Cross-platform global weed gene database unified building method

Abstract

The invention relates to the technical field of information, and provides a cross-platform global weed gene database unified building method which comprises the following steps: screening out a weed population having gene exchange characteristics with a target crop, identifying gene segments of gene exchange between the crop and weeds by adopting a gene comparison method, and determining a transmission path of resistance genes in interspecific hybridization; according to a gene comparison result, extracting transmission path data of the resistance gene, and analyzing the influence of resistance gene transmission on the genetic structure of the weed population in combination with the genetic structure characteristics of the weed population; calculating the frequency distribution of the resistance genes in different weed populations, judging the diffusion trend of the resistance genes in the weed populations according to the frequency distribution, generating a resistance gene diffusion map, and predicting the variation trend of the future weed population genetic structure based on a weed population genetic structure variation trend chart and the resistance gene diffusion map. And a global weed gene database is established, so that an effective weed management strategy can be formulated.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and in particular to a cross-platform global weed gene database construction method. Background Art

[0002] Related weeds refer to wild relatives of modern crops and weeds that are closely related to crops, including transitional types between cultivated types and wild types. In the study of the global weed gene database, it is necessary to integrate weed genomes, functional analysis, and related bioinformatics data to facilitate weed genome research, functional prediction, and data analysis. In the process of integrating weed genomes, there is a phenomenon of gene introgression from closely related weeds. At this time, it is necessary to focus on the characteristics of gene exchange between crops and weeds. Due to the complex evolutionary relationship between crops and weeds, gene exchange between them may occur at multiple levels, including interspecific hybridization, introgression hybridization, and transgenic drift. In these processes, resistance genes may be transmitted from crops to weed populations, causing weeds to become resistant to herbicides. However, the transmission pathways of resistance genes between different weed populations are still unclear. In addition, the transmission of resistance genes may have a significant impact on the genetic structure of weed populations, resulting in a decrease in genetic diversity and evolutionary potential of weed populations. Therefore, in-depth analysis of the mechanism by which resistance gene transmission reshapes the genetic structure of weed populations is of great significance for understanding the evolutionary dynamics of weeds and formulating effective weed management strategies. This requires the comprehensive use of multidisciplinary research methods such as genomics, population genetics, and evolutionary biology to clarify the characteristics of gene exchange between crops and weeds, reveal the transmission and spread mechanism of resistance genes in weed populations, and the impact of these processes on the genetic structure of weed populations, providing a scientific basis for the sustainable management of weeds worldwide. Summary of the invention

[0003] The present invention provides a cross-platform global weed gene database construction method, which mainly includes: Weed populations with gene exchange characteristics with target crops are screened out, gene sequence data of closely related weed species are extracted from the global weed gene database, and a gene introgression analysis model is constructed; gene comparison methods are used to identify gene fragments of gene exchange between crops and weeds in the screened weed populations, and the transmission path of resistance genes in interspecific hybridization is determined; based on the gene comparison results, the transmission path data of resistance genes are extracted, and a genetic structure change model is constructed in combination with the genetic structure characteristics of the weed population to analyze the impact of resistance gene transmission on the genetic structure of the weed population; population genetics algorithms are used to calculate the frequency distribution of resistance genes in different weed populations, and the diffusion trend of resistance genes in the weed population is determined based on the frequency distribution, generating a resistance gene diffusion map; based on the resistance gene diffusion map and combined with the genetic structure change model of the weed population, the impact of resistance gene transmission on the genetic indicators of the remodeling of the genetic structure of the weed population is analyzed, including changes in allele frequency and linkage disequilibrium. changes, and generate a genetic structure remodeling map reflecting the changes in genetic indicators; through the genetic structure remodeling map, extract the key nodes of the genetic structure changes of the weed population and the corresponding genetic indicators, combined with the gene introgression analysis model, analyze the impact of gene introgression on the adaptive evolutionary potential of the weed population and the level of genetic diversity of the population; according to the impact of gene introgression on the genetic structure of the weed population, generate the changing trends of the genetic diversity index and the allele frequency of the adaptive site over time, establish the trend map of the genetic structure changes of the weed population, and combine with the resistance gene diffusion map to quantitatively analyze the correlation between the transmission of resistance genes and the changes in the genetic structure of the weed population; if the correlation is greater than or equal to the preset correlation threshold, then based on the weed population genetic structure change trend map and the resistance gene diffusion map, construct a weed population genetic structure prediction model to predict the future trend of the genetic structure of the weed population, and build a global weed gene database based on the future trend of the genetic structure of the weed population.

[0004] The technical solution provided by the embodiment of the present invention may have the following beneficial effects: The present invention discloses a cross-platform global weed gene database construction method. The method first screens weed populations that have gene exchanges with target crops from the global weed gene database and extracts their gene sequence data. By constructing a gene introgression analysis model and a gene comparison algorithm, the fragments of gene exchange between crops and weeds are identified, and the resistance gene transmission path is determined. Combined with the population genetics algorithm, the impact of resistance gene transmission on the genetic structure of the weed population is analyzed, and a resistance gene diffusion map and a genetic structure remodeling map are generated. Furthermore, the present invention analyzes the impact of gene introgression on the adaptive evolutionary potential and genetic diversity of the weed population from the perspective of evolutionary genetics, and constructs a trend map of the genetic structure change of the weed population. Finally, a machine learning algorithm is used to predict the changing trend of the genetic structure of the weed population in the future. The present invention realizes a comprehensive analysis and prediction of the changes in the genetic structure of the weed population, and provides a scientific basis for formulating effective weed management strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0005] Figure 1 The present invention is a flow chart of a method for building a cross-platform global weed gene database.

[0006] Figure 2 The figure is a schematic diagram of a cross-platform global weed gene database construction method of the present invention.

[0007] Figure 3 This is another schematic diagram of a cross-platform global weed gene database construction method of the present invention. DETAILED DESCRIPTION

[0008] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0009] like Figure 1-Figure 3 In this embodiment, a cross-platform global weed gene database construction method may specifically include: Step S101, screen out weed populations that have gene exchange characteristics with target crops, extract gene sequence data of closely related weed species from the global weed gene database, and construct a gene introgression analysis model.

[0010] The method comprises the following steps: obtaining genome data of a weed population in an area where a target crop is located, constructing a multidimensional feature vector including genome similarity and genetic distance according to the genome data; extracting genome colinearity regions according to the multidimensional feature vector, and sorting the genome data by a pedigree clustering algorithm to obtain a genome arrangement map; calculating a genetic distance matrix according to the genome arrangement map, determining the direction of gene flow through the genome colinearity regions of the genetic distance matrix to obtain a gene exchange frequency; constructing a genome recombination index scoring matrix according to the gene exchange frequency, and calculating a hybridization affinity coefficient through the scoring matrix to obtain a gene introgression risk level; and determining that there is a gene introgression risk between the target crop and the weed population if the gene introgression risk level exceeds a preset threshold.

[0011] Specifically, for the genome data of weed populations collected in the geographical area where the target crops are located, a multidimensional feature vector including genome similarity, genetic distance, and genome coverage is constructed, and homologous gene sequences are obtained from the global genome database. The genome data of closely related species whose genome coverage is greater than the preset threshold and whose sequence integrity meets the requirements are obtained. The genome colinearity region is extracted from the above homologous gene sequence library, and the sequence repetition and base substitution rate are counted to obtain the genome data arrangement map. The pedigree clustering algorithm is used to sort the genome data from high to low according to the value of the population genetic variation. The genetic distance matrix is ​​calculated using the sorted genome data, the direction of gene flow is determined by the base substitution frequency of the genome colinearity region, the gene exchange frequency is quantitatively calculated, and a numerical mapping relationship is constructed for the degree of homology between the genome data. Based on the above genetic distance matrix and gene exchange frequency data, a genome recombination index scoring matrix is ​​constructed, and a machine learning method is used to predict and model the genome recombination index. The hybridization affinity coefficient is calculated according to the genome recombination index scoring matrix, and the gene introgression risk level is determined by the numerical value of the affinity coefficient, and the gene introgression probability distribution map between the target crop and the weed population is obtained. According to the gene introgression probability distribution map data, a gene introgression risk assessment function was established, and the risk assessment function was trained using a regression algorithm to obtain a gene introgression risk prediction model. Gene introgression analysis is of great significance in agricultural ecosystems. The genetic characteristics of weed populations are characterized by constructing a multidimensional feature vector, which contains key indicators such as genome similarity, genetic distance, and genome coverage. For example, in the study of gene exchange between rice and wild rice, sequence data with a genome coverage of more than 95% were identified as high-quality data, which showed high sequence integrity. After collinearity analysis, the homologous gene sequences screened from the global genome database showed obvious collinear conservative blocks at the chromosome level. By analyzing the sequence repetition and base substitution rate in these conservative blocks, the genetic variation patterns between different populations can be found. For example, in the study of corn and its wild relatives, genome collinearity analysis showed that there were obvious sequence conservative regions on chromosomes 2, 4, and 7, and the base substitution rate in these regions was less than 0.5%. The calculation of the genetic distance matrix is ​​based on the sequence alignment results of the genome collinear region, and the directionality of gene flow is revealed by analyzing the base substitution frequency. In the hybridization study between wheat and Aegilops tauschii, it was found that the frequency of gene exchange reached a peak during the flowering period, among which pollen-mediated gene flow dominated, and the gene exchange frequency could reach 12%. The genome recombination index scoring matrix reflects the potential for gene exchange between different species, and this index is significantly positively correlated with hybrid affinity. For example, in the study of rapeseed and its wild relatives, populations with a recombination index of more than 0.8 showed higher hybrid affinity, and their hybrid fruiting rate could reach more than 25%.The introgression risk assessment function integrates multiple ecological factors, including population spatial distribution, overlap of flowering phenology, etc. In the symbiotic area of ​​soybean and wild soybean, analysis shows that when the overlap of the flowering period of the two species exceeds 60%, the risk of introgression increases significantly. The accuracy of the risk prediction model can reach 85%, providing an important reference for the ecological security assessment of crop planting areas. Among them, the introgression risk level is divided into low risk, with a risk index less than 0.3, moderate risk, with a risk index between 0.3 and 0.7, and high risk, with a risk index greater than 0.7. By constructing a probability distribution map of introgression, the distribution pattern of introgression risks in different geographical regions is intuitively displayed, providing a scientific basis for the formulation of corresponding risk prevention and control measures.

[0012] Step S102, using a gene comparison method to identify gene fragments of gene exchange between crops and weeds for the screened weed population, and determining the transmission path of resistance genes in interspecific hybridization.

[0013] A double sequence local alignment method is used to perform sequence alignment on crop and weed genome data, and a gene fragment sequence similarity score is calculated based on the alignment results to obtain a target gene fragment with a sequence similarity score higher than a preset threshold; a coding region sequence is extracted for the target gene fragment, and a gene expression amount is calculated based on the transcriptome sequencing data of the coding region sequence, and a genetic stability index is calculated by calculating the gene expression amount and the gene copy number; a time series expression profile is constructed based on the genetic stability index and RNA sequencing data, and a gradient descent method is used to fit the time series expression profile to obtain gene transfer path data; a gene interaction network is constructed for the gene transfer path data, and a gene transfer key node sequence is obtained through a graph structure algorithm, and the resistance gene action site is determined based on the comparison result of the key node sequence with a phenotypic database, so as to form a transfer path diagram of the resistance gene during interspecific hybridization.

[0014] Specifically, sequence alignment is performed based on crop and weed genome data, and the double sequence local alignment method is used to divide the gene fragment length interval. The sequence similarity score is calculated for the base composition within the gene fragment, and the target gene fragment with a sequence similarity score higher than the preset threshold is screened from the whole genome. The coding region sequence in the target gene fragment is extracted, and the gene expression is calculated using transcriptome sequencing data. The genetic stability index is calculated based on the gene copy number and gene mutation frequency. The location of the gene recombination breakpoint is determined by comparing the sequencing depth data. The transcriptome expression data of the target gene fragment is obtained using RNA sequencing data, and a time series expression profile is constructed for the transcriptome expression data. The gradient descent method is used to fit the gene expression regulation data to obtain the preliminary direction of the gene transfer path. A gene interaction network is constructed based on the preliminary direction data of gene transfer, and a graph structure algorithm is used to perform connectivity analysis on the gene interaction network to obtain the key node sequence of gene transfer. Resistance phenotypic association analysis is performed on the key node sequence of gene transfer, and the action site of the resistance gene is determined by comparing the phenotypic database, and a transmission pathway map of the resistance gene in the process of interspecific hybridization is constructed. The neural network model was trained using the data of the transmission pathway map to predict the transmission probability of resistance genes in different hybrid combinations, and the dominant path combination of resistance gene transmission was obtained. In the study of gene exchange between crops and weeds, sequence alignment is a basic step in identifying gene transmission. Taking rice and wild rice as an example, the genome sequence was divided into fragments with a length of 300 to 1000 bases by local sequence alignment method, and the sequence similarity score of each fragment was calculated. When the score exceeded 0.85, it was identified as a high-similarity region. These high-similarity regions often contain potential gene exchange sites. Among the identified target gene fragments, the coding region sequence is particularly important. Through transcriptome sequencing analysis of the target gene fragments, it was found that the expression of disease resistance genes in weeds often showed time dependence. For example, in the study of wheat and Aegilops tauschii, the expression of rust resistance genes reached a peak 4 hours after infection with pathogens, which was 8 times the basal level, and formed a clear expression hotspot region on the chromosome. RNA sequencing data reveals the dynamic changes in gene expression. In the study of corn and its wild relatives, the transcriptome expression profile of herbicide resistance genes showed significant circadian rhythmicity, reaching a peak expression when the light intensity was the highest, and the expression level could reach 12 times the basal level. This expression pattern is closely related to the expression regulation network of photosensitive pigment genes. The construction of gene interaction networks is based on expression correlation analysis. In the study of rapeseed and its wild relatives, by analyzing the expression correlation of 250 gene loci, a gene network containing 1,500 interactions was constructed, in which 35 key node genes were identified, which play a pivotal role in the transmission of resistance. Phenotypic association analysis of resistance genes showed that different resistance genes have specific sites of action.In the study of soybean and wild soybean, drought resistance-related genes are mainly distributed on chromosomes 3, 7, and 13, and these sites are closely related to root morphological development. By analyzing the phenotypic data of 80 hybrid combinations, it was found that there was an obvious directional preference for the transmission of drought resistance genes, among which the drought resistance alleles from the wild genome showed significant adaptive advantages under drought stress. In the constructed transmission pathway map, each node represents a gene locus, and the connection weights between nodes reflect the probability of gene transmission. Through the analysis and training of 1,000 sets of hybridization data, the accuracy of the prediction model reached 90%, and multiple advantageous transmission pathways were successfully predicted. Among them, in the genome after three generations of hybridization, the transmission frequency of resistance genes can reach 25%, and it shows stable inheritance in the offspring.

[0015] Step S103, extracting the transmission path data of the resistance gene according to the gene comparison result, building a genetic structure change model in combination with the genetic structure characteristics of the weed population, and analyzing the impact of the resistance gene transmission on the genetic structure of the weed population.

[0016] A gene transmission network map is constructed based on genome resequencing data, and the gene transmission network map contains transmission association information between gene sites; genome variation site data is obtained from the gene transmission network map, and a genetic structure characteristic matrix is ​​constructed using a population genetics algorithm, and the genetic structure characteristic matrix records population genotype distribution information; a population genetic structure dynamic change curve is constructed using the genetic structure characteristic matrix, and the population genetic structure dynamic change curve reflects the genetic variation accumulation rate in a time series; if the genetic variation accumulation rate exceeds a preset threshold, the resistance gene transmission probability distribution is calculated, and a genome recombination frequency heat map is constructed using the resistance gene transmission probability distribution, and the genome recombination frequency heat map is used to display a quantitative description of the effect of resistance gene transmission on population genetic differentiation.

[0017] Specifically, the position information of the resistance gene sequence is extracted according to the gene comparison results, the population genotype frequency distribution is calculated using molecular marker data, and the gene transmission network map is constructed through genome resequencing data to obtain the spatial distribution characteristics of the gene transmission pathway. For the gene transmission network map data, the population genetics algorithm is used to calculate the genetic diversity parameters, and the population genetic structure characteristic matrix is ​​constructed according to the genomic variation site data to obtain the genetic variation map of the weed population. Using the genetic variation map data, the dynamic change curve of the population genetic structure on the time series is constructed, the gene penetration rate is calculated by the genomic variation accumulation rate, and the population genetic structure evolution path is predicted by the deep learning method. Based on the population genetic structure evolution path data, the transmission probability distribution of resistance genes in the population is calculated, and the genome recombination frequency heat map is constructed to obtain a quantitative description of the genetic differentiation of the population caused by the transmission of resistance genes. According to the quantitative data of genetic differentiation, the support vector machine algorithm is used to construct a population fitness prediction model, and the amplitude of genetic structure changes caused by the transmission of resistance genes is calculated from the genome level. For the genetic structure change amplitude data, a gene penetration intensity scoring function is constructed, and the prediction results of the genetic structure change model are verified by the genomic variation accumulation curve. When analyzing the impact of resistance gene transmission on the genetic structure of weed populations, genome resequencing data provides comprehensive genetic variation information. Taking the study of rice and wild rice as an example, molecular marker data analysis found that the frequency distribution of disease resistance genes in weed populations showed obvious geographical gradient characteristics. The frequency of disease resistance genes in southern populations reached 0.35, while that in northern populations was only 0.12. This difference reflects the change in selection pressure under different geographical environments. The calculation of genetic diversity parameters reveals the dynamic changes in population genetic structure. In the study of wheat and Aegilops tauschii, genomic variation site data showed that the introduction of drought resistance genes increased the average heterozygosity of the population from 0.25 to 0.42, indicating that the transmission of resistance genes significantly changed the genetic structure of the population. Multiple variation hotspots appeared in the population genetic variation map, which were highly overlapped with the regulatory sequences of resistance genes. Time series analysis showed the evolutionary trajectory of population genetic structure. In the study of maize and its wild relatives, the penetration rate of insect resistance genes in the population showed nonlinear characteristics. The penetration rate was the fastest in the first three generations after introduction, reaching a frequency increase of 0.15 per generation, and then gradually stabilized. The population evolution path predicted by deep learning methods showed that the transmission of resistance genes led to the directional evolution of population genetic structure in the direction of increasing resistance. The heat map of genome recombination frequency reflects the spatial characteristics of resistance gene transmission. In the genome analysis of rapeseed and its wild relatives, cold resistance genes formed obvious recombination hotspots on chromosomes, and the recombination frequency of these regions was more than 5 times the background value. The population fitness prediction model showed that individuals carrying cold resistance genes had a 35% higher fitness in cold environments. Quantitative analysis of genetic structure changes showed that the transmission of resistance genes is often accompanied by significant population differentiation.In the study of soybeans and wild soybeans, the genetic differentiation index of the population increased from 0.18 to 0.32 after the introduction of herbicide-resistant genes, indicating that the transmission of resistance genes accelerated the genetic differentiation process of the population. The gene penetration intensity score reflects the degree of this change, and the population with a score of more than 0.8 shows obvious adaptive differentiation characteristics. The prediction results of the genetic structure change model were verified by the genomic variation accumulation curve. In the follow-up observation of 10 generations, the correlation coefficient between the variation accumulation curve predicted by the model and the actual observed value reached 0.92, confirming that the model has a high reliability in predicting the changes in the genetic structure of the population. This trend of change shows that the transmission of resistance genes not only changes the target traits, but also reshapes the genetic structure of the entire population.

[0018] Step S104, using a population genetics algorithm to calculate the frequency distribution of resistance genes in different weed populations, judging the diffusion trend of resistance genes in the weed populations based on the frequency distribution, and generating a resistance gene diffusion map.

[0019] The frequency distribution value of the resistance gene is calculated according to the population genome sequencing data to obtain the distribution density data of the resistance gene in the population; the spatial coordinate information is obtained by using the geographic information system, and the population geographical distribution matrix is ​​constructed through the resistance gene distribution density data; the gene flow flux between adjacent geographical regions is calculated for the population geographical distribution matrix, and the gene diffusion rate curve is obtained after predicting the population diffusion path through the deep learning method; the gene diffusion radius is calculated according to the gene diffusion rate curve using the spatial statistics method, and the diffusion trajectory map of the resistance gene in the geographical space is drawn through the gene diffusion radius, and the resistance gene diffusion map is generated based on the diffusion trajectory map.

[0020] Specifically, based on the genome sequencing data of weed populations, the frequency distribution value of resistance genes is calculated using population genetics calculation methods, gene linkage information is extracted through genome resequencing data, and the distribution density map of resistance genes in the population is constructed. Based on the distribution density data of resistance genes, the spatial coordinate information collected by the geographic information system is integrated to construct a population geographic distribution matrix, and the diffusion direction of the population in geographic space is calculated. For the population geographic distribution matrix data, the gene flow flux between adjacent geographical regions is calculated, and the population diffusion path is predicted using deep learning methods to obtain the gene diffusion rate curve. Using the gene diffusion rate curve data, a population genetic drift intensity index system is constructed to calculate the population density distribution in different geographical regions. Based on the population density distribution data, the spatial statistical method is used to calculate the gene diffusion radius and draw the diffusion trajectory of resistance genes in geographic space. For the gene diffusion trajectory data, the random forest algorithm is used to predict the population diffusion trend and generate a resistance gene diffusion map containing time and space dimensions. In the population genetics study of rice and wild rice, the distribution of resistance genes in weed populations presents complex spatial characteristics. Genome resequencing data showed that the frequency distribution value of drought resistance genes in weed populations in the Yangtze River Basin reached 0.45, while it was only 0.15 in the northern region. This difference in distribution density reflects the changes in selection pressures in different geographical environments. Gene linkage analysis found that there was significant linkage disequilibrium between drought resistance genes and flowering period genes, with a linkage value of 0.82. Studies on the geographical distribution of populations reveal the spatial laws of gene diffusion. In the case of wheat and Aegilops tauschii, geographic information system data showed a trend of cold resistance genes spreading from high altitudes to low altitudes, and the diffusion direction was highly correlated with the altitude gradient. The gene flow flux between adjacent altitude areas decreased with the increase of altitude difference. When the altitude difference exceeded 500 meters, the gene flow flux dropped to 25% of the initial value. Population diffusion path analysis showed the dynamic process of gene transmission. In the study of corn and its wild relatives, the gene diffusion rate showed periodic changes during the breeding season, with a peak at the flowering period, reaching a weekly diffusion radius of 8 kilometers. Deep learning predictions show that this periodic diffusion pattern is closely related to local climatic conditions. For every 1 degree increase in temperature, the diffusion radius increases by about 0.5 kilometers. The genetic drift intensity index system reflects the changes in the genetic structure of the population. In the study of rapeseed and its wild relatives, the genetic drift intensity of the marginal population was 2.5 times that of the core population, indicating that the population size has a significant impact on the fluctuation of gene frequency. The population density distribution in different geographical regions shows significant differences, and the population density in the core area can reach 5 times that of the marginal area. The spatial statistical analysis of the gene diffusion trajectory reveals the geographical pattern of diffusion. In the case of soybean and wild soybean research, the diffusion radius of insect-resistant genes increases with the replacement of generations. After 10 generations, the diffusion range covers an area three times the original distribution area.The diffusion trajectory showed a clear directional preference. The diffusion speed along the river was twice that in the vertical direction. The diffusion map of resistance genes in the spatiotemporal dimension integrated multi-dimensional information. The data of 1,000 sampling points were analyzed by the random forest algorithm, and the prediction accuracy reached 88%. The prediction results showed that in the next five generations, the frequency of resistance genes in suitable habitats will tend to be stable, with a stable value between 0.6 and 0.8. This prediction provides an important basis for formulating scientific prevention and control measures.

[0021] Step S105, based on the resistance gene diffusion map and the genetic structure change model of the weed population, analyze the impact of resistance gene transmission on the genetic indicators of the genetic structure remodeling of the weed population, including changes in allele frequency and linkage disequilibrium, and generate a genetic structure remodeling map reflecting the changes in genetic indicators.

[0022] A genotyping method is used to obtain population genotype data, and allele frequency distribution values ​​at the genome level are calculated based on the population genotype data; the linkage disequilibrium coefficient of the chromosome fragment is calculated based on the allele frequency distribution value, and a linkage block distribution map within the genome is constructed through the linkage disequilibrium coefficient; the spatial distribution of gene recombination breakpoints is statistically analyzed for the linkage block distribution map, and a population differentiation index matrix is ​​constructed based on the spatial distribution; the population differentiation index matrix is ​​processed using a kernel density estimation method to obtain a genetic diversity distribution function, and a population genetic differentiation index value is calculated based on the genetic diversity distribution function; combined with the population genetic differentiation index value, a support vector machine algorithm is used to predict the evolution direction of the genetic structure, and a genetic structure remodeling map containing allele frequencies and linkage disequilibrium is generated.

[0023] Specifically, based on the resistance gene diffusion map data, the genotyping method is used to obtain the population genotype data, calculate the allele frequency distribution value at the genome level, and construct a linkage disequilibrium heat map reflecting the genotype distribution. Using the linkage disequilibrium heat map data, the linkage disequilibrium coefficient of the chromosome fragment is calculated, the linkage block distribution map within the genome is constructed, and the location information of the genetic structure variation region is obtained. For the genetic structure variation region, the spatial distribution of gene recombination breakpoints is statistically analyzed, the population differentiation index matrix is ​​constructed, and the deep learning method is used to predict the dynamic process of genome reconstruction. Based on the dynamic data of genome reconstruction, the change curve of the population heterozygosity index over time is calculated, the selection coefficient distribution map is constructed, and the time series change data reflecting the gene penetration process are obtained. According to the gene penetration time series data, the kernel density estimation method is used to construct the genetic diversity distribution function and calculate the population genetic differentiation index value. The population genetic differentiation data is subjected to multi-dimensional dimensionality reduction processing, and the support vector machine algorithm is used to predict the direction of genetic structure evolution, and a genetic structure remodeling map containing allele frequency and linkage disequilibrium is generated. In the study of the remodeling of the genetic structure of weed populations by resistance gene transmission, the dynamic changes of allele frequency can be tracked through genotyping data. Taking the study of rice and wild rice as an example, the allele frequency of the herbicide resistance gene showed a rapid upward trend in the early stage of transmission, rising from 0.05 to 0.35 within three generations. The linkage disequilibrium heat map showed that the gene formed a stable linkage block with the four adjacent functional genes. The analysis of linkage disequilibrium blocks revealed the spatial characteristics of genome reconstruction. In the study of wheat and Aegilops tauschii, the linkage disequilibrium coefficient of chromosome fragments decreased with the increase of physical distance from the resistance gene. The linkage disequilibrium coefficient within 500kb remained above 0.8, forming a stable co-segregation unit. This linkage structure promoted the coordinated inheritance of the resistance gene and its surrounding favorable alleles. The distribution of recombination breakpoints in the region of genetic structure variation reflects the law of genome recombination. In the case of maize and its wild relatives, the genome recombination breakpoints showed obvious hotspot distribution on the chromosome, of which 35% of the recombination events were concentrated in the intergenic region. The population differentiation index showed that individuals carrying the resistance gene formed obvious genetic differentiation within five generations. The time series analysis of population heterozygosity showed the dynamic changes of genetic structure. In the study of rapeseed and its wild relatives, the average heterozygosity of the population showed a trend of first increasing and then decreasing after the transmission of the resistance gene, with the peak occurring in the fourth generation, reaching 0.65, and then gradually decreasing to 0.45 under selection pressure and stabilizing. The distribution map of the selection coefficient showed that individuals carrying the resistance gene had a 1.8-fold fitness advantage under the herbicide treatment environment, and the genetic diversity distribution function reflected the spatial pattern of genetic variation in the population.In the study of soybean and wild soybean, the kernel density estimation method showed that genetic diversity showed a significant gradient distribution in geographic space. The genetic diversity of the marginal population was 25% lower than that of the core population, and the population genetic differentiation index reached 0.32. This distribution pattern was closely related to the spatiotemporal dynamics of gene penetration. The genetic structure remodeling map integrated multidimensional genetic indicators. Through multidimensional data analysis of 1,000 sample points, the support vector machine algorithm predicted that the transmission of resistance genes led to the directional evolution of the population genetic structure. In terms of linkage disequilibrium, the linkage disequilibrium intensity within 250kb around the resistance gene increased by 2.5 times, forming a new adaptive evolution unit. Changes in the allele frequency spectrum showed that the frequency of allele combinations related to resistance in the population increased significantly, thereby reshaping the genetic structure of the population.

[0024] Step S106, extracting the key nodes of genetic structure changes in weed populations and corresponding genetic indicators through the genetic structure remodeling map, and combining the gene introgression analysis model to analyze the impact of gene introgression on the changes in the adaptive evolutionary potential of the weed population and the changes in the genetic diversity level of the population.

[0025] According to the genetic structure remodeling map data, the key sites of genome variation are extracted, the allele frequency change values ​​are calculated according to the key sites of genome variation, and the evolution rate distribution map is constructed through the allele frequency change values; the genome sequence variation accumulation curve is calculated according to the evolution rate distribution map, the genome key adaptive sites are identified by the variation accumulation curve, and the population fitness change data are obtained through the key adaptive sites; a genome selection pressure gradient map is constructed for the population fitness change data, the introgression pressure coefficient is calculated according to the selection pressure gradient map, the population fitness curve is constructed, and the population genetic diversity index is calculated; the adaptive evolution trend is predicted according to the population genetic diversity index, and the genome evolution potential is evaluated according to the predicted result of the adaptive evolution trend to obtain the evaluation result of the genome evolution potential.

[0026] Specifically, according to the genetic structure remodeling map data, the key sites of genomic variation are extracted, the allele frequency change values ​​are calculated for the key sites, and the evolution rate distribution map is constructed through the population genetic data. Based on the evolution rate distribution map, the variation accumulation curve of the genome sequence is calculated, and the key adaptive sites of the genome are identified by the deep learning method to obtain the data of population fitness change. According to the population fitness change data, the genomic selection pressure gradient map is constructed, and the introgression pressure coefficient is calculated according to the distribution of gene introgression sites to obtain the population genetic differentiation degree value. Using the population genetic differentiation data, the population fitness curve is constructed by the kernel density estimation method, and the population genetic diversity index is calculated. According to the population genetic diversity index, the diversity change time series map is drawn, and the adaptive evolution trend is predicted by the time series analysis method. According to the adaptive evolution trend data, the genome evolution potential evaluation function is constructed, and the support vector machine algorithm is used to predict the evolution direction of the population genetic structure. For the evolution direction prediction results, the genome reconstruction probability distribution is calculated to generate an evolutionary genetic feature map containing the adaptive evolution potential and the dynamic changes of population diversity. From the perspective of evolutionary genetics, the identification of key sites of genomic variation is crucial to analyze the genetic structure changes of weed populations. In the study of rice and wild rice, the analysis of genomic data revealed that the allele frequencies of the herbicide resistance gene loci and the surrounding 500kb regions changed significantly, with a change of more than 0.4. These key loci constituted hot spots on the evolution rate distribution map. The accumulation of genomic sequence variation reflects the adaptive changes of the population. In the case of wheat and Aegilops tauschii, the cumulative variation rate of the genomic sequence of the population after the introgression of disease resistance genes showed a nonlinear growth in 10 consecutive generations. The variation rate of the adaptive loci related to resistance was 3 times that of the neutral loci, indicating that these loci were under strong selection pressure. The distribution pattern of selection pressure reveals the scope of influence of gene introgression. In the study of maize and its wild relatives, the genomic selection pressure gradient map showed that the introgression pressure coefficient showed obvious spatial heterogeneity on the chromosome, of which 35% of the genomic regions showed significant selection signals, and the population genetic differentiation degree value reached 0.42. The population adaptability curve reflects the population's ability to respond to the environment. In the analysis of rapeseed and its wild relatives, the kernel density estimation method showed that the population fitness showed a bimodal distribution after the introduction of cold-resistant genes, indicating that two different adaptation strategies appeared in the population. The population genetic diversity index decreased from the original 0.65 to 0.48, indicating that gene introgression led to the homogenization of the population genetic background. The temporal change analysis predicted the evolutionary direction of the population. In the study of soybean and wild soybean, the diversity change time series map showed that the population experienced a typical selection and elimination process, and the adaptive evolution showed stage characteristics. In the stage with the strongest selection pressure, the genetic diversity of the population decreased by 0.05 per generation until the eighth generation when it stabilized. The evaluation of the evolutionary potential of the genome revealed the long-term evolutionary trend of the population.The data of 1,000 sample points were analyzed by the support vector machine algorithm, and the prediction accuracy reached 92%. The prediction results showed that the population carrying resistance genes had greater evolutionary potential, and the probability of genome reconstruction increased to 0.35 in the short term. This change increased the chance of the population to produce new adaptive traits. The dynamic evolution map of the population genetic structure integrated multidimensional evolutionary information. In terms of adaptive evolutionary potential, gene introgression led to a significant increase in the selection coefficient, with the average value increasing from 0.1 to 0.28. The dynamic changes in population diversity showed that although the diversity level decreased in the early stage of introgression, new mutation combinations continued to emerge as the genome recombination proceeded, maintaining the population's evolutionary potential.

[0027] Step S107, based on the impact of gene introgression on the genetic structure of the weed population, generate the changing trends of the genetic diversity index and the allele frequency of the adaptive loci over time, establish a trend graph of the genetic structure of the weed population, and analyze the correlation between the transmission of resistance genes and the changes in the genetic structure of the weed population in combination with the resistance gene diffusion map.

[0028] The allele frequencies of adaptive sites in genome resequencing data are obtained, and a population genetic variation distribution map is constructed according to the allele frequencies; a genetic distance dynamic curve is constructed according to the population genetic variation distribution map, and a deep learning method is used to obtain the time evolution trend of the degree of resistance gene introgression from the genetic distance dynamic curve; the genetic structure variation coefficient is calculated based on the time evolution trend of the degree of resistance gene introgression, and a genetic diversity dynamic distribution map is constructed from the genetic structure variation coefficient using a kernel density estimation method; a gene diffusion intensity matrix is ​​constructed according to the genetic diversity dynamic distribution map, and the time correlation coefficient between gene diffusion and genetic structure change is calculated.

[0029] Specifically, based on the genome data of weed populations, the population genetics method was used to calculate the genetic diversity index values ​​at different time points, the allele frequencies of adaptive sites were obtained through genome resequencing data, and a population genetic variation distribution map at the genome level was constructed. Based on the population genetic variation distribution data, a dynamic curve of genetic distance in the time dimension was constructed, and a deep learning method was used to predict the temporal evolution trend of the degree of resistance gene introgression. Using the resistance gene introgression trend data, the coefficient of genetic structure variation between generations was calculated, and the kernel density estimation method was used to construct a dynamic distribution map of genetic diversity over time. Based on the dynamic distribution data of genetic diversity, a resistance gene diffusion intensity matrix was constructed, and the temporal correlation coefficient between gene diffusion and genetic structure change was calculated. Based on the temporal correlation coefficient data, the time series decomposition method was used to extract the periodic characteristics of genetic structure changes and construct a time series map of population genetic structure evolution. Based on the time series map data, the random forest algorithm was used to predict the direction of genetic structure change, and a quantitative association map of resistance gene transmission and population genetic structure reconstruction was generated. In the study of the temporal evolution of the genetic structure of weed populations, it is crucial to track the dynamic changes of genetic diversity through population genome data. Taking the study of rice and wild rice as an example, the population genetic diversity index monitored continuously for 10 generations showed that after the introduction of herbicide resistance genes, the population genetic diversity decreased rapidly in the first three generations, from 0.82 to 0.65. This change was highly correlated with the change in allele frequency of adaptive sites. The distribution map of population genetic variation at the genome level revealed the variation hotspots, of which 45% of the variation was concentrated in the 2MB area around the resistance gene. The dynamic changes in genetic distance reflect the evolution of population structure. In the study of wheat and Aegilops tauschii, deep learning methods predicted that the introgression of resistance genes led to accelerated population differentiation, and the genetic distance growth rate peaked in the 4th to 6th generations, increasing by 0.08 units per generation. This rapid differentiation is closely related to the spread of resistance genes in the population. For every 0.1 increase in the degree of introgression, the genetic distance between populations increased by an average of 0.15. The variation of genetic structure between generations showed significant time dependence. In the analysis of maize and its wild relatives, the kernel density estimation method showed that the dynamic distribution of genetic diversity had obvious bimodal characteristics, reflecting the two adaptation strategies coexisting in the population. The coefficient of variation of genetic structure showed periodic fluctuations at different time points, and the fluctuation period was highly consistent with the crop growing season. There was a significant correlation between the diffusion intensity of resistance genes and the changes in population genetic structure. In the study of rapeseed and its wild relatives, the gene diffusion intensity matrix showed that the genetic structure changes were most drastic in the diffusion front area, and the time correlation coefficient reached 0.85, indicating that the gene diffusion process drove the reorganization of the population genetic structure. The temporal evolution pattern of the population genetic structure shows the trajectory of adaptive evolution. In the case of soybean and wild soybean, the time series decomposition method identified three major evolutionary cycles, each lasting 4 to 6 generations.The periodic changes are highly synchronized with the seasonal changes in environmental selection pressure, highlighting the rhythmic characteristics of adaptive evolution. The prediction results of the random forest algorithm reveal the long-term evolutionary trend of the genetic structure of the population. By analyzing the multidimensional data of 1,000 sample points, the prediction accuracy reached 90%. The prediction results show that under continuous selection pressure, the genetic structure of the population carrying resistance genes will evolve in a direction of increasing resistance. The quantitative association map shows the key turning points in this evolutionary process, among which the time point when the genome reconstruction rate exceeds 0.4 is often accompanied by a significant increase in the fitness of the population.

[0030] Step S108, if the correlation is greater than or equal to the preset correlation threshold, a weed population genetic structure prediction model is constructed based on the weed population genetic structure change trend map and the resistance gene diffusion map to predict the future change trend of the weed population genetic structure, and a global weed gene database is constructed according to the future change trend of the weed population genetic structure.

[0031] Obtain resistance gene diffusion maps and weed population genetic structure change trend maps, and extract gene diffusion intensity and genetic structure variation data on the time series. Calculate the Pearson correlation coefficient between the gene diffusion intensity and genetic structure variation data; if the Pearson correlation coefficient is greater than or equal to the preset correlation threshold, use a deep reinforcement learning algorithm to train a genetic structure prediction model, the input layer of which contains a gene diffusion intensity matrix and historical genetic structure change data. Use the gradient descent method to optimize the loss function of the genetic structure prediction model, and use the back propagation algorithm to update the weight parameters of the genetic structure prediction model to obtain a trained genetic structure prediction model. According to the trained genetic structure prediction model, input an environmental factor data set in the future time period, the environmental factor data set includes temperature, precipitation and soil type change data, and predict the future weed population genetic structure. Extract the genotype frequency distribution data of the future weed population from the predicted results, and calculate the genetic diversity index at the population genome level. A future weed population genetic structure change trend curve is drawn according to the genetic diversity index, the future weed population genetic structure change trend curve is integrated with a global weed gene database, and the genetic structure information and gene function annotation data in the global weed gene database are updated.

[0032] Specifically, according to the resistance gene diffusion map and the trend map of genetic structure change of weed population, the gene diffusion intensity and genetic structure variation data on the time series are extracted, and the Pearson correlation coefficient between the two is calculated. If the Pearson correlation coefficient is greater than or equal to the preset correlation threshold, the genetic structure prediction model is trained by the deep reinforcement learning algorithm, and the model input layer contains the gene diffusion intensity matrix and the historical genetic structure change data. The loss function of the genetic structure prediction model is optimized by the gradient descent method, and the model weight parameters are updated by the back propagation algorithm to obtain the trained genetic structure prediction model. According to the trained genetic structure prediction model, the environmental factor data set in the future time period, including temperature, precipitation and soil type change data, is input to predict the genetic structure of the future weed population. The genotype frequency distribution data of the future weed population is extracted from the prediction results, and the genetic diversity index at the population genome level is calculated. According to the genetic diversity index, the trend curve of the genetic structure change of the future weed population is drawn, and the degree of genetic differentiation in different geographical regions is analyzed. The spatial interpolation algorithm is used to process the genetic structure change data of the future weed population to construct a global genetic structure change prediction map. According to the prediction map of genetic structure changes, the allele frequency distribution data of key gene loci are extracted, and combined with the gene annotation information in the global weed gene database, important genes that may produce adaptive advantages in the future are identified. Distributed computing technology is used to integrate the prediction results of the genetic structure of future weed populations with the global weed gene database, and the genetic structure information and gene function annotation data in the database are updated. Assuming that the preset correlation threshold is 8, firstly, the SNP data is obtained by gene sequencing of historical weed samples, and the population genetic structure is analyzed using software such as Structure, and the trend of the genetic structure of the weed population over time is drawn. For example, it is found that from 2020 to 2023, the proportion of genetic subgroup K1 in the weed population in region A increased from 15% to 35%. At the same time, the known resistance genotype data are collected to draw a resistance gene diffusion map. For example, it is found that the glyphosate resistance gene grg23 was detected in region A from 5% of the weed samples in 2020 to 25% in 2023. Then, a generalized linear mixed model (GLMM) was used, with the year as a fixed effect, the proportion of genetic subpopulations and the frequency of resistance genes in the weed population as the response variable, and the geographical location information (latitude and longitude) as a random effect. If the correlation coefficients between the year and the proportion of genetic subpopulations and the frequency of resistance genes were greater than or equal to 8, it was considered that a prediction model could be constructed. For example, the calculated correlation coefficients were 85 and 92, respectively. Based on the existing trend chart of the genetic structure change of weed populations and the diffusion map of resistance genes, a time series analysis method, such as the autoregressive moving average model (ARIMA), was used to construct a prediction model of the genetic structure of weed populations, with the year as the time unit and the proportion of each genetic subpopulation and the frequency of resistance genes as the prediction target.For example, using the ARIMA (2,1,2) model, based on the data from 2020 to 2023, it is predicted that the proportion of the K1 subpopulation in the weed population in region A in 2024 will be 40%, and the frequency of the grg23 gene will be 30%. In order to ensure the accuracy of data prediction, herbicide usage data can be added to quantify the use of herbicides into numerical values. For example, the use of a certain herbicide is 1 liter per hectare, which is recorded as 1, 2 liters per hectare, which is recorded as 2, and so on. They are added to the ARIMA model as covariates to improve the prediction accuracy. According to the prediction results, the predicted genetic subpopulation proportion and resistance gene frequency information, together with the corresponding geographic location information, year information, and detailed gene sequence data, such as raw sequencing data in FASTQ format or variant information in VCF format, are stored in a unified data format, such as JSON format, to build a global weed gene database, and the prediction model and database are updated regularly, such as annually, using new monitoring data.

[0033] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

Claims

1. A cross-platform global weed gene database construction method, characterized in that: The method comprises: Weed populations with gene exchange characteristics with target crops are screened out, gene sequence data of closely related weeds are extracted from the global weed gene database, and a gene introgression analysis model is constructed; gene comparison methods are used to identify gene fragments of gene exchange between crops and weeds in the screened weed populations, and the transmission path of resistance genes in interspecific hybridization is determined; based on the gene comparison results, the transmission path data of resistance genes are extracted, and a genetic structure change model is constructed in combination with the genetic structure characteristics of the weed population to analyze the impact of resistance gene transmission on the genetic structure of the weed population; population genetics algorithms are used to calculate the frequency distribution of resistance genes in different weed populations, and the diffusion trend of resistance genes in the weed population is determined based on the frequency distribution, generating a resistance gene diffusion map; based on the resistance gene diffusion map and combined with the genetic structure change model of the weed population, the impact of resistance gene transmission on the genetic indicators of the remodeling of the genetic structure of the weed population is analyzed, including changes in allele frequency and linkage disequilibrium degree. Based on the changes in the degree of genetic infiltration, a genetic structure remodeling map reflecting the changes in genetic indicators is generated; through the genetic structure remodeling map, the key nodes of the genetic structure changes of the weed population and the corresponding genetic indicators are extracted, and combined with the gene introgression analysis model, the impact of gene introgression on the adaptive evolutionary potential of the weed population and the change in the genetic diversity level of the population is analyzed; according to the impact of gene introgression on the genetic structure of the weed population, the changing trends of the genetic diversity index and the allele frequency of the adaptive site over time are generated, and a trend map of the genetic structure changes of the weed population is established, and combined with the resistance gene diffusion map, the correlation between the transmission of resistance genes and the changes in the genetic structure of the weed population is analyzed; if the correlation is greater than or equal to the preset correlation threshold, a genetic structure prediction model for the weed population is constructed based on the genetic structure change trend map of the weed population and the resistance gene diffusion map, and the future trend of the genetic structure changes of the weed population is predicted, and a global weed gene database is built according to the future trend of the genetic structure changes of the weed population.

2. The cross-platform global weed gene database construction method according to claim 1, characterized in that: The method of screening out weed populations that have gene exchange characteristics with target crops, extracting gene sequence data of closely related weed species from a global weed gene database, and constructing a gene introgression analysis model includes: Acquire genome data of weed populations in the area where the target crop is located, and construct a multidimensional feature vector including genome similarity and genetic distance based on the genome data; Extracting genome colinear regions according to the multidimensional feature vector, and sorting the genome data using a pedigree clustering algorithm to obtain a genome arrangement map; Calculate a genetic distance matrix according to the genome arrangement map, and determine the gene flow direction through the genome colinearity region of the genetic distance matrix to obtain the gene exchange frequency; Constructing a genome recombination index scoring matrix according to the gene exchange frequency, and calculating the hybridization affinity coefficient through the scoring matrix to obtain the gene introgression risk level; If the gene introgression risk level exceeds a preset threshold, it is determined that there is a risk of gene introgression between the target crop and the weed population.

3. The cross-platform global weed gene database construction method according to claim 1, characterized in that: The method of using a gene comparison method to identify gene fragments of gene exchange between crops and weeds and determining the transmission path of resistance genes in interspecific hybridization includes: The double sequence local alignment method is used to align the crop and weed genome data, and the gene fragment sequence similarity score is calculated based on the alignment results to obtain the target gene fragment with a sequence similarity score higher than a preset threshold; Extracting the coding region sequence from the target gene fragment, calculating the gene expression amount according to the transcriptome sequencing data of the coding region sequence, and calculating the genetic stability index by the gene expression amount and the gene copy number; Constructing a time series expression profile based on the genetic stability index and RNA sequencing data, fitting the time series expression profile using a gradient descent method, and obtaining gene transfer path data; A gene interaction network is constructed based on the gene transfer pathway data, and the key node sequence of gene transfer is obtained through a graph structure algorithm. The action site of the resistance gene is determined based on the comparison result between the key node sequence and the phenotype database, forming a transmission pathway diagram of the resistance gene during interspecies hybridization.

4. The cross-platform global weed gene database construction method according to claim 1, characterized in that: The method extracts the transmission path data of the resistance gene according to the gene comparison results, constructs a genetic structure change model based on the genetic structure characteristics of the weed population, and analyzes the impact of the resistance gene transmission on the genetic structure of the weed population, including: constructing a gene transmission network map according to genome resequencing data, wherein the gene transmission network map includes transmission association information between gene sites; According to the gene transmission network map, genomic variation site data is obtained, and a genetic structure characteristic matrix is ​​constructed using a population genetics algorithm, wherein the genetic structure characteristic matrix records population genotype distribution information; Constructing a population genetic structure dynamic change curve through the genetic structure characteristic matrix, wherein the population genetic structure dynamic change curve reflects the genetic variation accumulation rate in the time series; If the accumulation rate of genetic variation exceeds a preset threshold, the probability distribution of resistance gene transmission is calculated, and a genome recombination frequency heat map is constructed using the resistance gene transmission probability distribution. The genome recombination frequency heat map is used to display a quantitative description of the effect of resistance gene transmission on population genetic differentiation.

5. The cross-platform global weed gene database construction method according to claim 1, characterized in that: The population genetics algorithm is used to calculate the frequency distribution of resistance genes in different weed populations, and the diffusion trend of resistance genes in weed populations is determined according to the frequency distribution to generate a resistance gene diffusion map, including: The frequency distribution value of the resistance gene is calculated based on the population genome sequencing data to obtain the distribution density data of the resistance gene in the population; The spatial coordinate information was obtained by using the geographic information system, and the population geographic distribution matrix was constructed through the resistance gene distribution density data; The gene flow flux between adjacent geographical regions is calculated for the population geographic distribution matrix, and the gene diffusion rate curve is obtained after the population diffusion path is predicted by deep learning method; The gene diffusion radius is calculated according to the gene diffusion rate curve using a spatial statistics method, a diffusion trajectory map of the resistance gene in the geographic space is drawn using the gene diffusion radius, and a resistance gene diffusion map is generated based on the diffusion trajectory map.

6. The cross-platform global weed gene database construction method according to claim 1, characterized in that: The method analyzes the effect of resistance gene transmission on genetic indicators of genetic structure remodeling of weed populations, including changes in allele frequency and linkage disequilibrium, based on the resistance gene diffusion map and combined with the genetic structure change model of weed populations, and generates a genetic structure remodeling map reflecting changes in genetic indicators, including: A genotyping method is used to obtain population genotype data, and allele frequency distribution values ​​at the genome level are calculated based on the population genotype data; Calculate the linkage disequilibrium coefficient of the chromosome segment according to the allele frequency distribution value, and construct a linkage block distribution map within the genome through the linkage disequilibrium coefficient; Counting the spatial distribution of gene recombination breakpoints based on the linkage block distribution map, and constructing a population differentiation index matrix according to the spatial distribution; The population differentiation index matrix is ​​processed by a kernel density estimation method to obtain a genetic diversity distribution function, and a population genetic differentiation index value is calculated according to the genetic diversity distribution function; Combined with the population genetic differentiation index value, a support vector machine algorithm is used to predict the evolution direction of the genetic structure, and a genetic structure remodeling map including allele frequency and linkage disequilibrium is generated.

7. The method for building a cross-platform global weed gene database according to claim 1, characterized in that: The method of extracting the key nodes and corresponding genetic indicators of the genetic structure changes of the weed population through the genetic structure remodeling map, and combining the gene introgression analysis model to analyze the impact of gene introgression on the change of the adaptive evolutionary potential of the weed population and the change of the genetic diversity level of the population, includes: Extracting key sites of genome variation according to the genetic structure remodeling map data, calculating allele frequency change values ​​according to the key sites of genome variation, and constructing an evolution rate distribution map through the allele frequency change values; Calculating a genome sequence variation accumulation curve according to the evolution rate distribution map, using the variation accumulation curve to identify key genome adaptive sites, and obtaining population fitness change data through the key adaptive sites; Constructing a genome selection pressure gradient map for the population fitness change data, calculating the introgression pressure coefficient according to the selection pressure gradient map, constructing a population fitness curve, and calculating the population genetic diversity index; The adaptive evolutionary trend is predicted according to the population genetic diversity index, and the genome evolutionary potential is evaluated according to the predicted result of the adaptive evolutionary trend to obtain the evaluation result of the genome evolutionary potential.

8. The cross-platform global weed gene database construction method according to claim 1, characterized in that: The method generates the change trend of genetic diversity index and allele frequency of adaptive loci over time according to the influence of gene introgression on the genetic structure of weed population, establishes the trend map of genetic structure change of weed population, and analyzes the correlation between resistance gene transmission and genetic structure change of weed population in combination with resistance gene diffusion map, including: Obtaining the allele frequencies of adaptive sites in genome resequencing data, and constructing a population genetic variation distribution map based on the allele frequencies; Constructing a genetic distance dynamic curve according to the population genetic variation distribution map, and using a deep learning method to obtain the time evolution trend of the resistance gene introgression degree from the genetic distance dynamic curve; Calculating the genetic structure variation coefficient according to the temporal evolution trend of the resistance gene introgression degree, and constructing a genetic diversity dynamic distribution map from the genetic structure variation coefficient using a kernel density estimation method; A gene diffusion intensity matrix is ​​constructed according to the genetic diversity dynamic distribution map, and the time correlation coefficient between gene diffusion and genetic structure change is calculated.

9. The method for building a cross-platform global weed gene database according to claim 1, characterized in that: If the correlation is greater than or equal to a preset correlation threshold, a weed population genetic structure prediction model is constructed based on the weed population genetic structure change trend map and the resistance gene diffusion map to predict the future change trend of the weed population genetic structure, and a global weed gene database is constructed according to the future change trend of the weed population genetic structure, including: Obtain resistance gene diffusion maps and weed population genetic structure change trend maps, and extract gene diffusion intensity and genetic structure variation data in time series; Calculating the Pearson correlation coefficient between the gene diffusion intensity and the genetic structure variation data; If the Pearson correlation coefficient is greater than or equal to a preset correlation threshold, a deep reinforcement learning algorithm is used to train a genetic structure prediction model, wherein an input layer of the genetic structure prediction model includes a gene diffusion intensity matrix and historical genetic structure change data; The loss function of the genetic structure prediction model is optimized by using a gradient descent method, and the weight parameters of the genetic structure prediction model are updated by using a back propagation algorithm to obtain a trained genetic structure prediction model; According to the trained genetic structure prediction model, an environmental factor data set in a future time period is input, wherein the environmental factor data set includes temperature, precipitation and soil type change data, and the future genetic structure of the weed population is predicted; Extract the genotype frequency distribution data of the future weed population from the predicted results, and calculate the genetic diversity index at the population genome level; A future weed population genetic structure change trend curve is drawn according to the genetic diversity index, the future weed population genetic structure change trend curve is integrated with a global weed gene database, and the genetic structure information and gene function annotation data in the global weed gene database are updated.

Images