A method for screening chitinolytic enzyme-producing bacteria for nematicidal agents

Abstract

The application provides a chitinase-producing bacterial screening method for nematicides, and belongs to the technical field of nematicides.The application adopts high-precision liquid chromatography-mass spectrometry technology to detect enzyme products, establishes an accurate mass spectrometry characteristic database, uses a chitin degradation fitting model to accurately analyze the primary, secondary and tertiary characteristics of mass spectrometry data, replaces inaccurate judgment of artificial experience, constructs an accurate double-layer game mathematical model composed of an upper model with the maximum enzyme activity as the target and a lower model with the optimal killing effect of root-knot nematodes as the target, accurately solves the best enzyme strain screening parameter combination through a multi-objective optimization algorithm, tests the root-knot nematode killing activity of the candidate bacterial strains, selects high-precision screening strains with a mortality rate greater than 85%, and performs morphological identification to establish a bacterial strain preservation library, so that the technical problem that the prior art is often based on artificial experience and has insufficient accuracy is solved.

Description

Technical Field

[0001] This invention belongs to the field of nematicide technology, and more specifically, relates to a method for screening chitin-degrading enzyme-producing bacteria for use in nematicides. Background Technology

[0002] Chitin-degrading enzymes have significant application value in agricultural biological control and pest management. Current screening techniques primarily employ the traditional clear zone method for initial screening. Researchers assess enzyme activity by visually observing the size of the clear zone formed by strains on chitin-containing media. Subsequently, candidate strains are selected based on personal experience and subjective judgment for further enzyme activity assays and bioactivity verification. This experience-driven screening method is widely used in agricultural applications such as root-knot nematode control, plant pathogen control, and insect control. However, existing screening techniques suffer from significant accuracy problems. The clear zone observation method is affected by various factors such as media homogeneity, light conditions, and observation angle, leading to substantial errors in measurement results. Furthermore, human experience-based judgment is highly subjective, and evaluation results from different researchers often differ significantly, lacking objective and unified evaluation standards. Traditional methods cannot deeply analyze the enzyme degradation mechanism and product characteristics, providing only a rough activity assessment and failing to accurately identify truly high-biological-activity superior enzyme strains. The core problem that traditional technologies cannot solve is the lack of precise molecular-level analysis methods and objective quantitative evaluation systems. The subjectivity of human experience and the limitations of traditional detection methods result in generally low accuracy of screening results, leading to the wrong elimination of many enzyme strains with excellent performance, while some enzyme strains with poor actual performance are mistakenly selected, which seriously affects the reliability and effectiveness of chitin-degrading enzyme screening. Summary of the Invention

[0003] In view of this, the present invention provides a method for screening chitin-degrading enzyme-producing bacteria for nematicides, which solves the technical problem that the screening of chitin-degrading enzymes in the prior art is often based on human experience and lacks accuracy.

[0004] This invention is implemented as follows: This invention provides a method for screening chitin-degrading enzyme-producing bacteria for nematicides, comprising collecting tobacco planting soil samples and preparing a bacterial suspension; mixing the tobacco planting soil samples with sterile distilled water and shaking, then serially diluting to the specified concentrations and spreading them onto a screening medium containing colloidal chitin for cultivation; observing the formation of clear zones on the screening medium, selecting strains with a clear zone diameter to colony diameter ratio greater than 3.0 as primary screening strains, and purifying the primary screening strains using the streak plate method; inoculating the purified primary screening strains into a liquid fermentation medium for cultivation, collecting the fermentation supernatant and obtaining a crude enzyme solution by centrifugation, and determining the chitinase activity and protein concentration of the crude enzyme solution; and performing liquid chromatography-mass spectrometry. The enzymatic hydrolysis products of crude enzyme solution were detected using advanced technology, and a mass spectrometry feature database of the hydrolysis products was established. The primary, secondary, and tertiary features of the mass spectrometry data were analyzed using a chitin degradation adaptation model. A two-layer game model was constructed, consisting of an upper-layer model aiming to maximize enzyme activity and a lower-layer model aiming to optimize the root-knot nematode killing effect. A multi-objective optimization algorithm was used to solve for the optimal combination of enzyme strain screening parameters. The root-knot nematode killing activity of candidate strains was tested using the optimal combination of enzyme strain screening parameters. The root-knot nematode suspension was mixed with crude enzyme solution of different concentrations and cultured, and the root-knot nematode mortality rate was statistically analyzed. Strains with a root-knot nematode mortality rate greater than 85% were selected as chitin-degrading enzyme producers with high nematicidal activity. Morphological identification was performed, and a strain preservation library was established.

[0005] The preparation step of the colloidal chitin specifically involves mixing chitin powder with 6 mol / L hydrochloric acid at a mass-volume ratio of 1:20 and stirring for 2 hours, adjusting the pH value to 7.0 with sodium hydroxide solution, and obtaining colloidal chitin after centrifugation and washing.

[0006] The liquid fermentation medium is composed of 15 g / L glucose, 8 g / L peptone, 3 g / L yeast extract, 2 g / L potassium dihydrogen phosphate, 0.5 g / L magnesium sulfate, and 10 g / L colloidal chitin. The culture conditions are 28°C, 180 rpm, and pH 6.5.

[0007] The chitin degradation adaptation model is specifically based on a sequence processing network with a Transformer-XL architecture, which includes an 8-layer encoder and a 4-layer decoder. Each encoder layer has 512-dimensional hidden states and 8 attention heads. The number of memory fragments is dynamically adjusted according to the complexity index of mass spectrometry data and the concentration of tobacco samples.

[0008] Specifically, the training dataset establishment step for the chitin degradation adaptation model involves collecting mass spectrometry data of 1000 different chitin-degrading enzymes as positive samples and collecting mass spectrometry data of 500 non-chitin-degrading enzymes as negative samples. The mass spectrometry data are then standardized according to mass-to-charge ratio and abundance to construct feature vectors.

[0009] Specifically, the training steps for the chitin degradation adaptation model involve updating model parameters using the Adam optimizer, setting the learning rate to 0.001, the batch size to 32, the number of training rounds to 200, calculating the model prediction error using the cross-entropy loss function, and preventing overfitting through an early stopping mechanism.

[0010] Specifically, the primary characteristic is the mass-to-charge ratio distribution of the molecular ion peak, which reflects the initial cleavage sites of chitin-degrading enzymes on chitin substrates and the molecular weight distribution of products. It is determined by the mass spectrum peak intensity integral value and retention time window.

[0011] Specifically, the secondary feature is the mass difference feature of fragment ions, which reflects the site preference and regularity of glycosidic bond breakage during chitin degradation and is obtained by calculating the mass difference between the parent ion and the daughter ion.

[0012] Specifically, the third-level feature is the chitin oligosaccharide chain length distribution feature, which characterizes the degradation ability of chitin-degrading enzymes on chitin chains with different degrees of polymerization and the uniformity of product chain length distribution, and is identified by the mass decrease law of continuous fragment ions.

[0013] Specifically, the memory fragment adjustment function is used to adjust the memory length parameter of the chitin degradation adaptation model. The memory fragment adjustment value is calculated based on the mass spectrometry data complexity index, tobacco sample concentration, enzymatic hydrolysis product concentration, and chitinase activity. Different numbers of memory fragments are used according to the range of the memory fragment adjustment value.

[0014] Specifically, the objective function of the upper-level model is a linear combination of enzyme activity and protein concentration, plus a logarithmic function of culture time, minus the square root function of chitin concentration, and then multiplied by a sine function of enzyme hydrolysis product concentration. The objective function of the lower-level model is a linear combination of root-knot nematode mortality rate, plus a squared term of enzyme solution concentration, minus an exponential function of treatment time, plus a cosine function of enzyme hydrolysis product concentration, and then multiplied by a coupling term of enzyme activity and protein concentration.

[0015] Specifically, the multi-objective optimization algorithm uses a non-dominated sorting genetic algorithm to solve the optimal solution of the two-level game model, and determines the optimal combination of enzyme screening parameters through Pareto front analysis.

[0016] The morphological identification specifically includes colony morphology observation, cell morphology examination, and physiological and biochemical characteristic determination, and the taxonomic position of the strain is determined by comparing morphological characteristics.

[0017] The mass spectrometry data complexity index is specifically calculated using the number of mass spectrometry peaks, the uniformity of peak intensity distribution, and the complexity of peak shape, and is used to guide the parameter settings of the memory fragment adjustment function.

[0018] Specifically, the gradient dilution concentration is... The dilution concentration is obtained by continuously diluting the strain suspension by 5 orders of magnitude, which is used to reduce the strain density to facilitate single colony isolation and purification.

[0019] This invention significantly improves the accuracy and reliability of chitin-degrading enzyme screening by establishing a high-precision screening system based on precise molecular analysis and intelligent algorithms, completely overcoming the subjectivity and inaccuracy of manual experience-based judgment. This invention uses liquid chromatography-mass spectrometry (LC-MS) to obtain precise molecular structure information of enzymatic hydrolysis products, and employs a chitin degradation adaptation model to perform objective multi-level feature analysis of mass spectrometry data. First-level features identify the mass-to-charge ratio distribution of molecular ion peaks; second-level features analyze the mass differences of fragment ions; and third-level features characterize the oligosaccharide chain length distribution, achieving precise quantitative evaluation of enzyme degradation characteristics, replacing the rough observation and subjective judgment of traditional methods. Simultaneously, a two-layer game model establishes a precise mathematical correlation between enzyme activity and biocidal effect, and a multi-objective optimization algorithm determines the optimal screening parameters, ensuring high accuracy and reproducibility of the screening results. This invention solves the technical problem that the screening of chitin-degrading enzymes in existing technologies often relies on manual experience, resulting in insufficient accuracy. Attached Figure Description

[0020] Figure 1 This is a flowchart of the method of the present invention.

[0021] Figure 2 The diagram shows the distribution of enzyme activity and protein concentration of different strains in the examples.

[0022] Figure 3 This is a graph showing the relationship between the mass spectrometry data complexity index and memory fragment adjustment in the examples.

[0023] Figure 4 This is a diagram illustrating how chitin-degrading enzymes disrupt the body wall of nematodes in the example. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0025] like Figure 1 The diagram shown is a flowchart of a method for screening chitin-degrading enzyme-producing bacteria for nematicides provided by the present invention. This method includes the following steps:

[0026] S01. Collect tobacco planting soil samples and prepare strain suspensions. Mix tobacco planting soil samples with sterile distilled water at a mass ratio of 1:10 and shake for 30 minutes. After serial dilution to the serial dilution concentration, spread the suspensions on screening medium containing colloidal chitin and incubate for 48 hours.

[0027] S02. Observe the formation of clear zones on the screening medium, select strains with a clear zone diameter to colony diameter ratio greater than 3.0 as primary screening strains, purify the primary screening strains using the streak plate method and store them at 4℃.

[0028] S03. The purified primary screening strain was inoculated into liquid fermentation medium and cultured for 72 hours. The fermentation supernatant was collected and the crude enzyme solution was obtained by centrifugation. The chitinase activity and protein concentration of the crude enzyme solution were determined.

[0029] S04. The enzymatic hydrolysis products of the crude enzyme solution were detected by liquid chromatography-mass spectrometry, a mass spectrometry feature database of the enzymatic hydrolysis products was established, and the primary, secondary and tertiary features of the mass spectrometry data were analyzed by chitin degradation adaptation model.

[0030] S05. Construct a two-layer game model with an upper-layer model aimed at maximizing enzyme activity and a lower-layer model aimed at optimizing the killing effect of root-knot nematodes. Solve the optimal combination of enzyme strain screening parameters through a multi-objective optimization algorithm.

[0031] S06. The candidate strains were tested for their root-knot nematode killing activity using the optimal enzyme strain screening parameter combination. The root-knot nematode suspension was mixed with crude enzyme solution of different concentrations and cultured for 24 hours. The root-knot nematode mortality rate was then counted.

[0032] S07. Select strains with a root-knot nematode mortality rate greater than 85% as chitin-degrading enzymes with high nematicidal activity, perform morphological identification, and establish a strain preservation library.

[0033] The colloidal chitin is prepared by neutralizing chitin powder with concentrated hydrochloric acid. The preparation process of colloidal chitin includes mixing chitin powder with 6 mol / L hydrochloric acid at a mass-volume ratio of 1:20 and stirring for 2 hours, adjusting the pH value to 7.0 with sodium hydroxide solution, and obtaining colloidal chitin after centrifugation and washing.

[0034] The liquid fermentation medium consists of 15 g / L glucose, 8 g / L peptone, 3 g / L yeast extract, 2 g / L potassium dihydrogen phosphate, 0.5 g / L magnesium sulfate, and 10 g / L colloidal chitin. The culture conditions are 28°C, 180 rpm, and pH 6.5.

[0035] The chitin degradation adaptation model is specifically structured as a sequence processing network based on the Transformer-XL architecture, comprising an 8-layer encoder and a 4-layer decoder. Each encoder layer has 512-dimensional hidden states and 8 attention heads. The number of memory segments is dynamically adjusted based on the complexity index of the mass spectrometry data and the concentration of tobacco samples, with a memory length ranging from 64 to 512 segments. The steps for establishing the training dataset for the chitin degradation adaptation model specifically include collecting mass spectrometry data of 1000 different chitin-degrading enzymes as positive samples and collecting mass spectrometry data of 500 non-chitin-degrading enzymes as negative samples. The mass spectrometry data are standardized according to mass-to-charge ratio and abundance to construct feature vectors. Data augmentation techniques are used to expand the training samples to 10,000. The training steps for the chitin degradation adaptation model specifically include using the Adam optimizer to update model parameters, setting the learning rate to 0.001, the batch size to 32, and the number of training epochs to 200. The cross-entropy loss function is used to calculate the model prediction error, and an early stopping mechanism is used to prevent overfitting.

[0036] In the chitin degradation adaptation model, the primary feature is the mass-to-charge ratio distribution of the molecular ion peak. This primary feature reflects the initial cleavage sites of the chitin-degrading enzyme on the chitin substrate and the molecular weight distribution of the products, determined by the mass spectrum peak intensity integral value and retention time window. The secondary feature is the mass difference feature of fragment ions. This secondary feature reflects the site preference and regularity of glycosidic bond breakage during chitin degradation, calculated by the mass difference between the parent ion and the daughter ion. The tertiary feature is the chitin oligosaccharide chain length distribution. This tertiary feature characterizes the degradation ability of the chitin-degrading enzyme on chitin chains with different degrees of polymerization and the uniformity of product chain length distribution, identified by the mass decrease pattern of continuous fragment ions.

[0037] The objective function of the upper-level model is a linear combination of enzyme activity and protein concentration, plus a logarithmic function of culture time, minus the square root function of chitin concentration, multiplied by a sine function of enzymatic hydrolysis product concentration. Specifically, it is expressed as follows: The objective function of the lower-level model is a linear term representing the root-knot nematode mortality rate, plus a squared term representing the enzyme concentration minus an exponential term representing the treatment time, plus a cosine term representing the concentration of the enzymatic hydrolysis products, multiplied by a coupling term between enzyme activity and protein concentration. Specifically, it is expressed as follows: The constraints of the upper-level model include enzyme activity E greater than 100 U / mL, protein concentration P less than 50 mg / mL, culture time T within the range of 24 to 96 hours, chitin concentration C within the range of 5 to 20 g / L, and enzyme hydrolysis product concentration D within the range of 1 to 100 μg / mL. The constraints of the lower-level model include root-knot nematode mortality M greater than 80%, enzyme concentration S within the range of 0.1 to 10 mg / mL, and treatment time T within the range of 12 to 48 hours. The coupling terms... This indicates the synergistic effect of enzyme activity and protein concentration on the killing effect of root-knot nematodes.

[0038] The feature grading recognition function is used to distinguish chitin degradation features at different levels in mass spectrometry data. The inputs include the mass spectrum peak intensity integral value, retention time window, mass-to-charge ratio distribution features, fragment ion mass difference, and chitin oligosaccharide chain length distribution features. The outputs are feature level identifiers and feature weight coefficients.

[0039] The memory fragment adjustment function is used to adjust the memory length parameter of the chitin degradation adaptation model. The memory fragment adjustment function calculates the memory fragment adjustment value based on the mass spectrometry data complexity index, tobacco sample concentration, enzymatic product concentration, and chitinase activity. When the memory fragment adjustment value is in the range of 0 to 0.25 (excluding 0.25), 64 memory fragments are used to improve the model's efficiency in processing simple mass spectrometry data. When the memory fragment adjustment value is in the range of 0.25 to 0.5 (excluding 0.5), 128 memory fragments are used to balance computational complexity and recognition accuracy. When the fragment adjustment value is in the range of 0.5 to 0.75 (excluding 0.75), 256 memory fragments are used to process mass spectrometry features of medium complexity. When the memory fragment adjustment value is in the range of 0.75 to 1.0, 512 memory fragments are used to ensure accurate analysis of high-complexity mass spectrometry data. The memory fragment adjustment function optimizes the chitin degradation adaptation model's ability to process mass spectrometry data of different complexities by dynamically adjusting the memory length parameter, thereby improving the accuracy of primary, secondary and tertiary feature identification, and thus improving the accuracy of enzyme strain screening and the reliability of root-knot nematode killing activity prediction.

[0040] The multi-objective optimization algorithm uses a non-dominated sorting genetic algorithm to solve the optimal solution of the two-layer game model, and determines the optimal combination of enzyme screening parameters through Pareto front analysis.

[0041] The morphological identification includes colony morphology observation, cell morphology examination, and physiological and biochemical characteristic determination. The taxonomic position of the strain is determined by comparing morphological characteristics.

[0042] The mass spectrometry data complexity index is calculated by the number of mass spectrometry peaks, the uniformity of peak intensity distribution, and the complexity of peak shape. The mass spectrometry data complexity index is used to guide the parameter setting of the memory fragment adjustment function.

[0043] The tobacco sample concentration is the concentration of tobacco root extract in the detection system, and the tobacco sample concentration is determined by the ratio of the dry weight of tobacco roots to the volume of extraction solvent.

[0044] The root-knot nematode is a type of nematode that damages the roots of tobacco plants, and the root-knot nematode suspension is obtained by isolating and purifying it from tobacco growing soil.

[0045] The mathematical expression for the objective function of the upper-level model is: ,in Indicates enzyme activity, Indicates protein concentration. Indicates the cultivation time. Indicates chitin concentration. Indicates the concentration of the enzymatic hydrolysis product. to These are the weighting coefficients.

[0046] The mathematical expression for the objective function of the lower-level model is as follows: ,in Indicates the mortality rate of root-knot nematodes. Indicates enzyme concentration. Indicates processing time. Indicates the concentration of the enzymatic hydrolysis product. Indicates enzyme activity, Indicates protein concentration. to and These are the weighting coefficients.

[0047] Wherein, the gradient dilution concentration is The gradient dilution concentrations are obtained by continuously diluting the strain suspension by 5 orders of magnitude. These gradient dilution concentrations are used to reduce the strain density to facilitate single colony isolation and purification.

[0048] The specific implementation methods of the above steps are described in detail below.

[0049] The specific implementation of step S01 involves first pretreating the tobacco planting soil sample. After removing plant debris and large particulate impurities, the collected soil sample is sieved and then mixed with sterile distilled water at a mass ratio of 1:10 to prepare a soil suspension. The suspension is then shaken at 200 rpm for 30 minutes using a horizontal shaker to fully disperse the microbial cells in the soil, releasing the strains attached to the soil particle surface into the liquid phase. Subsequently, a serial dilution method of 10-fold dilution is used to perform gradient dilution, continuously diluting the sample by five orders of magnitude. A 1:1 dilution was performed, which effectively reduced the strain density and facilitated subsequent single-colony isolation. The diluted strain suspension was evenly spread on the surface of a selection medium containing colloidal chitin. Colloidal chitin, as the sole carbon source, selectively cultured strains with chitin-degrading capabilities. The culture was incubated at 28°C for 48 hours, a duration that ensured sufficient strain growth while avoiding over-culturing and resulting colony confluence.

[0050] The specific implementation of step S02 involves assessing the chitin degradation ability of the strain by observing the formation of a clear zone. The mechanism of clear zone formation is that chitinase secreted by the strain hydrolyzes colloidal chitin in the culture medium into soluble oligosaccharides, resulting in a clear area forming around the colony in the originally turbid culture medium. The diameter of the clear zone and the colony diameter are precisely measured using calipers, and their ratio is calculated as a preliminary screening indicator for enzyme activity. Strains with a clear zone diameter to colony diameter ratio greater than 3.0 are selected as initial screening strains; this threshold ensures the selection of strains with strong chitin degradation ability. The initial screening strains are purified using the streak plate method. Single colony isolation is achieved through continuous streak dilution. After obtaining pure cultures, they are stored at 4°C to maintain strain viability and prevent contamination by other microorganisms.

[0051] The specific implementation of step S03 involves inoculating the purified, initially screened strain into a liquid fermentation medium for submerged fermentation. Liquid culture provides more abundant nutrient supply and oxygen transfer, which is beneficial for the large-scale reproduction of the strain and enzyme synthesis. The culture conditions were set at 28°C, 180 rpm, and pH 6.5, and cultured in a shaker for 72 hours to obtain maximum enzyme activity. After culture, the fermentation supernatant was collected by centrifugation at 8000g for 15 minutes to remove bacterial cells and obtain a crude enzyme solution containing extracellular enzymes. The chitinase activity of the crude enzyme solution was determined using the o-nitrophenol method, which is based on the principle of a colorimetric reaction between the reducing sugar produced by chitinase hydrolysis of the substrate and o-nitrophenol. Simultaneously, the Bradford protein quantification method was used to determine protein concentration, providing a reference for subsequent enzyme activity standardization.

[0052] The specific implementation of step S04 involves using liquid chromatography-mass spectrometry (LC-MS) to perform qualitative and quantitative analysis of the enzymatic hydrolysis products of the crude enzyme solution. This technique can accurately identify the molecular structure and concentration distribution of chitin degradation products. First, a mass spectrometry feature database of the hydrolysis products is established, collecting characteristic ion peak information generated by different chitin-degrading enzymes as a standard reference. Then, a chitin degradation adaptation model is used for in-depth analysis of the mass spectrometry data. This model, based on the Transformer-XL architecture, can handle complex serialized mass spectrometry data. The primary feature of the model analysis of the mass spectrometry data is mainly the mass-to-charge ratio distribution of molecular ion peaks, reflecting the initial cleavage sites of the chitin substrate and the molecular weight distribution of the products. The secondary feature is the mass difference of fragment ions, reflecting the site preference and regularity of glycosidic bond breakage patterns. The tertiary feature is the length distribution of chitin oligosaccharide chains, characterizing the degradation ability of chitin-degrading enzymes on chitin chains with different degrees of polymerization.

[0053] The specific implementation of step S05 involves constructing a two-layer game model to achieve multi-objective optimization of enzyme strain screening parameters. This model includes an upper-layer model aiming to maximize enzyme activity and a lower-layer model aiming to optimize the root-knot nematode killing effect. The objective function of the upper-layer model comprehensively considers factors such as enzyme activity, protein concentration, culture time, chitin concentration, and enzyme hydrolysis product concentration, achieving multivariate optimization through mathematical combinations of linear combination terms, logarithmic function terms, square root function terms, and sine function terms. The objective function of the lower-layer model uses the root-knot nematode mortality rate as the main indicator, combining enzyme solution concentration, treatment time, enzyme hydrolysis product concentration, and coupling terms of enzyme activity and protein concentration for comprehensive evaluation. A non-dominated sorting genetic algorithm is used to solve the two-layer game model. This algorithm can simultaneously handle multiple conflicting objective functions, and the optimal combination of enzyme strain screening parameters is determined through Pareto front analysis.

[0054] The specific implementation of step S06 involves using the obtained optimal enzyme strain screening parameter combination to test the root-knot nematode-killing activity of candidate strains. This test directly evaluates the biocontrol effect of chitin-degrading enzymes on target pests. Root-knot nematode suspensions are mixed with crude enzyme solutions of different concentrations at a set ratio and cultured at room temperature for 24 hours before nematode viability is detected. Microscopic observation is used to count the mortality rate of root-knot nematodes, and their life status is determined by their movement and morphological changes. Chitin-degrading enzymes cause nematode death by destroying the chitin structure of the nematode's body wall; this insecticidal mechanism is highly selective and environmentally friendly.

[0055] The specific implementation of step S07 involves screening highly effective strains based on the root-knot nematode killing activity test results and establishing a strain preservation library. Strains with a root-knot nematode mortality rate greater than 85% are selected as chitin-degrading enzymes with high nematicidal activity. This threshold ensures that the screened strains have significant biocontrol effects. Morphological identification of the selected superior strains is performed, including colony morphology observation, cell morphology examination, and physiological and biochemical characteristic determination. The taxonomic position of the strains is determined through morphological characteristic comparison and biochemical reaction patterns. When establishing the strain preservation library, a combination of freeze-drying and liquid nitrogen preservation methods is used to ensure the long-term stability and viability maintenance of the strains.

[0056] It should be noted that the chitin degradation adaptation model employs a sequence processing network based on the Transformer-XL architecture, which is specifically designed for processing long sequence data and possesses memory capabilities. The model comprises a deep network structure with 8 encoder layers and 4 decoder layers. Each encoder layer has 512 hidden states and 8 attention heads, enabling it to capture complex feature relationships in mass spectrometry data. The number of memory fragments is dynamically adjusted based on the complexity index of the mass spectrometry data and the concentration of the tobacco sample, with the memory length ranging from 64 to 512 fragments. Adaptive optimization is achieved through a memory fragment adjustment function.

[0057] The training dataset was first built by collecting mass spectrometry data of 1000 different chitin-degrading enzymes as positive samples. This data covered chitin-degrading enzymes from different sources and types, ensuring the model's generalization ability. Simultaneously, mass spectrometry data of 500 non-chitin-degrading enzymes were collected as negative samples, providing a basis for comparative learning. The mass spectrometry data were standardized according to mass-to-charge ratio and abundance to eliminate systematic errors caused by different instruments and experimental conditions, and then a multi-dimensional feature vector was constructed. Data augmentation techniques, including noise addition, peak shifting, and intensity adjustment, were used to expand the training samples to 10,000, improving the model's robustness and recognition accuracy.

[0058] The model training process uses the Adam optimizer for parameter updates, with a learning rate set to 0.001 to ensure convergence stability. A batch size of 32 is used to balance computational efficiency and memory consumption. The training epochs are set to 200, and the cross-entropy loss function is used to calculate the model's prediction error. An early stopping mechanism is employed to monitor validation set performance and prevent overfitting.

[0059] The chitin degradation adaptation model can deeply analyze multi-level feature information in mass spectrometry data. Compared with existing support vector machine classification methods, the chitin degradation adaptation model has stronger sequence modeling capabilities, capturing the temporal relationships and structural dependencies between mass spectrometry peaks. Compared with traditional artificial neural network methods, the model's Transformer-XL architecture introduces a memory mechanism, enabling it to handle longer mass spectrometry sequences and maintain long-distance feature correlations. The model's self-attention mechanism can automatically learn the importance weights between different mass spectrometry peaks, avoiding the subjectivity and limitations of manual feature selection. The dynamic memory fragment adjustment function allows the model to adaptively adjust the processing strategy according to data complexity, optimizing computational efficiency while ensuring recognition accuracy.

[0060] It should be noted that the first key technical idea of ​​this invention is to achieve multi-objective optimization of enzyme strain screening using a two-level game model. Traditional methods typically use a single indicator or simple weighting to evaluate strains, making it difficult to balance the relationship between enzyme activity and actual application effects. The two-level game model, by constructing a hierarchical structure where the upper-level model focuses on maximizing enzyme activity and the lower-level model focuses on optimizing insecticidal effect, can better simulate the complex decision-making process in actual applications. This model considers the coupling effect between enzyme activity and protein concentration, as well as the comprehensive influence of culture conditions on the final application effect, and has stronger systematicity and practicality compared to traditional linear optimization methods.

[0061] The second key technological approach is the design of a chitin degradation adaptation model based on the Transformer-XL architecture. Existing mass spectrometry data analysis methods mainly rely on peak matching and database retrieval, lacking in-depth mining of the inherent patterns in mass spectrometry data. The chitin degradation adaptation model, by introducing a self-attention mechanism and memory function, can automatically learn complex patterns and feature correlations in mass spectrometry data, achieving an end-to-end mapping from raw mass spectrometry data to degradation features. The model's dynamic memory adjustment mechanism adaptively adjusts the processing strategy according to data complexity, exhibiting better adaptability and accuracy compared to traditional neural networks with fixed parameters.

[0062] The third key technical approach is to establish a multi-level mass spectrometry feature identification system. Traditional mass spectrometry analysis typically focuses on single-level feature information, lacking a comprehensive consideration of the multi-dimensional characteristics of the degradation process. The first-, second-, and third-level feature identification systems constructed in this invention analyze chitin degradation characteristics from three levels: molecular ion peak distribution, fragment ion differences, and oligosaccharide chain length distribution, forming a complete feature description framework. This system can comprehensively reflect the catalytic properties and product characteristics of chitin-degrading enzymes, exhibiting higher identification accuracy and stronger discriminative power compared to single-feature analysis methods.

[0063] The synergistic effect of these key technological approaches forms a complete intelligent enzyme strain screening system. The chitin degradation adaptation model provides accurate mass spectrometry feature analysis capabilities, the multi-level feature recognition system ensures the comprehensiveness and accuracy of feature information, and the two-level game theory model achieves optimal configuration of screening parameters. These three elements support and promote each other, constructing a complete technological chain from data acquisition, feature extraction, pattern recognition to parameter optimization. Compared to traditional screening methods based on experience and single indicators, this synergistic system has a higher degree of automation, stronger systematicity, and better practicality, significantly improving the efficiency and accuracy of chitin-degrading enzyme screening and providing strong technical support for the industrial application of biocontrol technologies.

[0064] Traditional chitin-degrading enzyme screening methods rely heavily on manual experience and simple qualitative detection for enzymatic hydrolysis product analysis. This fails to accurately identify and quantify the complex composition of enzymatic hydrolysis products, leading to significant biases and inaccuracies in understanding the enzyme's catalytic properties. This invention establishes a high-precision liquid chromatography-mass spectrometry (LC-MS) analytical technique and constructs a standardized mass spectrometry feature database. This enables precise identification and quantification of the molecular structure of enzymatic hydrolysis products. First-level features accurately determine the mass-to-charge ratio distribution parameters of molecular ion peaks; second-level features accurately calculate the mass difference values ​​of fragment ions; and third-level features accurately characterize the length distribution characteristics of chitin oligosaccharide chains. This completely replaces the rough estimations and empirical judgments of traditional methods, establishing an enzyme characteristic analysis system based on precise data, significantly improving the accuracy and reliability of enzymatic hydrolysis product analysis. Existing screening technologies lack precise mathematical modeling methods when dealing with multi-factor optimization problems, relying mainly on researchers' empirical speculation and trial-and-error experiments to determine screening parameters. This results in low accuracy of parameter settings and unstable screening effects. This invention constructs a precise two-layer game theory mathematical model, transforming a complex multi-objective optimization problem into a mathematically solvable expression. The upper-layer model uses a polynomial function to accurately describe the quantitative relationship between enzyme activity and various influencing factors, while the lower-layer model accurately characterizes the changing patterns of biokilling effects using exponential and trigonometric functions. The model parameters are globally and precisely optimized using a non-dominated sorting genetic algorithm to obtain the Pareto optimal solution set, ensuring the optimality and accuracy of the screening parameters and providing a scientific and reliable technical guarantee for the high-precision screening of chitin-degrading enzymes.

[0065] Specifically, the principle of this invention is as follows: This invention solves the technical problem of insufficient accuracy in screening based on human experience. Its fundamental principle lies in establishing a high-precision analysis and evaluation system based entirely on objective data and precise algorithms, eliminating the inaccuracy of human experience-based judgment at its source. The fundamental reason for the low accuracy of traditional human experience-based screening methods is the lack of objective and accurate detection methods and scientifically reasonable evaluation standards. Visual observation of the transparent zone is affected by various environmental factors, resulting in significant measurement errors. Human experience-based judgment has obvious subjectivity and arbitrariness, failing to guarantee the accuracy and consistency of results. This invention, through liquid chromatography-mass spectrometry (LC-MS), can obtain precise molecular mass data of enzymatic hydrolysis products, with a detection accuracy reaching four decimal places, completely eliminating the measurement errors and human interference factors of traditional methods. The primary characteristics of the mass spectrometry data reflect the cleavage sites of the enzyme on the chitinous substrate and the molecular weight distribution of the products through precise mass-to-charge ratio determination. The secondary characteristics reveal the specific patterns of glycosidic bond breakage through precise mass difference calculations between the parent ion and daughter ion. The tertiary characteristics accurately identify the degree of polymerization distribution of oligosaccharide chains through the mass decrease pattern of continuous fragment ions. These precise molecular-level data provide an objective and reliable basis for enzyme performance evaluation. The chitin degradation adaptation model based on the Transformer-XL architecture learns from a large amount of training data using deep learning algorithms, establishing a precise mapping relationship between input mass spectrometry features and enzyme degradation performance. The model comprises an 8-layer encoder and a 4-layer decoder. Each encoder layer has 512 hidden states and 8 attention heads, enabling simultaneous processing of multi-dimensional feature vectors and automatic identification of complex feature patterns. Precise parameter tuning using the cross-entropy loss function and the Adam optimizer achieves a training accuracy of over 95%, significantly improving feature recognition accuracy. The two-layer game theory model couples enzyme activity optimization with biological effect optimization through precise mathematical modeling. The upper-layer model's objective function uses a multinomial combination form to comprehensively consider various influencing factors, while the lower-layer model uses exponential and trigonometric functions to precisely describe the changing patterns of biokilling effects. Model parameters are globally optimized using a non-dominated sorting genetic algorithm, ensuring optimal selection parameters and high accuracy of results.

[0066] The following provides a specific embodiment 1 of the present invention, and the specific implementation of each step in this embodiment 1 is described in detail below.

[0067] In this embodiment, the specific implementation of step S01 is the same as described above, and will not be repeated in detail here.

[0068] The specific implementation of step S02 is to assess the chitin degradation ability of the strain by evaluating the formation of the clear zone. The calculation method for the ratio of the clear zone diameter to the colony diameter is as follows:

[0069] ;

[0070] In the formula, This is the ratio of the diameter of the transparent zone to the diameter of the colony. The diameter of the transparent ring is in mm; This represents the colony diameter, in mm. The method for obtaining this parameter is as follows: The maximum diameter of the transparent ring was obtained by measuring it with vernier calipers. The maximum diameter of the colonies was obtained by measuring with vernier calipers. (Selection) Strains with a value greater than 3.0 were used as initial screening strains.

[0071] The specific implementation method of step S03 is the same as described above, and will not be repeated in detail here.

[0072] The specific implementation of step S04 involves using liquid chromatography-mass spectrometry (LC-MS) to analyze the enzymatic hydrolysis products of the crude enzyme solution. The calculation method for the mass spectrometry data complexity index is as follows:

[0073] ;

[0074] In the formula, Dimensionless, representing a measure of the complexity of mass spectrometry data; The number of mass spectrometry peaks is dimensionless. The peak intensity distribution uniformity is dimensionless. The degree of peak shape complexity is dimensionless; These are weighting coefficients, dimensionless, with values ​​ranging from 0.3 to 0.5, 0.2 to 0.4, and 0.1 to 0.3, respectively. The method for obtaining these parameters is as follows: Obtained automatically through mass spectrometry data processing software; The formula is obtained using the Shannon entropy calculation method. ,in For the first The relative intensity of each peak, dimensionless. The peak number; The peak width at half maximum (FWHM) and peak asymmetry factor are used to calculate the peak shape. The calculation formula is as follows: ,in The peak width at half maximum (FWHM) is expressed in seconds (s). The peak asymmetry factor is dimensionless. These are the weighting coefficients, with units of [missing information]. and dimensionless, with values ​​ranging from 0.5 to 0.8 and 0.2 to 0.5, respectively.

[0075] The specific calculation method for the memory segment adjustment function is as follows:

[0076] ;

[0077] In the formula, The adjustment value for the memory segment is dimensionless. Dimensionless, representing a measure of the complexity of mass spectrometry data; The concentration of the tobacco sample is expressed in mg / mL. The concentration of the enzymatic hydrolysis product is expressed in μg / mL. Chitinase activity is expressed in U / mL. These are weighting coefficients, with units of dimensionless, [missing information]. , , The values ​​range from 0.4 to 0.6, 0.1 to 0.3, 0.2 to 0.4, and 0.1 to 0.3, respectively. The method for obtaining these parameters is as follows: The value was calculated using the ratio of the dry weight of tobacco roots to the volume of the extraction solvent, and the formula is as follows: ,in This refers to the dry weight of tobacco roots, expressed in mg. The volume of the extraction solvent is expressed in mL. Quantitative analysis was performed using liquid chromatography-mass spectrometry. It was determined using the o-nitrophenol method.

[0078] The specific implementation of step S05 is to construct a two-layer game model, and the objective function of the upper-layer model is specifically expressed as follows:

[0079] ;

[0080] In the formula, This represents the objective function value of the upper-level model. Enzyme activity, expressed in U / mL; This refers to protein concentration, expressed in mg / mL. The incubation time is expressed in hours (h). Chitin concentration, in g / L; The concentration of the enzymatic hydrolysis product is expressed in μg / mL. These are weighting coefficients, with values ​​ranging from 0.3 to 0.5, 0.2 to 0.4, 0.1 to 0.3, and 0.1 to 0.2, respectively.

[0081] The objective function of the lower-level model is specifically expressed as follows:

[0082] ;

[0083] In the formula, The objective function value of the lower-level model; The mortality rate of root-knot nematodes is expressed as % . This refers to the enzyme concentration, expressed in mg / mL. Processing time, in hours (h); The concentration of the enzymatic hydrolysis product is expressed in μg / mL. Enzyme activity, expressed in U / mL; This refers to protein concentration, expressed in mg / mL. These are weighting coefficients, with values ​​ranging from 0.4 to 0.6, 0.1 to 0.3, 0.05 to 0.15, 0.1 to 0.2, and 0.2 to 0.4, respectively. The method for obtaining these parameters is as follows: The ratio of dead nematodes to the total number was obtained by microscopic observation. The parameters were obtained using the Bradford protein quantification method; other parameters were obtained using the same method as the upper model.

[0084] The specific implementation methods for steps S06-S07 are the same as those described above, and will not be repeated in detail here.

[0085] The method for calculating the solubility of chitin during the preparation of colloidal chitin is as follows:

[0086] ;

[0087] In the formula, The mass of dissolved chitin is expressed in g, and the chitin content in the supernatant was determined by gravimetric method. The total mass of chitin is expressed in grams, representing the total mass of chitin powder added. The method for obtaining these parameters is as follows: The chitin content in the supernatant was determined by gravimetric method. The total mass of chitin powder added.

[0088] The specific method for calculating gradient dilution concentrations is as follows:

[0089] ;

[0090] In the formula, This represents the concentration after gradient dilution, with units equal to... same; The initial bacterial suspension concentration is expressed in CFU / mL. The dilution factor is dimensionless. The parameter acquisition method is as follows: Colony forming units in the initial strain suspension were determined using the plate count method.

[0091] It should be noted that the formula for the ratio of the diameter of the transparent zone to the diameter of the colony is... The principle is based on the positive correlation between chitinase activity and clear zone formation. This ratio can eliminate the influence of differences in strain growth rate on enzyme activity evaluation. Compared with traditional screening methods that rely solely on clear zone size, this ratio standardization improves the accuracy and comparability of screening, avoids evaluation bias caused by differences in colony size, and makes enzyme activity assessment more objective and accurate.

[0092] Mass spectrometry data complexity index formula The principle lies in comprehensively considering the multidimensional characteristics of mass spectrometry data. The number of peaks reflects the richness of compound species, the uniformity of intensity distribution reflects the relative content distribution of each component, and the complexity of peak shape reflects the separation effect and matrix interference. The Shannon entropy calculation term in this formula is:

[0093] ;

[0094] The calculation item for peak shape complexity is:

[0095] ;

[0096] Compared with a single feature parameter, this comprehensive index can more accurately assess the complexity of mass spectrometry data, provide a more reliable basis for memory fragment adjustment, and improve the model's adaptability and processing accuracy to data of different complexities.

[0097] Memory fragment adjustment function The principle is to dynamically adjust the model's memory capacity based on the complexity characteristics of the input data, thereby optimizing the allocation of computational resources. This function considers multiple influencing factors such as mass spectrometry data complexity, sample concentration, product concentration, and enzyme activity, and determines the optimal number of memory segments through a weighted combination. The calculation term for tobacco sample concentration is as follows:

[0098] ;

[0099] Compared to traditional methods with fixed memory length, this dynamic adjustment mechanism can optimize computational efficiency while ensuring recognition accuracy, avoiding problems such as wasted computing resources and insufficient processing power.

[0100] upper-level model objective function The principle behind this approach lies in describing the nonlinear influence of various parameters on enzyme activity using different mathematical functions. The linear term reflects the direct contributions of enzyme activity and protein concentration, the logarithmic term embodies the diminishing marginal returns of culture time, and the product of the square root term and the sine function captures the complex interaction between chitin concentration and product concentration. Compared to traditional linear models, this nonlinear combined function can more accurately describe the complex behavior of biological systems, improving the practicality and reliability of the optimization results.

[0101] Lower-level model objective function The principle behind this approach is to comprehensively consider multiple influencing factors and their nonlinear relationships regarding insecticidal efficacy. The linear term directly reflects the main contribution to mortality, the square term reflects the nonlinear enhancing effect of enzyme concentration, the exponential term describes the rapid decay characteristics of treatment time, the cosine function term captures the periodic influence of product concentration, and the product term characterizes the synergistic effect of enzyme activity and protein concentration. Compared to single-index evaluation methods, this polynomial combination can more comprehensively reflect the influencing mechanism of insecticidal efficacy, improving the accuracy and practicality of the screening results.

[0102] To better understand and implement this invention, the following is an embodiment 2 of a specific application scenario: First, soil samples from 15 different plots were collected from a tobacco planting base, each weighing 200g. After pretreatment, the soil samples were mixed with sterile distilled water at a mass ratio of 1:10 to prepare a soil suspension. After horizontal shaking at 200 rpm for 30 minutes, a serial dilution method of 10-fold was used for gradient dilution, continuously diluting for 5 orders of magnitude to... The concentration was diluted several times. Colloidal chitin was prepared by mixing chitin powder with 6 mol / L hydrochloric acid at a mass-to-volume ratio of 1:20 and stirring for 2 hours. The pH was then adjusted to 7.0 with sodium hydroxide solution, followed by centrifugation and washing to obtain colloidal chitin. The diluted bacterial suspension was evenly spread onto selection medium containing colloidal chitin and incubated at 28°C for 48 hours.

[0103] After initial screening, 286 colonies were observed on the culture medium forming clear zones of varying sizes. The diameters of the clear zones and colonies were precisely measured using calipers, and the ratio was calculated. Forty-two strains with a clear zone diameter to colony diameter ratio greater than 3.0 were selected for preliminary screening. As shown in Table 1, the clear zone characteristics of these preliminary screening strains indicate good chitin degradation ability.

[0104] Table 1. Statistics on the clear zone characteristics of initially screened strains

[0105]

[0106] Forty-two initially screened bacterial strains were purified using the streak plating method, and the pure cultures were stored at 4°C. The purified strains were inoculated into liquid fermentation medium containing 15 g / L glucose, 8 g / L peptone, 3 g / L yeast extract, 2 g / L potassium dihydrogen phosphate, 0.5 g / L magnesium sulfate, and 10 g / L colloidal chitin. The culture conditions were set at 28°C, 180 rpm, and pH 6.5, and incubated on a shaker for 72 hours. After incubation, the fermentation supernatant was collected by centrifugation at 8000g for 15 minutes to obtain a crude enzyme solution containing extracellular enzymes.

[0107] The chitinase activity of the crude enzyme solution was determined using the o-nitrophenol method, and the protein concentration was determined using the Bradford protein quantification method. Figure 2 As shown, there were significant differences in enzyme activity among different strains, with strain TC-08 exhibiting the highest enzyme activity at 458 U / mL and a protein concentration of 42.6 mg / mL. Subsequently, liquid chromatography-mass spectrometry (LC-MS) was used to analyze the enzymatic hydrolysis products of strain TC-08, and a mass spectrometric database of the hydrolysis products was established.

[0108] The chitin degradation adaptation model is designed based on the Transformer-XL architecture, consisting of an 8-layer encoder and a 4-layer decoder. Each encoder layer has 512-dimensional hidden states and 8 attention heads. The training dataset contains mass spectrometry data of 1000 chitin-degrading enzymes as positive samples and mass spectrometry data of 500 non-chitin-degrading enzymes as negative samples, which are augmented to 10,000 training samples using data augmentation techniques. The model is trained using the Adam optimizer with a learning rate of 0.001, a batch size of 32, and 200 training epochs.

[0109] Mass spectrometry data of strain TC-08 were analyzed using a chitin degradation adaptation model. Primary characteristics showed that the molecular ion peaks were mainly distributed in the mass-to-charge ratio range of 200-800, reflecting the initial cleavage sites of the chitin substrate. Secondary characteristics indicated that the mass differences of fragment ions were concentrated at characteristic values ​​of 162, 203, and 365, reflecting a site preference for glycosidic bond breakage. Tertiary characteristics analysis showed that the length of chitin oligosaccharide chains was mainly distributed between 2 and 8 sugar units, characterizing the degradation ability of strain TC-08 for chitin chains with different degrees of polymerization. Figure 3 As shown, the mass spectrometry data complexity index is 0.76. The memory fragment adjustment function sets the memory length to 512 fragments to ensure accurate analysis of high-complexity mass spectrometry data.

[0110] A two-layer game theory model was constructed to optimize enzyme strain screening parameters. The objective function of the upper-layer model is: The enzyme activity (E) was 458 U / mL, the protein concentration (P) was 42.6 mg / mL, the culture time (T) was 72 hours, the chitin concentration (C) was 10 g / L, and the concentration of the enzymatic hydrolysis product (D) was 68.5 μg / mL. (Weighting coefficients are not specified.) to The values ​​are set to 0.6, 0.3, 0.8, and 0.4 respectively. The objective function of the lower-level model is... Weighting coefficient to and The values ​​were set to 1.2, 0.5, 0.2, 0.7, and 0.001 respectively.

[0111] A non-dominated sorting genetic algorithm was used to solve the two-level game model. The population size was set to 100, the number of generations to 300, the crossover probability to 0.8, and the mutation probability to 0.1. Pareto front analysis determined the optimal combination of screening parameters for the enzyme strain: 68 hours of culture time, chitin concentration to 12 g / L, pH to 6.8, temperature to 29℃, and rotation speed to 185 rpm. Under these optimized parameters, the enzyme activity of strain TC-08 increased to 526 U / mL, and the protein concentration to 45.3 mg / mL.

[0112] The optimized parameters were used to test the root-knot nematode killing activity of all candidate strains. Tobacco root-knot nematode suspension was isolated and purified from tobacco planting soil, and the nematode density was adjusted to 500 nematodes / mL. Different concentrations of crude enzyme solution were mixed with the nematode suspension at a 1:1 volume ratio, and the nematode mortality rate was recorded after culturing at 25℃ for 24 hours. The results showed that strain TC-08 achieved a 92.4% killing rate against tobacco root-knot nematodes at an enzyme concentration of 5 mg / mL, far exceeding the screening threshold of 85%.

[0113] As shown in Table 2, after testing for the killing activity against root-knot nematodes, the mortality rate of 8 strains exceeded 85%, among which strain TC-08 performed the best.

[0114] Table 2 Results of root-knot nematode killing activity test of selected strains

[0115]

[0116] Morphological identification was performed on the eight highly nematicidal bacterial strains selected. Colony morphology observation showed that strain TC-08 exhibited round, smooth, milky-white colonies with regular edges and a moist, glossy surface. Cell morphology examination indicated that this strain was a rod-shaped bacterium, with cell lengths of 2.8-4.2 μm and widths of 0.8-1.2 μm, and was Gram-positive. Physiological and biochemical characterization results showed that this strain was positive for catalase, amylase, gelatin liquefaction, and nitrate reduction. Based on morphological comparison, it was preliminarily identified as belonging to the genus Bacillus.

[0117] The bacterial strain bank was established using a combination of freeze-drying and liquid nitrogen preservation. For freeze-drying, 10% skim milk powder was used as a cryoprotectant; the strains were pre-frozen at -80°C for 2 hours and then vacuum freeze-dried for 12 hours. For liquid nitrogen preservation, 20% glycerol was used as a cryoprotectant, and the strains were stored long-term in liquid nitrogen at -196°C. The viability of the preserved strains was periodically monitored, and the results showed that the survival rate remained above 95% after 6 months of preservation.

[0118] The nematicidal mechanism of chitin-degrading enzymes in strain TC-08 was further verified. Scanning electron microscopy revealed significant damage and dissolution of the body wall of tobacco root-knot nematodes treated with the enzyme, and severe disruption of the chitin fiber structure. Figure 4 This study confirmed the mechanism by which chitin-degrading enzymes cause nematode death by destroying chitin in the nematode's body wall.

[0119] This invention represents a significant technological advancement over traditional nematode control methods. Traditional chemical nematicides kill nematodes through toxicity, easily leading to resistance and causing environmental pollution. In contrast, the chitin-degrading enzyme used in this invention achieves its insecticidal effect by specifically degrading the chitin structure of the nematode's body wall, exhibiting high selectivity and environmental friendliness. The chitin degradation adaptation model, based on a deep learning architecture, can accurately identify multi-level feature information in mass spectrometry data, offering higher accuracy and reliability compared to traditional enzyme activity assays. A two-level game theory model achieves synergistic optimization of enzyme activity and insecticidal effect, avoiding local optima problems that may result from single-objective optimization. The memory fragment adjustment function enables the model to adaptively adjust its processing strategy based on data complexity, improving screening efficiency and accuracy.

[0120] It should be noted that the variables involved in this invention are explained in detail in Table 3 below.

[0121] Table 3. Variable Explanation Table

[0122]

[0123] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for screening chitin-degrading enzyme-producing bacteria for use in nematicides, characterized in that, The process included collecting tobacco-growing soil samples and preparing bacterial suspensions. The soil samples were mixed with sterile distilled water, shaken, and serially diluted to specific concentrations before being plated onto a selection medium containing colloidal chitin. The formation of clear zones on the selection medium was observed, and strains with a clear zone diameter to colony diameter ratio greater than 3.0 were selected as primary screening strains. These strains were purified using the streak plate method. The purified strains were inoculated into liquid fermentation medium, and the fermentation supernatant was collected and centrifuged to obtain crude enzyme solution. The chitinase activity and protein concentration of the crude enzyme solution were measured. Liquid chromatography-mass spectrometry (LC-MS) was used to detect the enzymatic hydrolysis products of the crude enzyme solution, establishing an enzymatic hydrolysis... A mass spectrometry feature database of the product was established, and the primary, secondary, and tertiary features of the mass spectrometry data were analyzed using a chitin degradation adaptation model. A two-layer game model was constructed, consisting of an upper-layer model aimed at maximizing enzyme activity and a lower-layer model aimed at optimizing the root-knot nematode killing effect. The optimal combination of enzyme strain screening parameters was solved using a multi-objective optimization algorithm. The root-knot nematode killing activity of candidate strains was tested using the optimal combination of enzyme strain screening parameters. After mixing root-knot nematode suspension with crude enzyme solutions of different concentrations and culturing, the root-knot nematode mortality rate was statistically analyzed. Strains with a root-knot nematode mortality rate greater than 85% were selected as chitin-degrading enzyme producing strains with high nematicidal activity. Morphological identification was performed, and a strain preservation library was established.

2. The method for screening chitin-degrading enzyme-producing bacteria for nematicides according to claim 1, characterized in that, The preparation steps of the colloidal chitin are as follows: chitin powder is mixed with 6 mol / L hydrochloric acid at a mass-volume ratio of 1:20 and stirred for 2 hours. Then, the pH value is adjusted to 7.0 with sodium hydroxide solution, and the colloidal chitin is obtained after centrifugation and washing.

3. The method for screening chitin-degrading enzyme-producing bacteria for nematicides according to claim 2, characterized in that, The liquid fermentation medium is composed of 15 g / L glucose, 8 g / L peptone, 3 g / L yeast extract, 2 g / L potassium dihydrogen phosphate, 0.5 g / L magnesium sulfate, and 10 g / L colloidal chitin. The culture conditions are 28°C, 180 rpm, and pH 6.

5.

4. The method for screening chitin-degrading enzyme-producing bacteria for nematicides according to claim 3, characterized in that, The chitin degradation adaptation model is specifically based on a sequence processing network with a Transformer-XL architecture, which includes an 8-layer encoder and a 4-layer decoder. Each encoder layer has 512-dimensional hidden states and 8 attention heads. The number of memory fragments is dynamically adjusted according to the complexity index of mass spectrometry data and the concentration of tobacco samples.

5. The method for screening chitin-degrading enzyme-producing bacteria for nematicides according to claim 4, characterized in that, The steps for establishing the training dataset for the chitin degradation adaptation model are as follows: collect mass spectrometry data of 1000 different chitin-degrading enzymes as positive samples, collect mass spectrometry data of 500 non-chitin-degrading enzymes as negative samples, and construct feature vectors after standardizing the mass spectrometry data according to mass-to-charge ratio and abundance.

6. The method for screening chitin-degrading enzyme-producing bacteria for nematicides according to claim 5, characterized in that, The training steps for the chitin degradation adaptation model specifically involve updating model parameters using the Adam optimizer, setting the learning rate to 0.001, the batch size to 32, and the number of training rounds to 200. The model prediction error is calculated using the cross-entropy loss function, and overfitting is prevented through an early stopping mechanism.

7. The method for screening chitin-degrading enzyme-producing bacteria for nematicides according to claim 6, characterized in that, The primary characteristic mentioned above is specifically the mass-to-charge ratio distribution of the molecular ion peak, which reflects the initial cleavage sites of chitin-degrading enzymes on chitin substrates and the molecular weight distribution of products. It is determined by the mass spectrum peak intensity integral value and retention time window.

8. The method for screening chitin-degrading enzyme-producing bacteria for nematicides according to claim 7, characterized in that, The secondary characteristic, specifically the mass difference characteristic of fragment ions, reflects the site preference and regularity of glycosidic bond breakage during chitin degradation, and is obtained by calculating the mass difference between the parent ion and the daughter ion.

9. The method for screening chitin-degrading enzyme-producing bacteria for nematicides according to claim 8, characterized in that, The aforementioned third-level characteristic is specifically the chitin oligosaccharide chain length distribution characteristic, which characterizes the degradation ability of chitin-degrading enzymes on chitin chains with different degrees of polymerization and the uniformity of product chain length distribution, and is identified by the mass decrease law of continuous fragment ions.

10. The method for screening chitin-degrading enzyme-producing bacteria for nematicides according to claim 9, characterized in that, The memory fragment adjustment function is specifically used to adjust the memory length parameter of the chitin degradation adaptation model. The memory fragment adjustment value is calculated based on the mass spectrometry data complexity index, tobacco sample concentration, enzymatic hydrolysis product concentration, and chitinase activity. Different numbers of memory fragments are used according to the range of the memory fragment adjustment value.

Images