Loofah stock and bitter gourd scion matching and screening method

Abstract

The invention provides a loofah rootstock and bitter gourd scion matching and screening method. The loofah rootstock and bitter gourd scion matching and screening method comprises the following steps that S101, a preliminary genetic difference classification result is obtained; step S102, determining a potential adaptive rootstock candidate set; s103, aiming at each individual in the rootstock candidate set, obtaining an expression potential evaluation result of a disease-resistant genome; step S104, according to the expression potential evaluation result of the disease-resistant genome, determining the disease-resistant ability sequence of each rootstock individual; s105, obtaining a prediction index of the plant growth potential; step S106, determining a priority sequence of the yield performance values; step S107, obtaining an evaluation result of quality stability; step S108, determining a final preferred stock list; and step S109, obtaining a continuously optimized stock screening basis. According to the method, the towel gourd rootstock and the bitter gourd scion are accurately matched, the grafting success rate and the plant growth performance are improved, and a scientific rootstock selection method is provided for bitter gourd planting.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of agricultural production, in particular to a matching screening method for luffa stock and balsam pear scion. BACKGROUND

[0002] In agricultural production, vegetable grafting technology as an important means to improve crop yield and stress resistance has attracted widespread attention, especially in the cultivation of economic crops such as balsam pear, the application of grafting technology has key significance to guarantee yield and quality.

[0003] Through grafting, the excellent traits of balsam pear can be combined with the disease resistance and adaptability of the stock, thereby achieving higher economic benefits. However, in current production, the selection and matching of different stocks often have many uncertainties, which directly affect the grafting effect and the final crop performance. The existing grafting methods mostly rely on experience judgment or simple phenotypic observation to select stocks. This approach often seems inadequate when faced with complex environmental conditions and disease pressure, especially in the genetic compatibility of stock and scion and the stability of resistance gene expression, lacking in-depth scientific basis, leading to different performances in yield and disease resistance of grafted plants, and even large fluctuations. This uncertainty not only increases production costs, but also limits the application of grafting technology in a wider range.

[0004] In this field, the core challenge is how to accurately identify and select the most suitable stock type for grafting to ensure the growth potential and disease resistance of balsam pear plants after grafting. First, due to the large genetic background difference of different luffa stocks, it is difficult to accurately judge the matching degree of luffa stocks and balsam pear scions in the early stage, which directly affects the survival rate of grafting and the overall performance of the plants in the later stage. Secondly, this genetic difference further leads to uncertainty in the effective expression of resistance genes, making some stocks unable to provide adequate protection when facing specific diseases, thereby affecting the yield and quality of balsam pear. SUMMARY

[0005] The technical problem to be solved by the present application is to provide a matching screening method for luffa stock and balsam pear scion to solve the problem of how to accurately identify and select the most suitable stock type for grafting.

[0006] To solve the above technical problems, the present application provides the following technical solutions: A matching screening method for luffa stock and balsam pear scion, comprising the following steps: Step S101: obtaining genetic difference related data from luffa stock samples by genome sequencing, constructing a genotype feature database for different stock individuals, and obtaining a preliminary genetic difference classification result; Step S102: Analyze the characteristics of the rootstock matching degree of the luffa rootstock and the bitter gourd scion using a data comparison method, combine the pre-set matching degree threshold for preliminary screening, and determine the potential adaptive rootstock candidate set; Step S103: For each individual in the rootstock candidate set, extract the disease-resistant genome-related information from the genotype feature database, analyze the distribution of disease-resistant genes through information comparison technology, and obtain the expression potential evaluation result of the disease-resistant genome; Step S104: According to the expression potential evaluation result of the disease-resistant genome, combine the environmental adaptability-related data, and use data analysis method to predict the disease resistance value of the rootstock in a specific environment, and determine the disease resistance ranking of each rootstock individual; Step S105: According to the disease resistance value ranking result, obtain the historical data related to plant growth potential, simulate the growth potential of the grafted bitter gourd plant through data modeling technology, and obtain the prediction index of plant growth potential; Step S106: According to the prediction index of plant growth potential, combine the yield performance value related data, and use comprehensive evaluation method to quantitatively analyze the yield potential of each rootstock individual after grafting, and determine the priority ranking of yield performance value; Step S107: According to the priority ranking of yield performance value, obtain the quality stability related data, predict the quality fluctuation range of the grafted bitter gourd through data comparison technology, and obtain the evaluation result of quality stability; Step S108: According to the evaluation result of quality stability, combine the screening accuracy related data, and use information integration technology to comprehensively score all rootstock candidate individuals, and determine the final preferred rootstock list; Step S109: For the final preferred rootstock list, use data recording technology to store the screening process and results in the database, and through the automatic updating mechanism, supplement the data of subsequent grafting experiments, and obtain the continuously optimized rootstock screening basis.

[0007] Preferably, the step S101 specifically comprises the following steps: S1011: Complete preliminary data collection by performing genome sequencing on the luffa rootstock sample to obtain raw genomic sequence data; S1012: According to the obtained genomic sequence data, use data extraction process to separate the key fragments related to genetic differences, and obtain the difference fragment set; S1013: For the difference fragment set, apply support vector machine algorithm for feature selection, extract genotype features with significant discrimination, and determine the feature identifier set; S1014: If the data distribution in the feature identifier set meets the pre-set threshold condition, it is recorded in the pre-established feature database, and the storage of the genotype features is completed; S1015: Obtain genetic difference grouping of different rootstock individuals by clustering analysis on the genotype features in the feature database, and obtain a preliminary individual classification result; S1016: According to the individual classification result, analyze the genetic difference mode between groups in combination with the difference fragment set, and judge whether there is a significant classification boundary; S1017: If the classification boundary is clear, use the classification result to generate the structured data of genetic difference analysis, and complete the final classification mapping.

[0008] Preferably, the step S102 specifically comprises the following steps: S1021: Obtain the initial feature set of the luffa rootstock and the bitter gourd scion from the classification result through the genetic difference data, and use the data comparison method to extract features to obtain the preliminary distribution of the matching degree features; S1022: According to the preliminary distribution of the matching degree features, use the support vector machine algorithm to classify the features for the genetic difference of the luffa rootstock and the bitter gourd scion, determine the correlation weight between the features, and obtain the classified feature combination; S1023: Obtain the classified feature combination, and perform quantitative analysis on the matching degree features in combination with the preset threshold value, if the matching degree feature value is higher than the preset threshold value, it is determined as a potential adaptive combination, and a preliminary screened matching pair is obtained; S1024: For the preliminary screened matching pair, perform deep comparison on the genetic difference of the luffa rootstock and the bitter gourd scion through the data comparison method, judge whether there is hidden matching degree feature, and obtain the matching pair set after deep analysis; S1025: According to the matching pair set after deep analysis, use the clustering analysis method to group the potential adaptive combination, determine the matching degree feature distribution of the rootstock candidate and the scion in each group, and obtain the grouped candidate set; S1026: Through the grouped candidate set, for the matching degree feature distribution in each group, if the feature distribution consistency of a group is higher than the preset threshold value, it is determined as a rootstock candidate with higher adaptability, and a final adaptive candidate set is obtained; S1027: Obtain the final adaptive candidate set, sort each group of rootstock candidates in combination with the genetic difference and the matching degree feature, judge the stability of the matching degree feature in the sorting result, and determine the optimal rootstock candidate combination.

[0009] Preferably, the step S103 specifically comprises the following steps: S1031: Obtain the genotype feature data of the rootstock candidate from the database, and perform preliminary information sorting for each individual to obtain a basic data set of individual genotypes; S1032: According to the basic data set, the distribution of disease-resistant genes of each individual is analyzed by using information comparison technology, and the position distribution information of the disease-resistant genes in the genome is determined; S1033: If the distribution density of the disease-resistant genes in the position distribution information is higher than the preset threshold, the corresponding individual is marked as a high-potential individual, and the genomic information data set of the high-potential individual is obtained; S1034: For the genomic information data set of the high-potential individual, the correlation between the disease-resistant genes and the genotype characteristics is analyzed, and the expression potential strength of the disease-resistant genes is judged; S1035: According to the analysis result of the expression potential strength, the disease-resistant genome of each high-potential individual is comprehensively scored in combination with the distribution situation data, and a comprehensive potential evaluation value is obtained; S1036: The comprehensive potential evaluation value is used to sort all high-potential individuals, and a final disease-resistant gene expression potential priority list is determined; S1037: The individuals in the priority list are verified and analyzed by using a random forest algorithm, the reliability of the evaluation result is judged in combination with the correlation data of the genotype characteristics and the disease-resistant genes, and a final disease-resistant genome potential analysis data set is output.

[0010] Preferably, the step S104 specifically includes the following steps: S1041: The preliminary evaluation result of the expression potential is obtained by obtaining the related data of the disease-resistant genes from the database and combining the original records of gene expression; S1042: According to the preliminary evaluation result of the expression potential, the data of environmental adaptation and adaptation information are fused to construct a comprehensive influence factor data set in a specific environment; S1043: The comprehensive influence factor data set is processed by using a random forest algorithm, the prediction value of disease resistance is calculated, and the resistance value of each rootstock individual is obtained; S1044: The resistance value of each rootstock individual is standardized to obtain a quantitative index of individual ability; S1045: The ability of all rootstock individuals is sorted by using the quantitative index of individual ability and combining data analysis methods to determine the sorting result; S1046: If there are rootstock individuals with similar resistance values in the sorting result, the final individual ability sorting is determined by comparing the environmental adaptation data in a specific environment; S1047: After obtaining the final individual ability sorting, a corresponding disease-resistant ability distribution map is generated to determine the disease-resistant ability positioning of each rootstock individual in a specific environment.

[0011] Preferably, the step S105 specifically includes the following steps: S1051: Construct an initial dataset by obtaining relevant data from disease resistance ranking results; S1052: Preprocess the ranking results using data cleaning techniques to remove outliers and missing values, obtaining a sorted resistance dataset; S1053: Extract historical data associated with plant growth from the sorted resistance dataset; S1054: Perform feature selection on the historical data to filter out highly correlated feature variables with growth potential, and determine the feature dataset; S1055: Construct input data for growth simulation by combining the feature dataset with the grafting treatment information of the bitter gourd plant; S1056: If some variable values in the feature dataset are below the preset threshold, standardize them to obtain normalized simulation input data; S1057: Train the normalized simulation input data using a random forest model to simulate the growth potential of the bitter gourd plant and obtain preliminary growth potential prediction values; S1058: Correct the prediction results based on the preliminary growth potential prediction values and the plant growth trend in the historical data; S1059: If the deviation between the prediction value and the historical trend exceeds the preset range, smooth the prediction value to obtain the corrected growth potential value; S10510: Generate a prediction index of plant growth potential based on the corrected growth potential value; S10511: Perform multi-dimensional analysis on the prediction index to determine its stability under different environmental conditions and obtain the final prediction result.

[0012] Preferably, the step S106 specifically includes the following steps: S1061: Form an initial dataset by obtaining plant growth potential index data and yield performance data from the database, and complete the preliminary integration of data; S1062: Based on the initial dataset, use a pre-established comprehensive evaluation model to quantitatively process the grafting yield potential of each rootstock individual, obtaining a potential evaluation result; S1063: For the potential evaluation result, if the potential value of a rootstock individual exceeds the preset threshold, mark it as a high-potential individual to complete the preliminary screening; S1064: Obtain the correlation coefficient between the high-potential individual's yield performance data and growth indicators by performing in-depth analysis, and determine the key influencing factors; S1065: Based on the key influencing factors, use a random forest model to predict the yield performance value and obtain the predicted yield ranking of each rootstock individual; S1066: For the predicted yield ranking, if the ranking of a certain rootstock individual is in the top, further extract its growth analysis data to determine its stability characteristics; S1067: Through the combination of stability characteristics and yield potential data, the final priority ranking result is generated, and the whole process analysis is completed.

[0013] Preferably, the step S107 specifically comprises the following steps: S1071: By obtaining the yield performance data and quality data of grafted bitter gourd from historical records, an initial data set is constructed to obtain basic information for subsequent analysis; S1072: According to the initial data set, the yield performance value is prioritized, and a high performance value data subset is selected by using a preset threshold to determine the key analysis object; S1073: By standardizing the quality data in the high performance value data subset, the distribution characteristics of the quality data are obtained, and the preliminary fluctuation range is determined; S1074: If the preliminary fluctuation range exceeds the preset threshold, further data comparison analysis is performed on the quality data, and the support vector machine algorithm is used to accurately predict the fluctuation range to obtain the predicted fluctuation value; S1075: According to the predicted fluctuation value, combined with the historical quality stability record of grafted bitter gourd, the intermediate index of stability evaluation is obtained, and the influence degree of fluctuation range on stability is determined; S1076: By comparing the intermediate index with the preset stability evaluation standard, it is judged whether the quality stability of grafted bitter gourd meets the expectation, and the final evaluation result is obtained.

[0014] Preferably, the step S108 specifically comprises the following steps: S1081: Obtain the quality stability data and screening accuracy data of the rootstock candidate individual, and integrate the related indexes from multiple sources through the pre-established data acquisition system to form an initial data set; S1082: For the initial data set, the quality stability data and screening accuracy data are standardized by using information integration technology to obtain a unified evaluation data matrix; S1083: According to the evaluation data matrix, each rootstock candidate individual is scored by using a preset scoring system, and if the score of a certain candidate individual is lower than a preset threshold, it is removed from the candidate list to obtain a preliminary screening result; S1084: Through individual analysis of the preliminary screening result, combined with data fusion technology, the weighted comprehensive score value of the remaining rootstock candidate individual is determined by weighted calculation of multiple indexes; S1085: According to the weighted comprehensive score value, the remaining stock candidate individuals are sorted, if the comprehensive score value of a candidate individual is higher than the preset upper threshold value, it is marked as a priority object, and a sorted candidate list is obtained; S1086: For the sorted candidate list, a decision tree algorithm is applied to the priority object for final verification to determine whether it meets the requirements of the preferred list, and a verified preferred individual set is output; S1087: By data comparison of the verified preferred individual set, an information integration technology is used to generate a final preferred stock list, and the final result meeting the quality stability and screening accuracy is determined.

[0015] Preferably, the step S109 specifically comprises the following steps: S1091: Related information of the stock list is obtained from the screening process by a data acquisition tool, and is stored in a pre-established database in a structured format to obtain an initial stock screening data set; S1092: According to the initial stock screening data set, the screening process information recorded therein is classified and processed, and the key indicators are extracted by using the preset field rules to determine the core basis data in the screening process; S1093: If the core basis data meets the preset threshold condition, the latest experimental data is obtained from the grafting experiment by an automatic updating mechanism, and is supplemented to the database to determine whether the existing stock list is adjusted; S1094: If the experimental data is supplemented and found that the indicators of part of the stock list change, a random forest algorithm is used to predict and analyze the change trend to obtain the potential optimization direction of the stock list; S1095: According to the optimization direction obtained by the prediction analysis, the subsequent experimental data stored in the database is compared and verified to obtain experimental results consistent with the optimization direction, and an adjustment scheme of the stock list is determined; S1096: The stock list in the database is updated by the adjustment scheme, the updated list is associated with the screening basis to obtain the latest screening reference data in combination with the continuous improvement goal; S1097: According to the latest screening reference data, an automatic data recording method is used to save it to the database to form a closed-loop data supplement and optimization process, and the continuous improvement effect of the stock screening basis is determined.

[0016] Compared with the prior art, the present application has at least the following beneficial effects: The application realizes precise matching of silk gourd rootstock and bitter gourd scion, improves the grafting success rate and plant growth performance, and provides a scientific rootstock selection method for bitter gourd planting. BRIEF DESCRIPTION OF DRAWINGS

[0017] The accompanying drawings, which are incorporated herein and constitute part of the specification, illustrate embodiments of the present disclosure and, together with the description, further serve to explain the principles of the present disclosure and to enable a person skilled in the relevant art to implement and use the present disclosure.

[0018] Fig. 1 A flowchart of the process of the silk gourd rootstock and bitter gourd scion matching screening method.

[0019] Fig. 2 A flowchart of step S102 of the silk gourd rootstock and bitter gourd scion matching screening method.

[0020] Fig. 3 A flowchart of step S109 of the silk gourd rootstock and bitter gourd scion matching screening method. DETAILED DESCRIPTION

[0021] A silk gourd rootstock and bitter gourd scion matching screening method provided by the present application will be described in detail below in combination with the drawings and specific embodiments. It should be noted that, in order to make the embodiments more detailed, the following embodiments are the best, preferred embodiments, and other alternative ways can also be implemented by those skilled in the art for some known technologies; and the drawings are only used to more specifically describe the embodiments, and are not intended to specifically limit the present application.

[0022] It should be noted that in the specification, “one embodiment”, “an embodiment”, “exemplary embodiment”, “some embodiments” and the like indicate that the described embodiment can include a specific feature, structure or property, but not necessarily every embodiment includes the specific feature, structure or property. In addition, when a specific feature, structure or property is described in combination with an embodiment, it should be within the knowledge of those skilled in the relevant art to realize this feature, structure or property in combination with other embodiments (whether or not explicitly described).

[0023] As shown in Figs. 1-3 The embodiment of the present application provides a silk gourd rootstock and bitter gourd scion matching screening method, which comprises the following steps: Step S101: Obtain genetic difference related data from the genome sequencing of the luffa stock sample, construct a genotype feature database for different stock individuals, and obtain a preliminary genetic difference classification result; the specific steps include the following: S1011: Perform genome sequencing on the luffa stock sample to obtain raw genomic sequence data and complete preliminary data collection; S1012: According to the obtained genomic sequence data, use the data extraction process to separate the key fragments related to genetic differences, and obtain a difference fragment set; S1013: For the difference fragment set, apply the support vector machine algorithm for feature selection to extract genotype features with significant discrimination, and determine the feature identifier set; S1014: If the data distribution in the feature identifier set meets the preset threshold condition, it is recorded in the pre-established feature database to complete the storage of genotype features; S1015: Cluster the genotype features in the feature database to obtain the genetic difference grouping of different stock individuals and obtain the preliminary individual classification result; S1016: According to the individual classification result, analyze the genetic difference patterns between groups in combination with the difference fragment set, and determine whether there is a significant classification boundary; S1017: If the classification boundary is clear, use the classification result to generate structured data for genetic difference analysis, and complete the final classification mapping.

[0024] Specific application examples of step S101: For example, when performing genome sequencing on the luffa stock sample, raw sequence data can be obtained through high-throughput sequencing technology.

[0025] Suppose that one sequencing covers more than 95% of the sample genome, and about 1 billion short read sequences are obtained, which lays a foundation for subsequent analysis and ensures the comprehensiveness and reliability of the data.

[0026] In the data extraction process, for the separation of genetic difference related fragments, the sequence can be matched with the reference genome using alignment tools to screen out variant sites such as single nucleotide polymorphism fragments.

[0027] Suppose that about 5 million difference fragments are extracted from the 1 billion sequences, which provides key materials for subsequent feature selection.

[0028] When applying the support vector machine algorithm for feature selection on the difference fragment set, a classification model can be constructed to identify genotype features with significant discrimination for stock genetic differences.

[0029] Assuming 1000 key features are selected from 5 million fragments, forming a feature identification set.

[0030] This step effectively reduces data dimensionality and improves analysis efficiency.

[0031] If the data distribution of the feature identification set meets the preset threshold condition, such as the correlation between features being less than 0.2, it will be recorded in the feature database.

[0032] This process ensures the standardized storage of data, facilitating subsequent calling and analysis.

[0033] Through clustering analysis of genotypic features in the feature database, the K-means algorithm can be used to divide different rootstock individuals into 3 genetic difference groups.

[0034] Assuming that 100 individuals in the sample are divided into groups A, B, and C, each containing 40, 30, and 30 individuals, respectively, the preliminary classification result provides a basis for subsequent difference pattern analysis.

[0035] Based on the individual classification results, combined with the difference fragment set analysis of genetic difference patterns between groups, it can be found that group A and group B have significant differences at certain genetic loci, while group C shows an intermediate state.

[0036] If statistical analysis determines that the classification boundary is clear, such as the group difference significance level reaching below 0.05, the effectiveness of classification can be confirmed.

[0037] If the classification boundary is clear, structured data for genetic difference analysis is generated, such as recording the feature distribution and difference site information of each group in table form, and finally completing the classification mapping.

[0038] This result can be used to guide rootstock breeding and improve breeding accuracy.

[0039] For example, from another perspective, during feature screening, if the support vector machine algorithm is combined with other machine learning methods such as random forest for verification, the accuracy of the feature identification set can be further improved.

[0040] Assuming that the overlap rate of the screening results of the two methods reaches 80%, the reliability of the features is enhanced.

[0041] For example, in clustering analysis, if combined with genetic distance matrix grouping verification, assuming that the genetic distance between group A and group B is 0.3, and the distance between group C is 0.15, the classification result has more biological significance.

[0042] These multi-angle verifications collectively support the scientificity of classification, providing a solid foundation for subsequent applications, while significantly improving the practical value of genetic difference analysis.

[0043] Step S102: According to the genetic difference classification result, the data comparison method is used to analyze the compatibility characteristics of the luffa rootstock and the bitter gourd scion, and a preliminary screening is performed combined with the preset compatibility threshold to determine the potential adaptive rootstock candidate set; Specifically, the following steps are included: S1021: From the classification result, the initial feature set of the luffa rootstock and the bitter gourd scion is obtained from the genetic difference data, and the data comparison method is used for feature extraction to obtain the preliminary distribution of the compatibility characteristics; S1022: According to the preliminary distribution of the compatibility characteristics, the genetic differences between the luffa rootstock and the bitter gourd scion are classified using the support vector machine algorithm to determine the correlation weight between the characteristics and obtain the classified feature combination; S1023: The classified feature combination is obtained, and the compatibility characteristics are quantitatively analyzed combined with the preset threshold; if the compatibility characteristic value is higher than the preset threshold, it is determined as a potential adaptive combination, and the preliminary screening of the matching pair is obtained; S1024: For the preliminary screening of the matching pair, the genetic differences between the luffa rootstock and the bitter gourd scion are compared in depth through the data comparison method to determine whether there are hidden compatibility characteristics, and the matching pair set after depth analysis is obtained; S1025: According to the matching pair set after depth analysis, the clustering analysis method is used to group the potential adaptive combinations to determine the compatibility characteristic distribution of the rootstock candidate and the scion in each group, and the grouped candidate set is obtained; S1026: Through the grouped candidate set, for the compatibility characteristic distribution in each group, if the consistency of the feature distribution of a certain group is higher than the preset threshold, it is determined as a rootstock candidate with higher adaptability, and the final adaptive candidate set is obtained; S1027: The final adaptive candidate set is obtained, and each group of rootstock candidates is sorted combined with the genetic difference and the compatibility characteristics to determine the stability of the compatibility characteristics in the sorting result and determine the optimal rootstock candidate combination.

[0044] Specific application examples of step S102: For example, in the study of the genetic adaptability of luffa rootstock and bitter gourd scion, the genetic difference data can be obtained, and the screening and optimization of the matching pair can be gradually deepened.

[0045] For the acquisition of the initial feature set, the genomic data of the luffa rootstock and the bitter gourd scion can be compared to extract key difference points such as single nucleotide polymorphism sites to form a preliminary set containing hundreds of features.

[0046] These features may include the length difference of gene fragments or the variation frequency of specific sites, laying a foundation for subsequent analysis.

[0047] For example, in feature extraction and matching degree distribution analysis, a preliminary statistics can be conducted on the extracted features. Assuming that the mutation rate of a certain gene fragment of luffa stock sample A is 0.05, and the corresponding mutation rate of balsam pear scion sample B is 0.07, the matching degree score of the two can be calculated by comparison method, which may be 0.85, and it is preliminarily judged that it has certain adaptation potential.

[0048] This method helps to quickly screen out possible adaptive combinations.

[0049] For example, for feature classification of support vector machine algorithm, the correlation between features can be weighted. Assuming that the influence weight of a certain feature on classification is 0.6, and that of another feature is only 0.2, the feature with higher weight is given priority.

[0050] This way can effectively distinguish which genetic differences have greater impact on adaptability, providing basis for subsequent screening.

[0051] In matching degree feature quantization analysis, a threshold such as 0.8 can be set. If the matching degree score of a certain stock and scion is 0.82, it will be included in the potential adaptive combination.

[0052] This quantitative method helps to reduce subjective judgment and improve screening efficiency.

[0053] For deep alignment and hidden feature mining, more detailed genome fragment alignment can be used to find some small differences not covered by the initial feature set, such as insertion or deletion of a certain gene, and then adjust the matching pair set.

[0054] This deep analysis can improve the comprehensiveness of adaptability judgment.

[0055] In cluster analysis and determination of grouping candidate set, combinations with similar matching degrees can be grouped. Assuming that the matching degree scores of 5 combinations in a group are between 0.80 and 0.85, and the consistency is high, it can be considered that the group has good adaptation potential.

[0056] This grouping method helps to focus on high adaptability combinations.

[0057] For sorting and stability judgment of the final candidate set, the matching degree features of each stock candidate can be evaluated in multiple dimensions, such as the stability of genetic differences. Assuming that the matching degree fluctuation of a certain combination in multiple tests is only 0.02, it is considered to have high stability, and is preferred as the optimal candidate.

[0058] This sorting method can ensure that the finally selected combination has high reliability.

[0059] Through the above multi-level analysis and screening, not only the adaptability of luffa stock and bitter gourd scion can be accurately identified from the perspective of genetic difference, but also data support can be provided for subsequent grafting experiments, and the adaptability efficiency and success rate of grafting combination can be significantly improved.

[0060] Step S103: For each individual in the candidate set of stock, extract the disease-resistant genome related information from the genotype feature database, analyze the distribution of disease-resistant genes by information matching technology, and obtain the expression potential evaluation result of disease-resistant genome; Specifically, the following steps are included: S1031: Obtain the genotype feature data of the candidate stock from the database, and preliminarily organize the data for each individual to obtain the basic data set of individual genotype; S1032: According to the basic data set, the distribution of disease-resistant genes in each individual is analyzed by information matching technology to determine the position distribution information of disease-resistant genes in the genome; S1033: If the distribution density of disease-resistant genes in the position distribution information is higher than the preset threshold, mark the corresponding individual as a high potential individual, and obtain the genome information data set of the high potential individual; S1034: For the genome information data set of the high potential individual, analyze the correlation between the disease-resistant genes and the genotype characteristics, and judge the expression potential strength of the disease-resistant genes; S1035: According to the analysis result of the expression potential strength, combined with the distribution situation data, the disease-resistant genome of each high potential individual is comprehensively scored to obtain a comprehensive potential evaluation value; S1036: Through the comprehensive potential evaluation value, all high potential individuals are sorted to determine the final disease-resistant gene expression potential priority list; S1037: The individuals in the priority list are verified and analyzed by using random forest algorithm, the correlation data of genotype characteristics and disease-resistant genes are combined to judge the reliability of the evaluation result, and the final disease-resistant genome potential analysis data set is output.

[0061] Specific application examples of step S103: For example, in the study of the adaptability of luffa stock and bitter gourd scion, when obtaining the genotype feature data of the candidate stock from the database, the specific fragment information in the gene sequence can be focused on first.

[0062] Suppose the database contains 100 luffa stock candidate individuals, and the genotype data of each individual includes marker point information on chromosomes. By organizing these data, a basic data set containing genotype markers and individual numbers is formed, laying a foundation for subsequent analysis.

[0063] Specifically, when analyzing the distribution of disease resistance genes using information comparison techniques, the location of disease resistance genes in the genome can be determined by comparing the genome sequence of each individual with a known database of disease resistance genes.

[0064] For example, a certain individual of the luffa rootstock has 5 disease resistance gene markers found in its genome, distributed on chromosomes 2 and 5, with 3 markers on chromosome 2, with a distribution density of 1.5 markers per 1 million base pairs.

[0065] If the preset threshold is 1.2 markers per 1 million base pairs, this individual is marked as a high-potential individual, and its genomic information will be further extracted for in-depth study.

[0066] For example, when analyzing the association between disease resistance genes and genotype characteristics of high-potential individuals, attention can be paid to the characteristics of gene expression regulation regions.

[0067] Suppose there is an enhanced signal in the promoter region near the disease resistance genes of a certain high-potential individual, indicating that its expression potential is relatively high.

[0068] By comparing the characteristics of the regulation regions of 10 high-potential individuals, it is found that 6 of them have similar enhanced signals, with expression potential scores above 8.5, while the scores of the other individuals are around 6.0, indicating that the former are more valuable for application.

[0069] Specifically, the calculation of the comprehensive potential evaluation value can combine the distribution density and the expression potential score.

[0070] For example, set the distribution density proportion to 40% and the expression potential proportion to 60%. The comprehensive score of a certain individual is calculated to be 8.2, while the score of another individual is 7.5. Finally, according to the score from high to low, the former has a higher priority.

[0071] This approach helps to screen out rootstock candidates with higher disease resistance potential.

[0072] For example, when verifying the priority list using the random forest algorithm, a feature matrix can be constructed, including genotype markers and disease resistance gene association data, to make a reliability judgment on the evaluation results.

[0073] Suppose the verification result shows that 4 out of the top 5 individuals have highly consistent predicted results with actual genotype characteristics, with a reliability score of more than 85%, indicating that the evaluation system is relatively reliable.

[0074] This verification method can improve the scientificity of the screening results and provide reliable basis for subsequent research on the adaptability of rootstock and scion.

[0075] Specifically, the output of the disease resistance genome potential analysis dataset can be presented in tabular form, including information such as individual ID, comprehensive score, and priority.

[0076] For example, the individual ranked first in the output dataset has a comprehensive score of 8.8, and its disease resistance gene distribution density and expression potential are both excellent, making it a key candidate rootstock.

[0077] This intuitive presentation method helps researchers quickly identify target individuals, improving research efficiency.

[0078] Step S104: Based on the evaluation results of the expression potential of the disease-resistant genome and combined with environmental adaptability-related data, the disease resistance value of rootstocks under specific environments is predicted using data analysis methods to determine the ranking of disease resistance ability for each individual rootstock; specifically, this includes the following steps: S1041: By obtaining relevant data on disease resistance genes from the database and combining them with the original records of gene expression, a preliminary assessment of expression potential is compiled. S1042: Based on the preliminary assessment results of expression potential, construct a comprehensive impact factor dataset for specific environments by integrating data on environmental adaptation and adaptation information; S1043: The random forest algorithm is used to process the comprehensive impact factor dataset, calculate the predicted values ​​of disease resistance, and obtain the resistance value of each rootstock individual; S1044: For the resistance values ​​of each individual rootstock, perform data standardization processing to obtain quantitative indicators of individual capabilities; S1045: By using quantitative indicators of individual capabilities and combining data analysis methods, rank all rootstock individuals by capability and determine the ranking results; S1046: If there are rootstock individuals with similar resistance values ​​in the ranking results, the final individual ability ranking is determined by comparing environmental adaptation data under a specific environment. S1047: After obtaining the final individual ability ranking, generate the corresponding disease resistance distribution map to determine the disease resistance location of each rootstock individual in a specific environment.

[0079] Specific application example of step S104: For example, when obtaining disease resistance gene-related data from a database, one can first focus on the genotype information in the rootstock candidate set and extract gene fragment data that are directly related to disease resistance.

[0080] Suppose the database stores the genetic information of 1,000 rootstock individuals, including specific gene loci data related to disease resistance. By screening out these loci, the number and distribution characteristics of disease resistance genes in each individual can be preliminarily identified, laying the foundation for subsequent expression potential assessment.

[0081] In one possible implementation, when combining the original records of gene expression, the gene expression level data measured in the laboratory can be matched with the database information.

[0082] For example, for a certain rootstock individual, it is found that the expression level of its disease resistance gene under a certain disease pressure is twice the conventional level, indicating that it has stronger potential disease resistance.

[0083] Such data integration helps to more accurately evaluate the expression potential.

[0084] For example, when constructing a comprehensive impact factor dataset by integrating environmental adaptation information, environmental variables such as temperature and humidity can be considered to affect gene expression.

[0085] Suppose that under high temperature and high humidity environment, the expression of disease resistance gene of a certain rootstock individual is inhibited, while under mild environment, the expression is significantly enhanced. By recording these difference data, a correlation model between environment and disease resistance is constructed, providing basis for subsequent prediction.

[0086] In one possible implementation, when processing the comprehensive impact factor dataset using a random forest algorithm, environmental variables and gene expression data can be used as input features to predict the disease resistance value of each rootstock individual.

[0087] Suppose that after algorithm analysis, the predicted resistance value of a certain rootstock individual is 85, while that of another individual is 60, which provides a quantitative basis for subsequent ranking.

[0088] This method can effectively integrate multi-dimensional data and improve the reliability of prediction.

[0089] For example, when standardizing the individual resistance value data, all individual values can be mapped to the range of 0 to 100, facilitating intuitive comparison.

[0090] Suppose that the original resistance value of a certain rootstock individual is 75, and after standardization, it is 80. Such processing helps to eliminate the dimensional difference of data and improve the fairness of comparison.

[0091] In one possible implementation, when ranking the ability of rootstock individuals, if it is found that the resistance values of two individuals are both 82, further comparison of their adaptation data under certain environment, such as drought tolerance or low temperature tolerance data, is needed. Suppose that one individual performs more stably under high temperature environment, then it has higher priority.

[0092] Such detailed comparison can ensure that the ranking result is more suitable for actual application scenarios.

[0093] For example, when generating the disease resistance distribution map, the resistance values of all rootstock individuals can be displayed in a column chart, intuitively presenting the positioning of each individual's ability.

[0094] If a rootstock individual is located in the top 10% in the distribution map, it indicates that its disease resistance ability in a specific environment is outstanding. This visualization method facilitates the rapid identification of potential individuals and provides a reference for subsequent breeding selection.

[0095] Through the above multi-dimensional analysis and processing, the comprehensiveness and practicality of rootstock disease resistance ability evaluation can be effectively improved.

[0096] Step S105: For the disease resistance value sorting result, obtain the historical data related to plant growth potential, simulate the growth potential of grafted bitter gourd plants through data modeling technology, and obtain the prediction index of plant growth potential; Specifically, the following steps are included: S1051: Obtain relevant data from the disease resistance sorting result, and construct an initial data set; S1052: Preprocess the sorting result using data cleaning technology to remove outliers and missing values, and obtain the sorted resistance data set; S1053: Extract historical data associated with plant growth from the sorted resistance data set; S1054: Feature selection is performed on the historical data to filter out highly correlated feature variables related to growth potential, and a feature data set is determined; S1055: Through the feature data set, combine the grafting information of bitter gourd plants to construct the input data for growth simulation; S1056: If the value of some variables in the feature data set is lower than the preset threshold, standardize it to obtain normalized simulation input data; S1057: Train the normalized simulation input data using a random forest model, simulate the growth potential of bitter gourd plants, and obtain preliminary growth potential prediction values; S1058: Correct the prediction results based on the preliminary growth potential prediction values and the plant growth trend in the historical data; S1059: If the deviation of the prediction value from the historical trend exceeds the preset range, smooth the prediction value to obtain the corrected growth potential value; S10510: Generate the prediction index of plant growth potential through the corrected growth potential value; S10511: Perform multi-dimensional analysis on the prediction index to determine its stability under different environmental conditions, and obtain the final prediction result.

[0097] Specific application examples of step S105: For example, when constructing the initial dataset, the resistance values of the bitter gourd rootstocks and the related environmental parameters can be extracted from the disease resistance ranking results. Assuming there are 100 rootstock samples, each containing a resistance value and environmental humidity data, a basic table containing multi-dimensional information is formed.

[0098] This approach ensures the integrity of the data for subsequent analysis.

[0099] In one possible implementation, the application of data cleaning techniques can handle outliers and missing values.

[0100] If the resistance value of a certain rootstock sample is 3 times higher than the average, it can be considered as an outlier and removed; For missing environmental data, the average value of adjacent samples can be used to fill in the missing data.

[0101] This preprocessing helps to improve the reliability of the dataset.

[0102] For example, when extracting historical data related to plant growth, the growth records of bitter gourd plants over the past 3 years can be retrieved from the database, including information such as plant height, leaf number, etc. Correlation analysis with the resistance dataset can be performed to select indicators that have a significant impact on growth.

[0103] This correlation analysis lays the foundation for subsequent feature selection.

[0104] In one possible implementation, feature selection can focus on variables that are highly related to growth potential, such as soil pH, light duration, etc. Assuming 5 key features are selected to form a feature dataset.

[0105] This selection can reduce the interference of irrelevant variables and improve simulation accuracy.

[0106] For example, when constructing growth simulation input data combined with the grafting information of bitter gourd plants, the grafting method, grafting time, etc. can be integrated with the feature dataset. Assuming that a batch of rootstocks uses top grafting method, the growth status 30 days after grafting is recorded as input parameters.

[0107] This integration approach helps to simulate a more realistic growth scenario.

[0108] In one possible implementation, for variables in the feature dataset that are below the threshold, such as samples with light duration less than 8 hours / day, standardization can be used to adjust them to a unified range.

[0109] This standardization ensures the consistency of the data and provides a stable foundation for model training.

[0110] For example, when training with a random forest model, the normalized input data can be divided into training and test sets to simulate the growth potential of bitter gourd plants in different environments, obtaining the predicted value of each sample.

[0111] This method can capture the complex relationship between data and improve the accuracy of prediction.

[0112] In one possible implementation, when correcting the prediction result, if the preliminary prediction value shows that the growth potential of a certain rootstock is 20 cm per year, but the historical trend is 15 cm, the weighted average method can be used for adjustment.

[0113] This correction method makes the prediction more in line with the actual growth pattern.

[0114] For example, when smoothing the predicted value, if the deviation exceeds the preset range of 10%, the moving average method can be used to smooth the abnormal fluctuations to obtain more stable growth potential values.

[0115] This processing helps to reduce the noise effect in prediction.

[0116] In one possible implementation, when generating the plant growth potential prediction index, the corrected potential value can be converted into a grade index, such as high, medium, and low, to facilitate intuitive judgment.

[0117] This index method provides a clear reference for subsequent decision-making.

[0118] For example, when performing multi-dimensional analysis, the growth stability of bitter gourd plants under different humidity and temperature conditions can be simulated, assuming that the growth potential decreases by 5% under a humidity of 80%, and performs best at a temperature of 25 degrees.

[0119] This analysis helps to identify the key points of environmental impact on growth and provides a basis for planting optimization.

[0120] Step S106: According to the prediction index of plant growth potential, combined with the yield performance value related data, the yield potential of each rootstock individual after grafting is quantitatively analyzed by using a comprehensive evaluation method, and the priority ranking of yield performance value is determined; Specifically, the following steps are included: S1061: Obtain the prediction index data and yield performance data of plant growth potential from the database to form an initial data set, and complete the preliminary integration of data; S1062: According to the initial data set, the comprehensive evaluation model established in advance is used to quantitatively process the grafting yield potential of each rootstock individual, and the potential evaluation result is obtained; S1063: For the potential evaluation result, if the potential value of a certain rootstock individual exceeds the preset threshold, it is marked as a high-potential individual, and the preliminary screening is completed; S1064: Obtain the correlation coefficient between the yield performance data of high-potential individuals and the growth indicators by in-depth analysis of the yield performance data of high-potential individuals, and determine the key influencing factors; S1065: According to the key influencing factors, use the random forest model to predict the yield performance value, and obtain the predicted yield ranking of each rootstock individual; S1066: For the predicted yield ranking, if the ranking of a rootstock individual is in the top, further extract its growth analysis data to judge its stability characteristics; S1067: Combine the stability characteristics with the yield potential data to generate the final priority ranking result, and complete the whole process analysis.

[0121] Specific application examples of step S106: For example, when obtaining plant growth potential prediction index data and yield performance data from the database, the relevant records of bitter gourd rootstocks in the past two years can be preferentially called, including information such as plant height growth rate, leaf expansion speed, and fruit yield per plant.

[0122] Suppose there are 200 rootstock samples, each containing growth potential prediction value and actual yield data, forming a multi-dimensional initial data set.

[0123] This way can ensure the comprehensiveness of the data, laying the foundation for subsequent integration.

[0124] For example, after completing the preliminary integration of data, the growth potential data and yield data can be matched according to the rootstock individuals, and incomplete records can be excluded.

[0125] Suppose that in a batch of data, 10 samples are excluded due to record loss, ensuring that the integrated data set contains 190 valid samples.

[0126] This integration method provides reliable data support for subsequent evaluation.

[0127] For example, the comprehensive evaluation model can be used to quantify the grafting yield potential based on a pre-set weight system, combining growth potential indicators and historical yield performance.

[0128] Suppose the growth potential prediction value of a rootstock individual is 85, and the historical average yield is 3.5 kg per plant, and the potential value calculated by the model is 88, which is higher than the average level.

[0129] This quantitative method helps to intuitively reflect the individual potential.

[0130] For example, when marking high-potential individuals in the potential evaluation results, a threshold of 80 can be set, and if the potential value of a rootstock is 88, it is marked as a high-potential individual.

[0131] Further screening 30 high-potential individuals to concentrate resources for follow-up analysis.

[0132] This screening mechanism can quickly lock in high-quality samples.

[0133] For example, when conducting in-depth analysis of yield performance data for high-potential individuals, the correlation coefficient with growth indicators can be calculated.

[0134] Assuming the height growth rate of a certain rootstock is 0.75, indicating that height is a key influencing factor.

[0135] This analysis helps to clarify the optimization direction.

[0136] For example, when using a random forest model to predict yield performance values, key factors such as height and leaf number can be used as input parameters.

[0137] Assuming that 30 high-potential individuals are predicted, the predicted yield of a certain rootstock is 4.2 kg per plant, ranking first.

[0138] This prediction method can provide a scientific basis for ranking.

[0139] For example, when extracting growth analysis data for rootstock individuals ranked first in predicted yield ranking to determine stability characteristics, attention can be paid to their performance in different environments.

[0140] Assuming that the height fluctuation of a certain rootstock under humidity changes is less than 5%, indicating that it has strong stability.

[0141] This stability analysis provides an important reference for final decision-making.

[0142] For example, when generating priority ranking results by combining stability characteristics and yield potential data, predicted yield and stability performance can be considered comprehensively.

[0143] Assuming that a certain rootstock has a predicted yield of 4.2 kg and a stability fluctuation of less than 5%, it ranks first in priority.

[0144] This full-process analysis ensures the comprehensiveness and reliability of the results.

[0145] Step S107: For priority ranking of yield performance values, obtain quality stability-related data, predict the quality fluctuation range of grafted bitter gourd through data comparison techniques, and obtain the evaluation results of quality stability; Specifically, the following steps are included: S1071: Obtain yield performance data and quality data of grafted bitter gourd from historical records to construct an initial data set and obtain basic information for subsequent analysis; S1072: According to the initial data set, the yield performance value is prioritized, and a high performance value data subset is screened out by using a preset threshold value, and a key analysis object is determined; S1073: The quality data in the high performance value data subset is standardized to obtain the distribution characteristics of the quality data, and the preliminary fluctuation range is determined; S1074: If the preliminary fluctuation range exceeds the preset threshold value, further data comparison and analysis of the quality data is performed, the support vector machine algorithm is used to accurately predict the fluctuation range, and the predicted fluctuation value is obtained; S1075: According to the predicted fluctuation value, combined with the historical quality stability record of grafted bitter gourd, the intermediate index of stability evaluation is obtained, and the influence degree of fluctuation range on stability is determined; S1076: By comparing the intermediate index with the preset stability evaluation standard, it is judged whether the quality stability of grafted bitter gourd meets the expectation, and the final evaluation result is obtained.

[0146] Step S108: According to the evaluation result of quality stability, combined with the screening accuracy related data, the information integration technology is used to comprehensively score all rootstock candidate individuals, and the final preferred rootstock list is determined; Specifically, the following steps are included: S1081: Obtain the quality stability data and screening accuracy data of the rootstock candidate individuals, integrate related indexes from multiple sources through a pre-established data acquisition system, and form an initial data set; S1082: According to the initial data set, the quality stability data and screening accuracy data are standardized by using information integration technology, and a unified evaluation data matrix is obtained; S1083: According to the evaluation data matrix, the preset scoring system is used to comprehensively score each rootstock candidate individual, if the score of a certain candidate individual is lower than the preset threshold value, it is removed from the candidate list, and the preliminary screening result is obtained; S1084: Through analysis of the individuals in the preliminary screening result, combined with data fusion technology, the weighted comprehensive score value of the remaining rootstock candidate individuals is determined by weighted calculation of multiple indexes; S1085: According to the weighted comprehensive score value, the remaining rootstock candidate individuals are sorted, if the comprehensive score value of a certain candidate individual is higher than the preset upper threshold value, it is marked as a priority consideration object, and the sorted candidate list is obtained; S1086: For the sorted candidate list, the decision tree algorithm is applied to the final verification of the priority consideration object, to determine whether it meets the requirements of the preferred list, and the verified preferred individual set is output; S1087: Generate the final preferred rootstock list by using information integration technology through data comparison of the verified preferred individual set, and determine the final result meeting the quality stability and screening accuracy.

[0147] Step S109: For the final preferred rootstock list, use data recording technology to store the screening process and results in the database, and supplement the data of subsequent grafting experiments through an automatic updating mechanism to obtain continuously optimized rootstock screening basis; specifically including the following steps: S1091: Obtain the relevant information of the rootstock list from the screening process through the data collection tool, and store it in the pre-established database after being arranged in a structured format to obtain the initial rootstock screening data set; S1092: According to the initial rootstock screening data set, classify the screening process information recorded therein, and extract key indicators using pre-set field rules to determine the core basis data in the screening process; S1093: If the core basis data meets the pre-set threshold condition, then the latest experimental data is obtained from the grafting experiment through an automatic updating mechanism, supplemented to the database, and whether the existing rootstock list is adjusted is determined; S1094: If the experimental data is supplemented and found that the indicators of part of the rootstock list have changed, then a random forest algorithm is used to predict and analyze the change trend to obtain the potential optimization direction of the rootstock list; S1095: According to the optimization direction obtained by the prediction analysis, the subsequent experimental data stored in the database is compared and verified to obtain experimental results consistent with the optimization direction, and an adjustment scheme for the rootstock list is determined; S1096: Update the rootstock list in the database through the adjustment scheme, associate the updated list with the screening basis to obtain the latest screening reference data in combination with the continuous improvement goal; S1097: For the latest screening reference data, use an automatic data recording method to save it to the database to form a closed-loop data supplement and optimization process, and determine the continuous improvement effect of the rootstock screening basis.

[0148] The present application encompasses any substitutions, modifications, equivalent methods and schemes made on the essence and scope of the present application. In order for the public to have a thorough understanding of the present application, specific details are described in the following preferred embodiments of the present application, and the present application can also be completely understood without these details by those skilled in the art. In addition, in order to avoid unnecessary confusion to the essence of the present application, well-known methods, processes, procedures, elements and circuits, etc. are not described in detail.

[0149] Those skilled in the art can understand that all or part of the steps in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a program, and the program can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, and the like.

[0150] The above only describes the preferred embodiments of the present application, and it should be pointed out that those skilled in the art can make several improvements and refinements without departing from the principles of the present application, and these improvements and refinements should also be considered as the protection scope of the present application.

Claims

1. A method for matching and screening loofah rootstock and bitter gourd scion, characterized in that, Includes the following steps: Step S101: By performing genome sequencing on loofah rootstock samples, genetic difference-related data are obtained, and a genotype characteristic database is constructed for different rootstock individuals to obtain preliminary genetic difference classification results; Step S102: Use data comparison method to analyze the rootstock matching characteristics of loofah rootstock and bitter gourd scion, and perform preliminary screening in combination with preset matching threshold to determine the potential candidate set of rootstocks. Step S103: For each individual in the rootstock candidate set, extract disease resistance genome-related information from the genotype feature database, analyze the distribution of disease resistance genes through information comparison technology, and obtain the expression potential assessment results of the disease resistance genome; Step S104: Based on the evaluation results of the expression potential of the disease-resistant genome and combined with environmental adaptability-related data, the disease resistance value of the rootstock under specific environments is predicted using data analysis methods, and the disease resistance ability ranking of each individual rootstock is determined. Step S105: Based on the disease resistance value ranking results, obtain historical data related to plant growth potential, simulate the growth potential of grafted bitter gourd plants using data modeling technology, and obtain predictive indicators of plant growth potential. Step S106: Based on the predicted indicators of plant growth vigor and combined with relevant data on yield performance, a comprehensive evaluation method is used to quantitatively analyze the post-grafting yield potential of each rootstock individual and determine the priority ranking of yield performance values. Step S107: Prioritize the yield performance values, obtain relevant data on quality stability, predict the quality fluctuation range of grafted bitter gourd through data comparison technology, and obtain the evaluation results of quality stability. Step S108: Based on the evaluation results of quality stability and combined with the relevant data on screening accuracy, use information integration technology to comprehensively score all candidate rootstocks and determine the final list of preferred rootstocks. Step S109: For the final list of preferred rootstocks, data recording technology is used to store the screening process and results in the database. The data of subsequent grafting experiments are supplemented through an automatic update mechanism to obtain continuously optimized rootstock screening criteria.

2. The method for matching and screening loofah rootstock and bitter gourd scion according to claim 1, characterized in that, Step S101 specifically includes the following steps: S1011: By performing genome sequencing on loofah rootstock samples, the raw genome sequence data was obtained, and preliminary data collection was completed; S1012: Based on the obtained genome sequence data, a data extraction process is used to separate key fragments related to genetic differences, resulting in a set of differentially expressed fragments; S1013: For the set of differential fragments, the support vector machine algorithm is applied to perform feature selection, extract genotype features with significant distinguishability, and determine the feature identifier set; S1014: If the data distribution in the feature identifier set meets the preset threshold conditions, then it is entered into the pre-established feature database to complete the storage of genotype features. S1015: By performing cluster analysis on the genotype features in the feature database, the genetic differences of individuals from different rootstocks are grouped to obtain preliminary individual classification results; S1016: Based on the individual classification results and combined with the set of differential fragments, analyze the genetic difference patterns between groups and determine whether there are significant classification boundaries. S1017: If the classification boundaries are clear, the classification results are used to generate structured data for genetic difference analysis, thus completing the final classification mapping.

3. The method for matching and screening loofah rootstock and bitter gourd scion according to claim 1, characterized in that, Step S102 specifically includes the following steps: S1021: Using genetic difference data, the initial feature sets of loofah rootstock and bitter gourd scion are obtained from the classification results. The data comparison method is used to extract features and obtain the preliminary distribution of matching degree features. S1022: Based on the preliminary distribution of matching degree features, and considering the genetic differences between loofah rootstock and bitter gourd scion, the support vector machine algorithm is used to classify features, determine the correlation weights between features, and obtain the feature combinations after classification. S1023: Obtain the classified feature combination, and perform quantitative analysis of the matching degree feature in combination with the preset threshold. If the matching degree feature value is higher than the preset threshold, it is determined to be a potentially suitable combination, and the preliminary screening of matching pairs is obtained. S1024: For the initially screened matching pairs, the genetic differences between the loofah rootstock and bitter gourd scion are compared in depth using data comparison methods to determine whether there are hidden matching degree features and obtain the matching pair set after in-depth analysis. S1025: Based on the matching pair set after deep analysis, cluster analysis is used to group the potential fit combinations, determine the matching degree feature distribution of rootstock candidates and scions in each group, and obtain the grouped candidate set; S1026: Based on the grouped candidate sets, for the matching degree feature distribution within each group, if the consistency of the feature distribution of a certain group is higher than the preset threshold, it is determined to be a rootstock candidate with high adaptability, and the final adaptation candidate set is obtained. S1027: Obtain the final candidate set of rootstocks, and sort each group of rootstock candidates by combining genetic differences and matching degree features. Determine the stability of the matching degree features in the sorting results and determine the optimal combination of rootstock candidates.

4. The method for matching and screening loofah rootstock and bitter gourd scion according to claim 1, characterized in that, Step S103 specifically includes the following steps: S1031: By obtaining the genotypic characteristic data of rootstock candidates from the database, preliminary information is organized for each individual to obtain the basic dataset of individual genotypes; S1032: Based on the basic dataset, information comparison technology is used to analyze the distribution of disease resistance genes for each individual and determine the location and distribution information of disease resistance genes in the genome; S1033: If the distribution density of disease-resistant genes in the location distribution information is higher than the preset threshold, then the corresponding individuals are marked as high-potential individuals, and the genome information dataset of high-potential individuals is obtained. S1034: Analyze the association between disease resistance genes and genotype characteristics in genomic information datasets of high-potential individuals to determine the expression potential intensity of disease resistance genes; S1035: Based on the analysis results of expression potential intensity and combined with distribution data, the disease resistance genome of each high-potential individual is comprehensively scored to obtain a comprehensive potential assessment value; S1036: Based on the comprehensive potential assessment value, all high-potential individuals are sorted to determine the final priority list of disease resistance gene expression potential; S1037: The random forest algorithm is used to perform validation analysis on individuals in the priority list. Combined with the association data between genotype characteristics and disease resistance genes, the reliability of the evaluation results is judged, and the final disease resistance genome potential analysis dataset is output.

5. The method for matching and screening loofah rootstock and bitter gourd scion according to claim 1, characterized in that, Step S104 specifically includes the following steps: S1041: By obtaining relevant data on disease resistance genes from the database and combining them with the original records of gene expression, a preliminary assessment of expression potential is compiled. S1042: Based on the preliminary assessment results of expression potential, construct a comprehensive impact factor dataset for specific environments by integrating data on environmental adaptation and adaptation information; S1043: The random forest algorithm is used to process the comprehensive impact factor dataset, calculate the predicted values ​​of disease resistance, and obtain the resistance value of each rootstock individual; S1044: For the resistance values ​​of each individual rootstock, perform data standardization processing to obtain quantitative indicators of individual capabilities; S1045: By using quantitative indicators of individual capabilities and combining data analysis methods, rank all rootstock individuals by capability and determine the ranking results; S1046: If there are rootstock individuals with similar resistance values ​​in the ranking results, the final individual ability ranking is determined by comparing environmental adaptation data under a specific environment. S1047: After obtaining the final individual ability ranking, generate the corresponding disease resistance distribution map to determine the disease resistance location of each rootstock individual in a specific environment.

6. The method for matching and screening loofah rootstock and bitter gourd scion according to claim 1, characterized in that, Step S105 specifically includes the following steps: S1051: Construct an initial dataset by obtaining relevant data from the disease resistance ranking results; S1052: Data cleaning techniques are used to preprocess the sorting results, removing outliers and missing values ​​to obtain a cleaned resistant dataset. S1053: Based on the organized resistance dataset, extract historical data related to plant growth; S1054: Perform feature selection on historical data, filter out feature variables that are highly correlated with growth potential, and determine the feature dataset; S1055: Input data for growth simulation is constructed by combining feature datasets with grafting treatment information of bitter gourd plants; S1056: If the values ​​of some variables in the feature dataset are lower than the preset threshold, they are standardized to obtain normalized simulated input data; S1057: A random forest model is used to train the normalized simulated input data, and the growth potential of bitter gourd plants is simulated and calculated to obtain preliminary growth potential predictions. S1058: Based on the preliminary growth potential predictions and combined with the plant growth trends in historical data, the prediction results are corrected. S1059: If the deviation between the predicted value and the historical trend exceeds the preset range, the predicted value is smoothed to obtain the corrected growth potential value. S10510: A predictive index of plant growth potential is generated using the corrected growth potential value. S10511: Perform multi-dimensional analysis on the prediction indicators to determine their stability under different environmental conditions and obtain the final prediction results.

7. The method for matching and screening loofah rootstock and bitter gourd scion according to claim 1, characterized in that, Step S106 specifically includes the following steps: S1061: By obtaining plant growth potential prediction index data and yield performance data from the database, an initial dataset is formed, and the initial data integration is completed. S1062: Based on the initial dataset, the grafting yield potential of each rootstock individual is quantified using a pre-established comprehensive evaluation model to obtain the potential evaluation results; S1063: Based on the potential assessment results, if the potential value of a certain rootstock individual exceeds the preset threshold, it will be marked as a high-potential individual, and the initial screening will be completed. S1064: By conducting in-depth analysis of the yield performance data of high-potential individuals, obtain their correlation coefficients with growth indicators, and identify key influencing factors; S1065: Based on key influencing factors, a random forest model is used to predict yield performance values, and the predicted yield ranking of each rootstock individual is obtained. S1066: For the ranking of predicted yield, if a certain rootstock individual ranks at the top, its growth analysis data is further extracted to determine its stability characteristics. S1067: By combining stability characteristics with yield potential data, the final priority ranking result is generated, completing the full-process analysis.

8. The method for matching and screening loofah rootstock and bitter gourd scion according to claim 1, characterized in that, Step S107 specifically includes the following steps: S1071: By obtaining yield and quality data of grafted bitter gourd from historical records, an initial dataset is constructed to obtain basic information for subsequent analysis; S1072: Based on the initial dataset, prioritize the output performance values, use a preset threshold to filter out the data subset with high performance values, and determine the key analysis objects; S1073: By standardizing the quality data in the high-performance data subset, the distribution characteristics of the quality data are obtained, and its initial fluctuation range is determined. S1074: If the initial fluctuation range exceeds the preset threshold, further data comparison and analysis of the quality data will be performed, and the support vector machine algorithm will be used to accurately predict the fluctuation range to obtain the predicted fluctuation value. S1075: Based on the predicted fluctuation value and combined with the historical quality stability record of grafted bitter gourd, obtain the intermediate index for stability assessment and determine the degree of influence of the fluctuation range on stability. S1076: By comparing intermediate indicators with preset stability evaluation standards, it is determined whether the quality stability of grafted bitter gourd meets expectations, and the final evaluation result is obtained.

9. The method for matching and screening loofah rootstock and bitter gourd scion according to claim 1, characterized in that, Step S108 specifically includes the following steps: S1081: Obtain quality stability data and screening accuracy data of rootstock candidate individuals, and integrate relevant indicators from multiple sources through a pre-established data collection system to form an initial dataset; S1082: For the initial dataset, information integration technology is used to standardize the quality stability data and screening accuracy data to obtain a unified evaluation data matrix; S1083: Based on the evaluation data matrix, each candidate rootstock is comprehensively scored using a preset scoring system. If the score of a candidate is lower than the preset threshold, it is removed from the candidate list to obtain preliminary screening results. S1084: By analyzing the individuals in the preliminary screening results and combining data fusion technology, multiple indicators of the remaining rootstock candidates are weighted and calculated to determine the weighted comprehensive score. S1085: Based on the weighted comprehensive score, sort the remaining rootstock candidates. If the comprehensive score of a candidate is higher than the preset upper limit threshold, mark it as a priority candidate and obtain the sorted candidate list. S1086: For the sorted candidate list, apply the decision tree algorithm to perform final verification on the priority objects, determine whether they meet the requirements of the preferred list, and output the set of the verified preferred individuals. S1087: By comparing the data of the verified preferred individual set, information integration technology is used to generate the final list of preferred rootstocks, and the final result that meets the requirements of quality stability and screening accuracy is determined.

10. The method for matching and screening loofah rootstock and bitter gourd scion according to claim 1, characterized in that, Step S109 specifically includes the following steps: S1091: Obtain relevant information about the rootstock list from the screening process using data acquisition tools, organize it in a structured format, and store it in a pre-established database to obtain the initial rootstock screening dataset. S1092: Based on the initial rootstock screening dataset, classify the screening process information recorded therein, extract key indicators using preset field rules, and determine the core basis data in the screening process. S1093: If the core data meets the preset threshold conditions, the latest experimental data will be obtained from the grafting experiment through an automatic update mechanism and added to the database to determine whether to adjust the existing rootstock list. S1094: If the indicators of some rootstocks change after the experimental data are supplemented, the random forest algorithm is used to predict and analyze the trend of change to obtain the potential optimization direction of the rootstock list. S1095: Based on the optimization direction obtained from the predictive analysis, compare and verify the subsequent experimental data stored in the database, obtain experimental results consistent with the optimization direction, and determine the adjustment plan for the rootstock list. S1096: Update the rootstock list in the database by adjusting the scheme, and in conjunction with the goal of continuous improvement, associate the updated list with the screening criteria to obtain the latest screening reference data; S1097: For the latest screening reference data, use an automated data recording method to save it to the database, forming a closed-loop data supplementation and optimization process to judge the continuous improvement effect of the rootstock screening criteria.

Images