Method for evaluating resistance of rice brown planthopper and application of method

Abstract

The invention discloses a method for evaluating resistance of rice brown planthopper and application of the method. The method can identify or predict the brown planthopper resistance based on the characteristics of volatile organic compounds (VOC) of rice, solves the problems of strong insect source dependence, complex operation, strong destructiveness, single evaluation dimension, low data acquisition efficiency and the like in the prior art, realizes rapid and nondestructive identification of the brown planthopper resistance in the adult-plant stage of rice, and has good application prospects. And technical support is provided for resistant variety screening, insect-resistant mechanism research and green prevention and control strategies.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of modern agricultural technology, in particular to a method for evaluating the resistance of rice to brown planthopper and application thereof. BACKGROUND

[0002] The existing resistance identification of rice to brown planthopper is divided into seedling stage identification and adult stage identification, among which, the seedling stage identification is the mainstream identification method. However, in actual production, the damage of brown planthopper to rice mainly occurs in the adult stage. It has not been confirmed in related research that the resistance of rice to brown planthopper in the seedling stage is necessarily consistent with that in the adult stage. Therefore, it is difficult to fully reflect the insect resistance ability of rice in the field environment by relying only on the seedling stage resistance identification. In comparison, the resistance identification in the adult stage is relatively more practical for guiding field production practice.

[0003] The existing adult stage resistance identification methods mainly include the following several methods, but these methods have certain limitations. For example, the rice stalk identification method mainly analyzes the resistance characteristics of a single tiller on the rice stalk to infer the resistance of the whole plant. However, due to the differences in growth and development and metabolic level between the single tiller and the whole plant, and the mechanical damage to the plant during the process of stripping the tiller, the identification results obtained by this method are difficult to accurately reflect the true resistance performance of the whole plant. The single plant identification method is to evaluate the insect resistance by transplanting single rice plants in pots and installing supports and net covers. This method is suitable for artificial climate chambers or greenhouse environments, and is complicated to operate and requires a large space, which is difficult to reflect the group resistance level of rice varieties in the actual field environment. The field insect introduction combined with molecular marker method combines the mortality evaluation by artificial insect introduction in the field net cover and the detection of resistance genes by molecular markers, but in this way, it is necessary to manage the insect source, handle the insect introduction, extract DNA and detect molecules, which is a complex operation process with high technical requirements, and is difficult to cover unknown or complex phenotypic resistance mechanisms.

[0004] Therefore, there is an urgent need for a new method that can effectively overcome the problems of strong dependence on insect source, complex operation, strong destructiveness, single evaluation dimension, low data acquisition efficiency and the like in the above-mentioned methods, so as to realize the rapid and non-destructive identification of the resistance of rice to brown planthopper in the adult stage, and provide technical support for resistance variety screening, insect resistance mechanism research and green prevention and control strategy. SUMMARY

[0005] The present application aims to at least solve one of the above technical problems existing in the prior art. To this end, the present application aims to provide a method for evaluating rice brown planthopper resistance and its application. In the present application, a non-destructive brown planthopper resistance identification or prediction method based on the characteristics of rice volatile organic compounds (VOC) is provided to solve the problems of strong dependence on insect sources, complex operation, strong destructiveness, single evaluation dimension, low data acquisition efficiency and the like in the prior art, and to realize rapid and non-destructive identification of rice brown planthopper resistance at the plant stage, thereby providing technical support for resistance variety screening, insect resistance mechanism research and green prevention and control strategies.

[0006] In a first aspect of the present application, a method for constructing a model for predicting rice brown planthopper resistance is provided, comprising the following steps:

[0007] (1) determining rice brown planthopper resistance based on field tests, establishing a resistance label, collecting rice volatile organic compounds (VOC), and using a partial least squares method to screen VOC markers based on the resistance label;

[0008] (2) using a machine learning algorithm to model based on the VOC markers and brown planthopper resistance to obtain a model for predicting rice brown planthopper resistance.

[0009] In some embodiments of the present application, the machine learning algorithm comprises at least one of K-nearest neighbors, support vector machines and random forests.

[0010] In some embodiments of the present application, the machine learning algorithm is a support vector machine.

[0011] In some embodiments of the present application, the support vector machine is adjusted for hyperparameters.

[0012] In some embodiments of the present application, the adjustment parameter categories of the support vector machine include C, gamma and kernel.

[0013] In the present application, the adjustment parameter C of the support vector machine represents a penalty coefficient for balancing the margin and the misclassified samples, and the default parameter is 1.0. The kernel represents the selection method of the kernel function, including RBF, Linear, Poly and Sigmoid. The gamma represents the coefficient of the kernel function (only effective for Poly, RBF and Sigmoid), and if gamma is not specifically set, it represents the value of the inverse of the number of sample features, i.e. 1 / n features.

[0014] In some embodiments of the present application, the VOC marker is a combination containing the VOCs shown in Table 2.

[0015] In some embodiments of the present application, the VOC markers are a combination of the VOCs shown in Table 2. Specifically, the VOC markers are (1S)-6,6-dimethyl-2-methylenebicyclo[3.1.1]heptane, isobutylcyclohexane, dodecyl nonyl ether, beta-pinene and 2,6-bis(1,1-dimethylethyl)-2,5-cyclohexadiene-1,4-dione.

[0016] In some embodiments of the present application, in step (1), the screening comprises: performing importance analysis on the rice VOCs, and selecting the first N rice VOCs according to the importance ranking as the VOC markers.

[0017] In some embodiments of the present application, N is selected from any integer between 2-7.

[0018] In some embodiments of the present application, N is 5.

[0019] In some embodiments of the present application, in step (2), the VOC markers are quantitative or qualitative data of the VOC markers.

[0020] In some embodiments of the present application, in step (2), the VOC markers used in the modeling are the TD-GC / MS detection peak areas of the VOC markers.

[0021] In some embodiments of the present application, the construction method is specifically:

[0022] (1) collecting rice volatile organic compounds (VOCs) using a dynamic headspace sampling device, obtaining VOC peak area data using thermal desorption-gas chromatography-mass spectrometry (TD-GC / MS), performing importance analysis on the VOCs based on the brown planthopper resistance data using partial least squares method, ranking according to the importance analysis parameters, and selecting the first N rice VOCs as the VOC markers;

[0023] (2) modeling based on the peak areas of the VOC markers and the brown planthopper resistance using support vector machine, to obtain a model for predicting the brown planthopper resistance of rice.

[0024] In some embodiments of the present application, the detection method of TD-GC / MS is:

[0025] The rice to be tested is sealed in a closed space, the gas therein is extracted into an adsorption tube, the adsorption tube is heated, and then GC / MS is used to detect the gas therein.

[0026] In a second aspect of the present application, the model constructed by the construction method of the above aspect is used to evaluate or predict the insect resistance of rice.

[0027] In some embodiments of the present application, the insect resistance comprises brown planthopper resistance.

[0028] In some embodiments of the present application, the rice is adult rice.

[0029] In a third aspect of the present application, a marker is provided, wherein the marker is rice VOC.

[0030] In some embodiments of the present application, the VOC marker is as defined in the above aspect.

[0031] In some embodiments of the present application, the VOC marker is used for evaluating or predicting the insect resistance of rice.

[0032] In some embodiments of the present application, the insect resistance comprises brown planthopper resistance.

[0033] In some embodiments of the present application, the rice is adult rice.

[0034] In a fourth aspect of the present application, a use of a reagent for detecting the marker in the above aspect in the preparation of a product for evaluating or predicting the insect resistance of rice is provided.

[0035] In some embodiments of the present application, the reagent for detecting the marker in the above aspect comprises a reagent used based on a method selected from the group consisting of:

[0036] gas chromatography, ionization detection, infrared spectroscopy, solid-phase microextraction, bag analysis, high-performance liquid chromatography (HPLC), continuous monitoring system, etc.

[0037] In some embodiments of the present application, the gas chromatography comprises but is not limited to: thermal desorption-gas chromatography (TD-GC), purge and trap-gas chromatography mass spectrometry (P&T-GC-MS), headspace gas chromatography.

[0038] In some embodiments of the present application, the ionization detection comprises but is not limited to: photoionization detection (PID), flame ionization detection (FID)

[0039] In some embodiments of the present application, the product comprises a reagent, a kit, a chip, and a detection kit.

[0040] In a fifth aspect of the present application, a method for evaluating or predicting the insect resistance of rice is provided, comprising the following steps:

[0041] collecting VOC of the rice to be tested, detecting based on the marker in the above aspect, and inputting the detection result into the model constructed by the construction method in the above aspect to obtain the evaluation or prediction result of the insect resistance of the rice.

[0042] In some embodiments of the present application, the detection is to obtain the peak area of the marker shown in the above aspect in the sample to be detected (i.e. the detection result) using TD-GC / MS.

[0043] In some embodiments of the present application, the evaluation or prediction result of the rice resistance is the resistance level of the rice (susceptible, resistant and highly resistant).

[0044] In a sixth aspect of the present application, a computer readable storage medium is provided, and the computer readable storage medium stores a computer program, and the computer program is executed by a processor to realize the steps in the model or method constructed by the construction method of the above aspect.

[0045] In a sixth aspect of the present application, a system is provided, and the system comprises an integrated VOC detection module and a model identification module, and the model identification module comprises the computer readable storage medium of the above aspect.

[0046] In some embodiments of the present application, the integrated VOC detection module comprises a sample collection device and a VOC analysis device.

[0047] The present application has the following beneficial effects:

[0048] The present application collects and analyzes the characteristics of volatile organic compounds (VOCs) released by different rice varieties at the tillering stage, explores the relationship between the characteristics and the resistance to brown planthopper, and constructs a resistance identification model based on the characteristics of VOCs, so as to ultimately realize the rapid, non-invasive and automatic prediction of the resistance of rice to brown planthopper. The model performs excellently in key indicators such as accuracy (0.89), precision (0.92), recall rate (0.89) and F1-score (0.89), and the AUC value is as high as 0.98, showing strong classification ability (prediction ability). BRIEF DESCRIPTION OF DRAWINGS

[0049] Figure 1 The results of using partial least squares discriminant analysis (PLS-DA) to distinguish high resistance (HR), resistance (R) and susceptible rice (S).

[0050] Figure 2 The fitting situation after 200 cross-validations. DETAILED DESCRIPTION

[0051] The content of the present application will be further described in detail through specific examples. Unless otherwise specified, the raw materials, reagents or devices used in the examples and comparative examples can be obtained from conventional commercial channels or can be obtained by existing technical methods. Unless otherwise specified, the test or test method is a conventional method in the art.

[0052] Example 1

[0053] Preparation of the brown planthopper source:

[0054] In early May 2024, healthy TN1 rice seeds that had been screened for no problems were taken, soaked in a greenhouse for 48 h, with water changed 2-3 times during the period, and then uniformly sown on the seedbed after being incubated in a 28°C incubator for 36 h.

[0055] In early June 2024, the seedlings were transplanted to the field plots, with a plant spacing and row spacing of 20 cm each, and a plot area of 9 m x 12 m.

[0056] At the end of June 2024, a 120-mesh white transparent nylon mesh cover was placed on the TN1 plot, and 1000 brown planthopper female adults carrying eggs were introduced to breed eggs. Normal water and fertilizer management was carried out in the field, and a sufficient amount of brown planthopper population was bred on TN1 for standby use.

[0057] Example 2

[0058] In late May 2024, 10 (10 of each) healthy rice seeds that had been screened for no problems were taken, soaked in a greenhouse for 48 h, with water changed 2-3 times during the period, and then uniformly sown on the seedbed after being incubated in a 28°C incubator for 36 h. After 30 days, the seedlings were transplanted to the field plots, with a plant spacing and row spacing of 20 cm each, and 40 plants (5 clusters x 8 rows) were planted per plot. The plots were cross-arranged, with 4 repeats for each variety. Normal water and fertilizer management was carried out in all plots, and no pesticides were applied in the test plots throughout the growing season.

[0059] After the rice entered the peak tillering stage, a 120-mesh white transparent nylon mesh cover was placed on each variety plot, and then the brown planthopper source bred in the above example was introduced. The amount of insects was about 1800 2-3 instar brown planthopper nymphs per 40 rice plants. Among the 4 repeats of each rice variety, 3 repeats were inoculated with brown planthoppers, and the 4th repeat was covered with a 120-mesh white transparent nylon mesh cover to avoid damage from brown planthoppers in the field. These plots without inoculation were used to determine the source information of volatile organic compounds (VOCs) to ensure that it could be compared to determine which VOCs reflected the metabolic characteristics of the variety itself, rather than changes caused by insect damage.

[0060] After the introduction of the insect source, normal water and fertilizer management was carried out in the field, and no pesticides were applied in all plots throughout the growing season.

[0061] When the feeding rate of TN1 plants in the plot reached about 95%, the plant death of other introduced varieties was investigated, and the plant mortality rate of each variety was calculated.

[0062] According to the "Rice Field Resistance Identification Evaluation Standard" (DB42 / T 1376-2018 Hubei Provincial Standard: Rice Brown Planthopper Resistance Identification Technical Specification), based on the plant mortality of different varieties, the resistance of the test materials to brown planthoppers at the adult stage was evaluated.

[0063] The results are shown in the following table.

[0064] Table 1: Identification results of brown planthopper resistance of different varieties of rice at the adult stage

[0065]

[0066]

[0067] Among them, varieties V1 and V7 used known high-susceptible rice TN1 and high-resistant rice IR56, respectively, to verify the accuracy of the identification results.

[0068] As can be seen from the results in Table 1, the susceptible control variety TN1 had a dead seedling rate of 95.8%, an average damage level of 9, and was susceptible (9 level); the resistant control variety IR56 had a dead seedling rate of 1.1%, an average damage level of 1, and was highly resistant (1 level), and the identification system was working normally. The 10 varieties of rice can be divided into 3 categories: susceptible (V1 and V2), resistant (V3, V4, V5 and V6), and highly resistant (V7, V8, V9 and V10).

[0069] Example 3

[0070] In this example, a model-based rice adult stage brown planthopper resistance prediction method is constructed.

[0071] The specific steps are as follows:

[0072] (1) In order to avoid the influence of sample imbalance on model training and evaluation, according to the field insect density of the above-mentioned 10 different varieties of rice, 6 of them were selected as the training set for model training. Among them, the selected rice varieties are V1-4 and V7-8, which represent 3 resistance levels of rice, and each level contains 2 rice varieties.

[0073] Through a dynamic headspace sampling device, volatile organic compounds (VOCs) of the above-mentioned rice varieties were collected, and the collected VOCs were quantitatively detected by thermal desorption-gas chromatography-mass spectrometry (TD-GC / MS) based on conventional methods, to obtain VOC quantitative detection data.

[0074] Among them, the detection method of TD-GC / MS is as follows:

[0075] The adult rice is covered with a sealing device, leaving two air inlets and outlets. A gas pump is connected to the two air inlets and outlets to make one end of the gas pump an air inlet and the other end an air outlet, so as to discharge the gas into the gas adsorption tube (the gas adsorption tube is purchased from Hangzhou Lanrui Scientific Instrument Co., Ltd., and the inner filler is 150 mg of Tenax TA and 200 mg of Carbograph 1TD) for storage. The adsorption tube is heated by a thermal desorption (TD) device, and the gas stored therein is discharged to a GC / MS device for detection. Compared with the general VOC collection method, the method can reduce the loss of VOC caused by solvent dissolution.

[0076] Then, the collected VOC quantitative detection data are subjected to partial least squares discriminant analysis (PLS-DA) until clear distinction between the three types of samples is achieved (as shown in FIG. 2). Figure 1

[0077] The obtained PLS-DA model is verified by 200 times of permutation test to confirm that the PLS-DA model has significance and stability, and overfitting is avoided (as shown in FIG. 3). Figure 2

[0078] Then, based on the obtained PLS-DA model, the variable importance projection value (VIP) is calculated to obtain key markers, wherein the top 5 key VOCs with significant contribution in sample classification are selected as potential insect resistance metabolic markers.

[0079] The selected 5 key VOCs are shown in the following table.

[0080] Table 2: VOCs with top 5 VIP values

[0081] (1S)-6,6-dimethyl-2-methylenebicyclo[3.1.1]heptane Dodecyl nonyl ether 2,6-bis(1,1-dimethylethyl)-2,5-cyclohexadiene-1,4-dione Isobutylcyclohexane Beta pinene

[0082] The TD-GC / MS detection results (peak area) of the above-mentioned 5 key VOCs are selected as input items, and the data of the 6 varieties are divided into a training set and a test set for the rice adult stage brown planthopper resistance prediction model according to a data amount ratio of 3:1. Different machine learning algorithms are used for modeling (using the default algorithm parameters), and the results are shown in the following table.

[0083] Among them, the model output item is the resistance level (susceptible, resistant and highly resistant).

[0084] Table 3: Model effects of different algorithms under default parameters

[0085]

[0086] The model is subjected to hyperparameter adjustment, and the adjustment parameters are shown in the following table.

[0087] ​​Table 4 Hyperparameter adjustment situation

[0088]

[0089]

[0090] The model effect after hyperparameter adjustment is shown in the following table.

[0091] Table 5 Model effect of different algorithms after hyperparameter adjustment

[0092]

[0093] It can be found that support vector machine (SVM) is most suitable for the construction of this type of model, which performs well in key indicators such as accuracy (0.89), precision (0.92), recall (0.89) and F1-score (0.89), and the AUC value is as high as 0.98, showing strong classification ability (prediction ability). At the same time, the accuracy of support vector machine in cross-validation is also high, and the standard deviation is only 0.008, which shows that the model has good stability and generalization ability under different data division. While other models, such as random forest (RF), have similar evaluation values in accuracy, recall, F1-score, cross-validation accuracy and standard deviation, but its precision (0.91) and AUC (0.75) are low. For example, the XGBoost model, although it has similar evaluation values in accuracy, recall and F1-score, its precision (0.91), AUC (0.75) and cross-validation accuracy (0.67) are low, and the standard deviation is high (0.025). KNN, logistic regression (LR) and LightGBM are significantly inferior in multiple indicators. Considering various indicators and model stability, the rice brown planthopper resistance prediction model based on support vector machine (SVM) is the optimal model, and this model is selected for further verification in the future.

[0094] Example 4

[0095] In order to further verify the effectiveness of the model, the above rice brown planthopper resistance prediction model based on support vector machine is used to predict the data of another 4 rice varieties, and the actual field results in Comparative Example 1 are used as a control to determine the accuracy of the model prediction effect.

[0096] The results are shown in the following table.

[0097] Table 6 Model effect of different algorithms

[0098] Variety Results of field resistance identification in the adult stage Model predicted resistance Whether the prediction is accurate V5 Susceptible Susceptible Yes V6 Susceptible Susceptible Yes V9 Highly resistant Highly resistant Yes V10 Highly resistant Highly resistant Yes

[0099] The results show that the resistance identification results based on the above model and the field resistance identification results in the adult stage are completely consistent, which further illustrates the reliability and accuracy of the above model for predicting the resistance of rice brown planthopper.

[0100] Embodiment 5

[0101] The embodiment discloses a set of rice adult stage resistance prediction or rapid identification system, which comprises an integrated VOC detection module and a model identification module.

[0102] The integrated VOC detection module comprises a sample collection device and a VOC analysis device.

[0103] The data (qualitative and / or quantitative) of (1S)-6,6-dimethyl-2-methylenebicyclo[3.1.1]heptane, isobutylcyclohexane, dodecyl nonyl ether, beta-pinene and 2,6-bis(1,1-dimethylethyl)-2,5-cyclohexadiene-1,4-dione in the sample are obtained by inputting the sample collected by the sample collection device into the VOC analysis device. Then the data are output to the model identification module.

[0104] The rice brown planthopper resistance prediction model in the above embodiment is loaded in the model identification module. By loading the data output by the integrated VOC detection module into the model, the resistance prediction result can be directly output.

[0105] The above embodiment is a preferred embodiment of the present application, but the embodiments of the present application are not limited by the above embodiment, and any changes, modifications, substitutions, combinations, simplifications made without departing from the spirit and principles of the present application should be equivalent replacement methods, which are all included in the protection scope of the present application.

Claims

1. A method for constructing a model for predicting resistance of rice to brown planthopper, comprising the following steps: (1) determining resistance of rice to brown planthopper based on field test, establishing a resistance label, collecting volatile organic compounds (VOC) of rice, and screening VOC markers based on the resistance label using a partial least squares method; (2) modeling based on the VOC markers and resistance to brown planthopper using a machine learning algorithm to obtain a model for predicting resistance of rice to brown planthopper. wherein The machine learning algorithm comprises at least one of K-nearest neighbors, support vector machines, and random forests.

2. The construction method of claim 1, wherein, The VOC markers are a combination of VOCs shown in Table 2. Preferably, the VOC markers are a combination of VOCs shown in Table 2.

3. The construction method of claim 1, wherein, In step (1), the screening comprises: performing importance analysis on VOCs of rice, and selecting the first N VOCs of rice according to the importance ranking as VOC markers; wherein N is an integer selected from any integer between 2 and 7. 4.Use of the model constructed by the method of any one of claims 1-3 in evaluating or predicting insect resistance of rice. Preferably, the insect resistance comprises resistance to brown planthopper.

5. Use according to claim 4, characterized in that, The rice is rice at the tillering stage.

6. A marker characterized by The markers are VOCs of rice. Preferably, the markers are a combination of VOCs shown in Table 2. Preferably, the markers are a combination of VOCs shown in Table 2. 7.Use of a reagent for detecting the markers of claim 6 in preparing a product for evaluating or predicting insect resistance of rice. Preferably, the product comprises a reagent, a kit, a chip, and a detection kit. 8.A method for evaluating or predicting insect resistance of rice, comprising the following steps: collecting VOCs of rice to be tested, obtaining TD-GC / MS peak areas based on the markers shown in claim 6, inputting the TD-GC / MS peak areas into the model constructed by the method of any one of claims 1-3 to obtain an evaluation or prediction result of insect resistance of rice.

9. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the method of claim 8 or the model constructed by the method of any one of claims 1-3.

10. A system, characterized by The system comprises an integrated VOC detection module and a model identification module, wherein the model identification module comprises the computer readable storage medium of claim 9. Preferably, the integrated VOC detection module comprises a sample collection device and a VOC analysis device.

Images