similarity between samples and construct a dendrogram. Cluster analysis comprehensively evaluated 126 soybean local varieties. V The vitality level of the value results is divided into 5 categories: 7 varieties with high vitality (level I, 0.848≤ V ≤1); 15 varieties with medium and high vitality (Level II, 0.662≤ V <0.848); 59 varieties with medium vitality (Level III, 0.388≤ V <0.662); 36 varieties with medium and low vitality (IV level, 0.195≤ V <0.388); 9 varieties with low vitality (V grade, 0≤ V <0.195) (Table 5).

[0067]

[0068] 5. Establishment of soybean seed vitality assessment model

[0069] The evaluation model was constructed by using stepwise regression analysis and by gradually adding or deleting independent variables that had a significant impact on the dependent variable. This study used IBM SPSS 25.0 statistical software to comprehensively evaluate soybean seed vitality. V The value was used as the dependent variable, and the 11 vitality evaluation indicators of soybean varieties were used as independent variables to establish 11 significant evaluation models. From model 1 to model 11, there are 1, 2, 3...11 vitality evaluation indicators respectively. The vitality evaluation indicators included in each model were selected and determined by statistical methods. Based on the significance test of the model fitting effect, the comprehensive evaluation indicators were gradually introduced or deleted. V The vitality evaluation index has a significant impact on the value, thus establishing the evaluation V The order of introducing vitality evaluation indicators into the model and the number of vitality evaluation indicators contained in each model are also selected and determined by statistical methods. This process is usually based on the contribution of each vitality evaluation indicator to the explanatory power of the model and the significance level of the model fitting effect.

[0070] With the addition of more vitality evaluation indicators, the model can evaluate V The explanatory power of the value gradually increases. At the same time, the residual standard error ( RSE ) gradually decreased, indicating that the model fit was gradually improved (Table 6). R adj ²Higher, better able to explain the variation in the data, the model is more effective in evaluatingV The predictive power of the value becomes stronger, and RSE Lower ( R adj ²>0.98; RSE <0.03), but the first three models provide robust evaluation performance while maintaining the simplicity of variables, making them more practical. Therefore, the present invention focuses on the analysis of the first three models.

[0071]

[0072] Among them, model 1 uses relative seedling length (RSL) as the only variable and obtains the regression equation V=- 0.074+0.948×RSL. Relative seedling length has a significant effect on the evaluation of seed vigor ( P <0.01). Model 2 is based on Model 1, with relative germination index (RGI) as the second variable. The coefficients of RSL and RGI are 0.625 and 0.485, respectively ( P <0.01), indicating that these two variables significantly improved the accuracy of evaluating seed vigor, and the regression equation was obtained. V=- 0.026+0.625×RSL+0.485×RGI. Model 3 introduces relative conductivity (REC) as the third independent variable based on model 2, and the regression equation is obtained. V =0.216+0.537×RSL+0.454×RGI-0.077×REC. In this model, RSL and RGI were still significant positive evaluation factors, while the coefficient of REC was -0.077, showing a significant negative correlation with seed vigor (Table 7).

[0073]

[0074] 6. Verification of the application effect of soybean seed vigor assessment model

[0075] The 10-fold cross validation (CV) method was used to evaluate the stability of the model and assess the performance. The dataset was randomly divided into 10 subsets. In each iteration, 9 subsets were used for model training, while the remaining subsets were used as test sets. This process was repeated 10 times to ensure that each subset was used as a test set once, maximizing data utilization and minimizing the risk of overfitting. The root mean square error ( RMSE ) is used to measure the standard deviation of the difference between the estimated value and the actual value, and the coefficient of determination ( R 2 ) represents the proportion of variance explained by the model in the data. R 2 ( R adj 2) is the number of independent variables in the model. R 2 The adjustment is an indicator of the model fitting effect, and the mean absolute error (MAE) is used to calculate the model evaluation V Value and comprehensive evaluation V The mean absolute difference between the values. These indicators are used to evaluate the model performance.

[0076] The above model evaluation index formula is as follows:

[0077] Evaluation of soybean seed vitality measured by vitality evaluation model V The type of soybean seed vitality to be tested is also determined according to the following standards: High vitality soybean seeds: Class I, 0.848≤ V ≤1; Medium and high vitality soybean seeds: Level II, 0.662≤ V <0.848; Medium vigor soybean seeds: Grade III, 0.388≤ V <0.662; Medium and low vitality soybean seeds: IV level, 0.195≤ V <0.388; Low vitality soybean seeds: V level, 0≤ V <0.195.

[0078] Classification Accuracy = Model Evaluation V Value classification results and comprehensive evaluation V The number of germplasms with consistent classification results / the total number of evaluated germplasms × 100%, where is the comprehensive evaluation of the i-th germplasm V value; is the evaluation of the i-th accession given by the model V value; Comprehensive evaluation of all germplasm V The average value of the values; n is the number of germplasm samples 126; k is the number of evaluation factors in the model, namely, 11 evaluation indicators of germplasm vitality traits.

[0079] The 10-fold cross validation was used to evaluate the evaluation accuracy and model fit goodness of the three regression models 1, 2, and 3 with different complexity (Table 8). The results show that model 2 provides good evaluation performance ( RMSE =0.055, R adj 2 =0.916), Figure 2 The scatter plot in the figure intuitively shows the comprehensive evaluation V Values ​​and Model Evaluation V Therefore, in order to balance the simplicity of the model and the accuracy of the evaluation, Model 2 is selected as the optimal model. Model 2 has high prediction accuracy. RMSEThe lower value is 0.055, which means that the model has a small prediction error for seed vigor, indicating that the model has good prediction accuracy. In addition, model 2 has strong explanatory power, and the index of model fitting effect R adj 2 It is 0.916, close to 1, indicating that RSL and RGI in the model can well explain the variation of seed vigor, and the model achieves a good fitting effect without introducing additional variables. In short, the evaluation results of Model 2 show that it has high accuracy and strong explanatory power in predicting soybean seed vigor, while maintaining low complexity, and is a very effective prediction model.

[0080] Practical application verification: Model 2 vitality evaluation V The vitality category of the value is consistent with the previous comprehensive evaluation V The results of the vigor classification of the hierarchical cluster analysis of the values ​​of the two materials were compared. Among the 126 materials, the evaluation classification of 110 materials was consistent with the actual classification, and the vigor classification accuracy was 87.30% (Tables 8 and 9). The results showed that Model 2 had high reliability and accuracy in the classification evaluation of seed vigor grades. In addition, the model only used two independent variables, maintaining a low complexity. Therefore, the actual application effect of Model 2 verified that the seed vigor of soybean germplasm can be effectively evaluated by measuring the two indicators of RSL and RGI.

[0081]

[0082] Note: RMSE: root mean square error; R adj ²: Adjusted R ² value; MAE: mean absolute error

[0083]

[0084] 3. Conclusion

[0085] This study invented a soybean seed vigor assessment method based on multi-index comprehensive evaluation. Using the artificial accelerated aging method, 11 seed vigor-related indicators of 126 soybean landraces selected from a micro-core germplasm bank containing 23,587 soybean germplasms were measured. Through principal component analysis, a comprehensive evaluation system for soybean seed vigor was established, and a concise and efficient seed vigor assessment model was constructed: V=- 0.026+0.625×RSL+0.485×RGI. In this model, relative seedling length (RSL) and relative germination index (RGI) have significant positive effects on seed vigor. This model provides good evaluation performance ( RMSE =0.055, Radj 2 =0.916) and 87.30% vitality classification accuracy, while only two independent variables were used, maintaining a low complexity and balancing the simplicity of the model and the accuracy of the evaluation. In practical applications, the present invention can effectively evaluate the seed vitality of soybeans before sowing, thereby accurately and quickly monitoring the changes in seed vitality and improving the sowing efficiency of soybeans. The evaluation method of the present invention has the advantages of high efficiency and simple operation, is convenient for seed vitality screening and evaluation of large-scale germplasm resources, helps to promote the breeding of high-vigor soybeans, and ultimately improves soybean production capacity. The present invention provides a new method for the scientific evaluation of soybean seed vitality, and helps to promote the establishment of seed vitality evaluation methods for other crops, and has broad application prospects.

Claims

1. A soybean seed vitality assessment method based on multi-index comprehensive evaluation, comprising: 1) Artificial accelerated aging of seeds The soybean seed samples to be tested were subjected to artificial accelerated aging treatment for 5 days. The parameters for artificial accelerated aging treatment of seeds were 41℃±1℃ and 95%±1% relative humidity. The treated seeds were placed at room temperature of 20℃~28℃ for 1 day and then subjected to standard germination test; 2) Standard germination test The sand culture method was used, and the river sand used had a particle size of less than 0.5 mm. After sterilization, 160 mL of double distilled water was added to each kg of dry sand and mixed thoroughly. Before the germination test, the seeds were soaked in 75% ethanol solution for 1 minute for disinfection, and then rinsed three times with sterile water. In the germination test, 40 plump seeds with intact seed coats were selected for each sample and evenly sown in a germination box containing 1 kg of sand, and each germplasm was sown 3 times. After sowing, the germination box was placed in an artificial climate chamber for 7 days for seed germination test. The temperature of the climate chamber was 23℃~28℃, the light intensity was 60000 Lux~80000 Lux, the photoperiod was 16 hours during the day and 8 hours at night, and the relative humidity was 50%~70%; 3) Determination of seed vigor-related traits The number of germinated seeds Gt was recorded every day from the second day to the seventh day after sowing, and the seedling length SL was recorded and measured on the seventh day; The germination index GI trait was calculated using the following formula: germination index GI = ∑ (Gt / Dt), where Gt is the number of germinated seeds within t days, and Dt is the corresponding number of germination days; the relative value RSL of the seedling length trait SL and the relative value RGI of the germination index trait GI were calculated using the following formula: relative value = trait value after accelerated aging treatment / trait value of control × 100%; 4) Soybean seed vitality assessment model Model evaluation using seed vigor V =−0.026+0.625×RSL+0.485×RGI to calculate the estimated value of soybean seed vitality to be tested V , in this model: RSL is the relative value of the seedling length trait SL; RGI is the relative value of the germination index trait GI; 5) Soybean seed vitality assessment results According to the evaluation value of soybean seed vitality measured in step 4) V , determine the type of soybean seed vitality to be tested according to the following standards: High-vigor soybean seeds: Grade I, 0.848 ≤ V ≤1; Medium and high vigor soybean seeds: Grade II, 0.662 ≤ V <0.848; Medium vigor soybean seeds: Grade III, 0.388 ≤ V <0.662; Medium and low vigor soybean seeds: Grade IV, 0.195 ≤ V <0.388; Low-vigor soybean seeds: Grade V, 0≤ V <0.195.

Images