Method and system for generating intervention strategy for vaginal micro-ecological imbalance based on multi-omics data

Abstract

The application provides a method and system for generating intervention strategies for vaginal micro-ecological imbalance based on multi-omics data, relating to the technical field of data analysis, by obtaining multi-omics data of the vaginal micro-ecological system, including proteomic data, microbiome data, metabolomic data, and transcriptomic data; performing dimension reduction and data cleaning processing on the multi-omics data to generate a standardized data set; performing feature correlation processing on the standardized data set to generate a feature correlation matrix; using causal reasoning technology to analyze the feature correlation matrix to extract key influencing factors that cause vaginal micro-ecological imbalance and determine the interaction relationships between the key influencing factors; performing collaborative analysis on the standardized data set, the key influencing factors, and the interaction relationships to generate a data analysis result, and further generating an intervention strategy for vaginal micro-ecological imbalance, thereby improving the precision targeting ability of the intervention scheme for the root cause of vaginal micro-ecological imbalance and the individualized diagnosis and treatment efficiency.

Description

Technical Field

[0001] This application relates to the field of data analysis technology, and in particular to a method and system for generating intervention strategies for vaginal microecological imbalance based on multi-omics data. Background Technology

[0002] In the field of women's reproductive health, vaginal microecological imbalance is the root cause of various gynecological diseases. Clinically, there is an urgent need for a technology that can integrate multi-dimensional biological information, deeply analyze pathogenic mechanisms, and generate precise, personalized intervention plans. Faced with complex microbial communities and host interaction networks, relying solely on single-dimensional detection data is insufficient to meet the need for in-depth analysis of disease development mechanisms. There is an urgent need to establish an intelligent analysis system that can systematically integrate multi-source heterogeneous data and guide clinical decision-making.

[0003] One current solution to this need is a microbiome assessment method based on machine learning classification models. This method primarily utilizes statistical learning algorithms to train microbiome abundance and a few metabolic indicators obtained from high-throughput sequencing. It constructs a classifier to identify characteristic patterns in healthy and imbalanced states, predicts disease risk based on statistical regularities in historical case data, and recommends general probiotic supplementation or drug treatment regimens accordingly, thus providing auxiliary diagnosis for common microbiome disorders. However, this approach only focuses on surface-level statistical correlation analysis and fails to delve into the causal driving mechanisms between different biomolecular levels. This makes it difficult to pinpoint the core pathogenic factors and their interaction pathways that cause imbalances when faced with complex and variable individual differences. Due to the lack of collaborative analysis and causal inference of multi-level data such as proteins, genes, and metabolites, the resulting intervention strategies often lack specificity and cannot fundamentally reverse the microbiome imbalance, failing to meet the deeper needs of precision medicine. Summary of the Invention

[0004] The purpose of this application is to provide a method and system for generating intervention strategies for vaginal microecological imbalance based on multi-omics data, in order to solve the problem that the lack of multi-level data collaborative analysis and causal inference in the existing technology makes it difficult to locate the core pathogenic factors and the intervention strategies are not targeted.

[0005] To address the aforementioned technical problems, in a first aspect, this application provides a method for generating intervention strategies for vaginal microecological imbalance based on multi-omics data, including:

[0006] Obtain multi-omics data of vaginal microecology, including proteomic data, microbial community data, metabolomic data, and transcriptomic data;

[0007] The multi-omics data are subjected to dimensionality reduction and data cleaning to generate a standardized dataset;

[0008] The standardized dataset is subjected to feature association processing to generate a feature association matrix;

[0009] The feature correlation matrix was analyzed using causal reasoning techniques to extract key influencing factors leading to vaginal microecological imbalance and to determine the interaction relationships among these key influencing factors.

[0010] The standardized dataset, the key influencing factors, and their interactions are subjected to collaborative analysis to generate data analysis results.

[0011] Based on the data analysis results, intervention strategies for vaginal microecological imbalance are generated.

[0012] Secondly, this application provides a system for generating intervention strategies for vaginal microecological imbalance based on multi-omics data, including:

[0013] The acquisition module is used to acquire multi-omics data of the vaginal microecology, including proteomic data, microbial community data, metabolomic data, and transcriptomic data.

[0014] The processing module is used to perform dimensionality reduction and data cleaning on the multi-omics data to generate a standardized dataset;

[0015] The association module is used to perform feature association processing on the standardized dataset to generate a feature association matrix;

[0016] The determination module is used to analyze the feature correlation matrix using causal reasoning techniques to extract key influencing factors that lead to vaginal microecological imbalance and to determine the interaction relationships between these key influencing factors.

[0017] The parsing module is used to perform collaborative parsing processing on the standardized dataset, the key influencing factors, and their interactions to generate data parsing results;

[0018] The generation module is used to generate intervention strategies for vaginal microecological imbalance based on the data analysis results.

[0019] Thirdly, this application provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement the steps of the method for generating intervention strategies for vaginal microecological imbalance based on multi-omics data as described in the first aspect above.

[0020] Fourthly, this application provides a computer storage medium storing a computer program, which, when executed by a computer, implements the steps of the method for generating intervention strategies for vaginal microecological imbalance based on multi-omics data as described in the first aspect above.

[0021] The method for generating intervention strategies for vaginal microecological imbalance based on multi-omics data provided in this application has the following beneficial effects: By acquiring and standardizing four-dimensional omics data of proteins, flora, metabolism and transcriptome, a unified analytical benchmark for high-dimensional heterogeneous data is constructed. The feature correlation matrix is ​​used to comprehensively capture the complex connections between different biomolecular layers. Then, by using causal reasoning technology to penetrate the fog of statistical correlation, the core driving factors that cause microecological imbalance and their interaction networks are accurately identified. Finally, through the collaborative analysis of multi-dimensional data, intervention strategies with deep mechanistic support are generated. This realizes the intelligent upgrade of the entire chain from multi-source data fusion to pathogenic mechanism tracing to precise plan formulation, and improves the adaptability of the diagnosis and treatment plan to the complex microecological environment of the individual and the ability to fundamentally repair it.

[0022] Furthermore, by establishing a mechanism for determining intervention priorities, levels, and directions based on trigger intensity and correlation, and by combining direct action relationships to dynamically adjust the linkage and intensity of intervention methods, abstract data analysis results are transformed into concrete, executable, and internally self-consistent comprehensive intervention plans. This avoids the limitations of single intervention methods and potential conflicts when multiple measures are implemented in parallel, ensuring precise matching and overall optimization of intervention strategies at multiple levels and dimensions, thereby maximizing the restoration of vaginal microecological balance and improving the pertinence and reliability of clinical treatment. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 A flowchart illustrating the method for generating intervention strategies for vaginal microecological imbalance based on multi-omics data, as provided in this application embodiment;

[0025] Figure 2 A schematic diagram illustrating the specific implementation of the method for generating intervention strategies for vaginal microecological imbalance based on multi-omics data provided in this application embodiment;

[0026] Figure 3 This is a schematic diagram of the structure of the vaginal microecological imbalance intervention strategy generation system based on multi-omics data provided in the embodiments of this application. Detailed Implementation

[0027] To address the challenge that existing machine learning solutions, limited by statistical correlation analysis, struggle to reveal causal driving mechanisms among multi-omics data, resulting in an inability to accurately pinpoint core pathogenic factors and a lack of targeted intervention strategies, this application provides a method for generating intervention strategies for vaginal microecological imbalance based on multi-omics data. The core idea of ​​this method is to overcome the limitations of single-dimensional or superficial correlation analysis by simultaneously integrating four-dimensional omics information: proteomics, microbiota, metabolism, and transcription. By constructing a cross-omics feature association matrix and introducing causal reasoning techniques, the key driving factors causing imbalance and their interaction pathways are extracted from the complex biological network. Then, standardized data and causal chains are synergistically analyzed to ultimately generate personalized intervention strategies that directly target the root cause of the lesions.

[0028] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0029] The core of this application is to provide a method for generating intervention strategies for vaginal microecological imbalance based on multi-omics data. A flowchart of one specific implementation method is shown below. Figure 1 As shown, the method includes:

[0030] Step 1: Obtain multi-omics data of vaginal microecology, including proteomic data, microbial community data, metabolomic data, and transcriptomic data.

[0031] In this step, multi-omics data refers to a dataset that integrates multiple biomolecular-level information related to the vaginal microecology, used to comprehensively reflect the physiological and pathological state of the vaginal microecology, and obtained based on multi-dimensional detection of biological samples.

[0032] Proteomic data refers to the expression levels and distribution of all proteins in the vaginal microecological environment. It is used to reflect the protein expression characteristics of vaginal tissues and microorganisms and is obtained based on protein detection of biological samples.

[0033] Vaginal community data refers to the types, abundance, and distribution of various microorganisms in the vagina. It is used to reflect the composition and structure of the vaginal microbial community and is obtained based on the detection of the vaginal community in biological samples.

[0034] Metabolomics data refers to the types and contents of various metabolites in the vaginal microecological environment, which are used to reflect the metabolic activity characteristics of microorganisms and the host, and are obtained based on metabolic detection of biological samples.

[0035] Transcriptome data refers to the expression level data of gene transcription products in vaginal tissues and microorganisms, which is used to reflect the transcriptional activity and expression characteristics of genes and is obtained based on the transcriptional level detection of biological samples.

[0036] Step 2: Perform dimensionality reduction and data cleaning on the multi-omics data to generate a standardized dataset.

[0037] In this step, the standardized dataset refers to a multi-omics dataset that has undergone dimensionality reduction and data cleaning to remove redundant information and invalid data, and whose data format and dimensions are unified.

[0038] In this embodiment, dimensionality reduction and data cleaning are performed sequentially on the multi-omics data of vaginal microecology to reduce redundant information dimensions that have no practical distinguishing value, and to filter out invalid content such as abnormal values ​​and missing key information. After continuous dimensionality reduction and cleaning operations, a standardized dataset with uniform format and valid information is generated.

[0039] Step 3: Perform feature association processing on the standardized dataset to generate a feature association matrix.

[0040] In this step, the feature association matrix refers to a matrix that integrates the first association degree value and the second association degree value to intuitively present the association strength between various omics features. This matrix includes the association of features within the same omics and the association of features between different omics.

[0041] In this embodiment of the application, step 3 specifically includes steps 301-303:

[0042] Step 301: Perform feature association analysis on the standardized proteome data, standardized microbiome data, standardized metabolome data and standardized transcriptome data in the standardized dataset to determine the first degree of association between each feature.

[0043] In this step, standardized proteome data, standardized microbiome data, standardized metabolome data, and standardized transcriptome data refer to protein-related data, microbiome-related data, metabolite-related data, and gene transcription-related data that have been processed by dimensionality reduction and data cleaning, and are in a unified format and contain valid information.

[0044] Each feature can be understood as various protein expression indicators included in standardized proteomics data, various microbial species and abundance indicators included in standardized microbial community data, various metabolite content indicators included in standardized metabolomics data, and various gene transcription level indicators included in standardized transcriptomics data. These are specific detection indicator units corresponding to the four types of standardized omics data, directly belonging to and corresponding to each type of standardized omics data.

[0045] The first correlation value refers to the quantitative value of the degree of correlation between different features within a single class of standardized omics data.

[0046] In this embodiment of the application, independent feature association analysis is performed on standardized proteome data, standardized microbiome data, standardized metabolome data, and standardized transcriptome data. All feature items within each type of standardized omics data are compared pairwise to analyze the degree of association between each pair of feature items, so as to calculate the quantitative value corresponding to each pair of feature items. This value is the first degree of association value between each feature item.

[0047] The first correlation value between each feature item is obtained by calculating the covariance of each pair of feature items divided by the product of their standard deviations. The covariance refers to the degree of linear correlation between the observations of each pair of feature items, reflecting the consistency of the changing trends of the two types of feature items; the standard deviation refers to the dispersion of all observations of a single feature item, reflecting the fluctuation of the detection data of that feature item; the observation value refers to the specific detection value corresponding to each feature item in the standardized omics data, that is, the detection value of each protein expression index in the standardized proteome data, the detection value of each microbial species and abundance in the standardized microbial community data, the detection value of the content of each metabolite in the standardized metabolome data, and the detection value of the transcription level of each gene in the standardized transcriptome data.

[0048] Step 302: Perform cross-correlation analysis on the standardized proteome data, standardized microbiome data, standardized metabolome data, and standardized transcriptome data respectively to determine the second correlation value between the feature items of different standardized omics data.

[0049] In this step, different standardized omics data refers to any two different types of standardized omics data in a standardized dataset.

[0050] The second correlation value refers to the quantitative value of the degree of correlation between the features contained in different types of standardized omics data.

[0051] In this embodiment, the four types of standardized omics data are first combined in pairs. Cross-correlation analysis is performed on each different set of standardized omics data. All features of one type of standardized omics data are compared with all features of another type of standardized omics data in pairs. The degree of correlation between each pair of cross-omics features is analyzed, and the corresponding quantitative value is calculated. This value is the second correlation value between the features of different standardized omics data.

[0052] The second correlation value between feature items of different standardized omics data is obtained by calculating the covariance of each pair of cross-omics feature items divided by the product of their standard deviations. Here, the covariance refers to the degree of linear correlation between the observations of each pair of cross-omics feature items, reflecting the consistency of the changing trends of the two types of feature items under different omics; the standard deviation refers to the dispersion of all observations of a single cross-omics feature item, reflecting the fluctuation of the detection data of that feature item; the observation value refers to the specific detection value corresponding to each feature item in different standardized omics data, namely the protein expression index value of standardized proteomics data, the microbial species and abundance value of standardized microbial community data, the metabolite content value of standardized metabolomics data, and the gene transcription level value of standardized transcriptomics data.

[0053] Step 303: Based on the characteristic correspondence of each standardized omics data, integrate the first correlation degree value and the second correlation degree value to form a feature correlation matrix.

[0054] In this step, feature correspondence refers to the pairing relationship of feature items within the same standardized omics data, as well as the pairing relationship of feature items between different standardized omics data. The feature association matrix is ​​a structured dataset that integrates the first and second association strength values, arranged in order according to feature correspondence, and is used to reflect the strength of association between all feature items.

[0055] In this embodiment, the feature correspondence of all feature items is first sorted out, and the feature item pairing corresponding to each first correlation degree value and second correlation degree value is clarified. Then, according to the feature correspondence of each standardized omics data, all first correlation degree values ​​and second correlation degree values ​​are sequentially filled into the corresponding positions, and the two types of correlation degree values ​​are integrated in an orderly manner to finally form a complete feature correlation matrix.

[0056] The embodiments of this application can comprehensively explore the inherent relationships among various features in standardized datasets, providing complete correlation data support for subsequent extraction of key influencing factors and analysis of interaction relationships, and improving the comprehensiveness and systematicness of multi-omics data utilization.

[0057] Step 4: Analyze the feature correlation matrix using causal reasoning techniques to extract key influencing factors that lead to vaginal microecological imbalance and determine the interaction relationships among these key influencing factors.

[0058] In this step, the key influencing factors refer to the core feature items that are screened and identified from the feature association matrix and can cause vaginal microecological imbalance. These are the main factors that cause the vaginal microecology to deviate from the normal state.

[0059] Interaction can be understood as the pattern of interaction between key influencing factors. This interaction includes direct and indirect interactions, and is used to reflect the mutual influence and mutual constraints among key influencing factors.

[0060] In this embodiment of the application, step 4 specifically includes steps 401-403, and step 403 includes steps 411-413:

[0061] Step 401: Select target first correlation degree values ​​and target second correlation degree values ​​that are higher than the preset threshold from the feature correlation matrix.

[0062] In this step, the preset threshold refers to a pre-set critical value used to determine whether the correlation between feature items has analytical value. In this embodiment, the size of the threshold is not limited, and it can be set according to the conventional correlation strength range of vaginal microecology multi-omics data.

[0063] The target first correlation value refers to the correlation value between feature items within the same standardized omics data that are selected and exceed the preset threshold.

[0064] The second correlation value refers to the correlation value between feature items of different standardized omics data that are selected and exceed the preset threshold.

[0065] In this embodiment, firstly, based on the association characteristics of vaginal microecology multi-omics data, a reasonable preset threshold is set, and then all first association degree values ​​and second association degree values ​​in the feature association matrix are extracted one by one; then, each association degree value is compared with the preset threshold, and all first association degree values ​​and second association degree values ​​higher than the preset threshold are selected and determined as target first association degree values ​​and target second association degree values, respectively. This step removes association degree values ​​that are lower than or equal to the preset threshold and retains association data with practical analytical value.

[0066] Step 402: Combine the feature terms corresponding to the first correlation degree value and the second correlation degree value of each target respectively to obtain multiple feature term combinations.

[0067] In this step, a feature combination refers to a set formed by pairing two feature items corresponding to the first correlation value or the second correlation value of the target. Each feature combination contains two feature items that are strongly correlated.

[0068] In this embodiment, firstly, for each target's first correlation degree value, two feature terms within the same omics are found, and these two feature terms are paired and combined to form a feature term combination; secondly, for each target's second correlation degree value, two feature terms between different omics are found, and these two feature terms are paired and combined to form a feature term combination; the pairing of feature terms corresponding to all target's first correlation degree values ​​and target's second correlation degree values ​​is completed sequentially, and all pairing results are integrated to obtain multiple feature term combinations.

[0069] Step 403: Use causal reasoning techniques to analyze the relationships between the combinations of features to identify the target features that cause changes. These target features are considered as key influencing factors leading to vaginal microecological imbalance, and the interaction relationships between these key influencing factors are analyzed.

[0070] In this step, the target feature refers to the core feature identified from the combination of features that can trigger changes in other features and is the key to causing vaginal microecological imbalance.

[0071] Interaction relationship refers to the direct or indirect correlation between key influencing factors, which is determined by the combination of characteristic items corresponding to each key influencing factor and the degree of correlation.

[0072] In this embodiment of the application, step 403 specifically includes steps 411-413:

[0073] Step 411: Use causal reasoning techniques to analyze the relationships between the combinations of features to identify the first feature that triggers the change. Combine the changes of each first feature in the corresponding combination of features to assign a trigger strength value to each first feature.

[0074] In this step, the first feature refers to the feature identified from the feature combination that can actively trigger changes in other feature items, and is the core triggering factor in the feature combination.

[0075] Change performance refers to the magnitude and trend of change of the first feature relative to another feature in the combination of corresponding feature items.

[0076] The trigger strength value refers to the numerical value used to quantify the ability of the first feature to cause vaginal microecological imbalance. The higher the value, the stronger the ability of the feature to cause microecological imbalance.

[0077] In this embodiment of the application, causal reasoning technology is used to analyze the causal relationship between the two feature terms in each feature term combination one by one, determine which feature term is the active factor that causes the change of the other feature term, and determine the feature term that actively causes the change as the first feature term.

[0078] Subsequently, the variation amplitude and trend of the first feature item in the corresponding feature item combination are extracted. Combined with its corresponding target correlation value, the trigger intensity value of each first feature item is obtained by calculating the product of the variation amplitude and the target correlation value. Here, the variation amplitude refers to the degree of deviation of the observed value of the first feature item from the normal range, the target correlation value refers to the target first correlation value or target second correlation value corresponding to the feature item combination in which the first feature item is located, and the observed value refers to the specific detection value corresponding to each feature item.

[0079] Step 412: Select the target feature whose trigger intensity value is not lower than the preset intensity value from each first feature item as the key influencing factor causing vaginal microecological imbalance.

[0080] In this step, the preset intensity value refers to a pre-set critical value used to determine whether the first feature can be used as a key influencing factor. In this embodiment of the application, the size of the threshold is not limited, but can be set according to the conventional trigger intensity range of vaginal microecological imbalance.

[0081] The target feature refers to the feature selected from the first feature with a trigger intensity value not lower than the preset intensity value, which is the core factor leading to the imbalance of the vaginal microecology.

[0082] In this embodiment, firstly, based on the clinical characteristics of vaginal microecological imbalance and the patterns of multi-omics data, a reasonable preset intensity value is set, and then the trigger intensity value corresponding to each first feature item is extracted one by one; then, each trigger intensity value is compared with the preset intensity value, and all first feature items with trigger intensity values ​​not lower than the preset intensity value are selected and identified as target feature items, and these target feature items are used as key influencing factors leading to vaginal microecological imbalance.

[0083] Step 413: Determine the interaction relationship between the key influencing factors based on the trigger intensity values ​​of each key influencing factor.

[0084] In this step, the interaction relationship refers to the direct and indirect interaction relationships between the key influencing factors. The direct interaction relationship means that the characteristic items corresponding to the two key influencing factors are directly related, that is, they belong to the same characteristic item combination. The indirect interaction relationship means that the two key influencing factors are related through other key influencing factors. The judgment criteria include the trigger intensity value of each key influencing factor and the corresponding characteristic item combination association.

[0085] In this embodiment of the application, the feature combination corresponding to all key influencing factors is first sorted out. If the feature items corresponding to two key influencing factors belong to the same feature combination, that is, there is a first correlation value or a second correlation value of the target, then it is determined that there is a direct relationship between the two key influencing factors. At the same time, the difference between the two trigger intensity values ​​is combined. The smaller the difference, the stronger the correlation of the direct relationship.

[0086] If the features corresponding to two key influencing factors do not belong to the same feature combination, but are indirectly related through one or more other key influencing factors, then it is determined that there is an indirect relationship between the two key influencing factors. Combined with the trigger strength value of the intermediate key influencing factor, the higher the trigger strength value of the intermediate factor, the stronger the indirect relationship.

[0087] The embodiments of this application can comprehensively explore the core factors and correlation patterns that lead to vaginal microecological imbalance, providing accurate core data support for subsequent collaborative data analysis and intervention strategy generation, solving the problem that single-dimensional analysis cannot locate the core pathogenic factors, and improving the accuracy and systematicness of the microecological imbalance mechanism analysis.

[0088] Step 5: Perform collaborative analysis on the standardized dataset, the key influencing factors, and their interactions to generate data analysis results.

[0089] In this step, the data analysis results refer to the comprehensive analysis results that integrate the omics characteristic information of the standardized dataset, the core attributes of key influencing factors, and the interaction relationships among key influencing factors, which are used to intuitively reflect the intrinsic mechanism and correlation of vaginal microecological imbalance.

[0090] In this embodiment, each type of omics feature in the standardized dataset is first matched with the corresponding key influencing factors one by one, and the omics feature performance of the key influencing factors is associated. At the same time, the interaction relationship type and correlation degree between the key influencing factors are combined to analyze the intrinsic relationship between omics features, key influencing factors and interaction relationships, and sort out the core logic related to vaginal microecological imbalance. After multiple rounds of association matching and integration analysis, the data analysis results that can clearly reflect the internal mechanism of vaginal microecological imbalance and the correlation law between omics features and key influencing factors are finally generated.

[0091] Step 6: Based on the data analysis results, generate intervention strategies for vaginal microecological imbalance.

[0092] In this step, the intervention strategy for vaginal microecological imbalance refers to a personalized regulatory plan that integrates the intervention levels, directions, methods, intensity, and linkage adaptation requirements of key influencing factors to specifically correct the state of vaginal microecological imbalance.

[0093] In the embodiments of this application, such as Figure 2 As shown, step 6 specifically includes steps 601-606, steps 611-614 (including step 611) of step 604, and steps 621-623 (including step 621) of step 605:

[0094] Step 601: Based on the data analysis results and the characteristic performance of the corresponding omics data in the standardized dataset, determine the intervention priority of each key influencing factor.

[0095] In this step, intervention priority refers to the order of interventions assigned to each key influencing factor. The higher the priority of a key influencing factor, the more likely it is to require intervention.

[0096] In this embodiment, the trigger intensity and interaction relationship of each key influencing factor are first extracted from the data parsing results. Combined with the feature performance of the corresponding omics data in the standardized dataset, the degree of deviation of the omics features corresponding to each key influencing factor from the normal state is determined. Key influencing factors with higher trigger intensity values, greater deviation of omics features, and wider range of associated influence are judged as having higher intervention priority.

[0097] For example, one way to determine intervention priority is to first classify the trigger intensity value into three levels: high, medium, and low; the degree of deviation of omics characteristics from the normal state into three levels: significant, moderate, and slight; and the scope of associated influence into three levels: broad, medium, and narrow, based on the number of other key influencing factors directly affected by the key influencing factor. For each key influencing factor, assign values ​​to the three dimensions: 3 points for high trigger intensity, significant deviation, and broad scope of influence; 2 points for medium level; and 1 point for low level. Calculate the total score for the three dimensions; the higher the total score, the higher the intervention priority. If the total scores are the same, the key influencing factor with the higher trigger intensity value is given higher priority.

[0098] Step 602: Based on the trigger intensity values ​​of each key influencing factor and the degree of correlation between the interactions among the key influencing factors, determine the intervention level of the key influencing factors and the intervention direction of the key influencing factors at different intervention levels.

[0099] In this step, intervention level refers to the level of intervention based on the importance of key influencing factors, and intervention direction refers to the direction of regulation taken for key influencing factors at different intervention levels.

[0100] In this embodiment, intervention levels are divided according to the trigger intensity value of each key influencing factor. The key influencing factor with the highest trigger intensity value is classified into the core intervention level, and the rest are classified into the secondary intervention level. Then, based on the degree of correlation between the interaction relationships between the key influencing factors, a negative regulation intervention direction is set for the key influencing factors that cause imbalance, and a positive regulation intervention direction is set for the key influencing factors that maintain ecological balance, thereby determining the intervention direction of key influencing factors at each level.

[0101] Step 603: Based on the omics data characteristics corresponding to each key influencing factor, determine the first intervention method for key influencing factors at different intervention levels and in different intervention directions.

[0102] In this step, omics data features refer to the specific features corresponding to the standardized omics data of each key influencing factor. These features include specific information such as the type, expression level, abundance, or content of the feature items, reflecting the biological attributes and abnormal performance of the key influencing factors.

[0103] The first intervention method refers to the basic intervention operation method that targets a single key influencing factor without linkage and adaptation, and matches it one by one with the omics data characteristics corresponding to the key influencing factor.

[0104] In this embodiment, the omics type of each key influencing factor is first identified, its corresponding omics data features are extracted, and the appropriate first intervention method is determined by combining the intervention level and intervention direction of the key influencing factor.

[0105] For example, if the key influencing factor belongs to standardized microbial community data, and its omics data characteristics show that the abundance of a certain type of harmful microorganism is too high and it is at the core intervention level, then the operation of inhibiting the growth of this type of harmful microorganism is determined as the first intervention method. If the key influencing factor belongs to standardized metabolomics data, and its omics data characteristics show that the content of a certain type of beneficial metabolite is too low and it is at the secondary intervention level, then the operation of supplementing this type of beneficial metabolite is determined as the first intervention method. If the key influencing factor belongs to standardized proteomics data or standardized transcriptomics data, combined with its omics data characteristics such as protein expression level and gene transcription level, the corresponding basic intervention operations that regulate protein expression and regulate gene transcription are determined as the first intervention method, ensuring that each first intervention method is precisely matched with the omics data characteristics, intervention level and intervention direction of the key influencing factor.

[0106] Step 604: Based on the trigger intensity values ​​of each key influencing factor and the interaction relationship between each key influencing factor, set the target intervention intensity corresponding to each first intervention method.

[0107] In this step, the target intervention intensity refers to the final regulatory intensity of the first intervention method, which is the final intervention intensity value obtained after initial setting and coefficient adjustment.

[0108] In this embodiment of the application, step 604 specifically includes steps 611-614:

[0109] Step 611: Based on the trigger intensity value of each key influencing factor, set the initial intervention intensity of the first intervention method corresponding to each key influencing factor.

[0110] In this step, the initial intervention intensity refers to the initial intervention intensity set directly based on the trigger intensity value of the key influencing factors, without any associated adjustments.

[0111] In this embodiment, a positive correlation is established between the trigger intensity value of each key influencing factor and the initial intervention intensity. The higher the trigger intensity value, the greater the initial intervention intensity of the corresponding first intervention method. The corresponding initial intervention intensity is set for all key influencing factors according to this mapping relationship.

[0112] For example, one approach is to categorize key influencing factors into three levels—high, medium, and low—based on their trigger intensity values. Key influencing factors with high trigger intensity values ​​are assigned a strong initial intervention level; those with medium trigger intensity values ​​are assigned a medium initial intervention level; and those with low trigger intensity values ​​are assigned a weak initial intervention level. For key influencing factors at the core intervention level, the initial intervention level is uniformly increased by one level based on the mapping results. For key influencing factors at the secondary intervention level, the mapped initial intervention level is directly used. This method completes the initial intervention level settings for all key influencing factors.

[0113] Step 612: For the first key influencing factors with direct interaction relationships, determine the adjustment coefficients corresponding to each first key influencing factor based on the degree of correlation between the interaction relationships among the first key influencing factors.

[0114] In this step, the first target key influencing factor refers to the key influencing factor that has a direct relationship with other key influencing factors.

[0115] The adjustment coefficient is a quantitative coefficient used to correct the initial intervention intensity. The higher the correlation, the larger the value of the adjustment coefficient. For key influencing factors that are not directly related to the interaction relationship, no adjustment coefficient is set, and the adjustment coefficient defaults to 1.

[0116] In this embodiment, for the first target key influencing factors whose interaction relationship is direct, an adjustment coefficient is set according to the degree of correlation of their interaction relationship. The higher the degree of correlation, the larger the adjustment coefficient. For key influencing factors whose interaction relationship is not direct, the adjustment coefficient is uniformly set to 1.

[0117] Step 613: Based on the initial intervention intensity and the adjustment coefficient, calculate the intermediate intervention intensity of the first intervention method corresponding to each key influencing factor of the first objective.

[0118] In this step, the intermediate intervention intensity refers to the intermediate intervention strength obtained by multiplying the initial intervention intensity by the adjustment coefficient, which is the intervention intensity value after a single adjustment.

[0119] In this embodiment, the initial intervention intensity corresponding to each key influencing factor of the first objective is multiplied by the adjustment coefficient to obtain the intermediate intervention intensity of the first intervention method corresponding to the key influencing factor of the first objective; for key influencing factors whose interaction relationship is not direct, the initial intervention intensity remains unchanged.

[0120] Step 614: For the key influencing factors of the first objective that have multiple direct relationships with the key influencing factors of the second objective, adjust the intermediate intervention intensity corresponding to each key influencing factor of the second objective according to the corresponding multiple adjustment coefficients to obtain the target intervention intensity corresponding to each first intervention method.

[0121] In this step, the key influencing factors of the second objective refer to the key influencing factors of the first objective that have multiple direct relationships with each other.

[0122] In this embodiment, for a second key influencing factor with multiple direct relationships, the average of its corresponding adjustment coefficients is calculated, and then the average is multiplied by the intermediate intervention intensity to obtain the target intervention intensity of the first intervention method corresponding to the second key influencing factor; for a first key influencing factor with only a single direct relationship, its initial intervention intensity is the target intervention intensity.

[0123] Step 605: For the first target key influencing factor that has a direct relationship with the key influencing factors, the first intervention method corresponding to the first target key influencing factor needs to be linked and adapted to obtain the second intervention method.

[0124] In this embodiment of the application, step 605 specifically includes steps 621-623:

[0125] Step 621: Pair the first key influencing factors that have a direct relationship with the target to obtain multiple pairs.

[0126] In this step, pairing refers to combining two primary target key influencing factors that have a direct relationship into a group, forming a key influencing factor combination unit.

[0127] In this embodiment of the application, all first target key influencing factors that have a direct relationship are traversed, and every two first target key influencing factors that have a direct relationship are combined into a pair. All combination operations are completed in sequence to obtain multiple first target key influencing factor pairs.

[0128] Step 622: Analyze whether there is mutual interference between the first intervention methods corresponding to the first key influencing factors of each pair during the execution process. For the first pair that has mutual interference, determine the adjustment priority of the first intervention method corresponding to the first pair based on the degree of correlation of the direct action relationship corresponding to the first pair and the trigger intensity value of each first key influencing factor.

[0129] In this step, mutual interference refers to the conflict or cancellation of effects that occur when two first intervention methods are executed simultaneously, and adjustment priority refers to the order in which the first intervention methods that are mutually interfering are adjusted.

[0130] In this embodiment, the execution logic of the first intervention method corresponding to the two first target key influencing factors in each pair is first analyzed to determine whether there is mutual interference with conflicting or canceling effects; then, for the first pair with mutual interference, the first intervention method with higher trigger intensity value and higher correlation is set as higher adjustment priority.

[0131] For example, a certain first pair contains two first target key influencing factors that have a direct relationship. The first key influencing factor belongs to standardized microbial community data, and the corresponding omics data characteristics are that the abundance of harmful microorganism A is too high, the trigger strength value is 8, and the maximum trigger strength value is 10. The degree of association between the first key influencing factor and the other key influencing factor in the pair is 0.9, and the degree of association is 1.0. The first intervention method is the local application of antimicrobial agents to inhibit the proliferation of microorganism A.

[0132] The second key influencing factor, identified in standardized metabolomics data, is characterized by low levels of beneficial metabolite B (trigger strength value 6). Its direct correlation with the first key influencing factor is 0.9, and the primary intervention is oral supplementation of metabolite B. Analysis revealed that antimicrobial agents inhibit the synthesis and absorption of metabolite B, leading to a counteracting effect when both are applied simultaneously. Given the higher trigger strength of the first key influencing factor and the consistent correlation between the two, the primary intervention corresponding to the first key influencing factor—"topical application of antimicrobial agents to inhibit the proliferation of microorganism A"—is given a higher priority. Subsequent adjustments to the other intervention will be based on the core requirements of this first intervention to eliminate interference.

[0133] Step 623: According to the adjustment priority and the omics data characteristics of the key influencing factors of the first target in the first pairing, adjust the first intervention method corresponding to the first pairing that has mutual interference to obtain the second intervention method, and take the first intervention method corresponding to the first pairing that does not have mutual interference as the second intervention method.

[0134] In this embodiment of the application, according to the determined adjustment priority, combined with the omics data characteristics corresponding to the key influencing factors of the first target in the first pairing, the parameters of the first intervention method that has mutual interference are adjusted, and the second intervention method is obtained after eliminating the interference.

[0135] For example, one adjustment method could be: if the first pairing contains two key influencing factors with a direct relationship, the key influencing factor with higher priority corresponds to standardized microbial community data characteristics, and the first intervention method is to inhibit the proliferation of harmful microorganisms; the key influencing factor with lower priority corresponds to standardized metabolomics data characteristics, and the first intervention method is to regulate the synthesis of metabolites. If the simultaneous execution of the two intervention methods will cause mutual interference of microenvironmental conflict, then the core parameters of the intervention method with higher priority, inhibiting the proliferation of harmful microorganisms, remain unchanged, while the implementation sequence and intensity of the intervention method with lower priority, regulating the synthesis of metabolites, are adjusted. The simultaneous implementation is changed to intermittent step-by-step implementation, and the intensity of metabolic regulation is appropriately reduced, thereby eliminating the execution conflict between the two intervention methods and forming a second intervention method without interference.

[0136] For the first pairing where there is no mutual interference, the corresponding first intervention method is directly determined as the second intervention method.

[0137] Step 606: Integrate the intervention methods, intensity, and linkage adaptation requirements of key influencing factors at all intervention levels to form an intervention strategy for vaginal microecological imbalance. The intervention methods include the first intervention method or the second intervention method.

[0138] In this step, the linkage and adaptation requirements refer to the collaborative operation requirements that must be followed when intervening in key influencing factors that have a direct relationship.

[0139] In this embodiment, the intervention levels, directions, first or second intervention methods, target intervention intensity, and corresponding linkage and adaptation requirements of all key influencing factors are first collected and integrated in the order of core intervention level first and secondary intervention level last to form a complete intervention strategy for vaginal microecological imbalance.

[0140] The embodiments of this application achieve multi-dimensional and coordinated precise intervention through intervention strategies, solving the problems of insufficient targeting and conflict in the implementation of traditional solutions, and improving the effectiveness and systematicness of vaginal microecological imbalance regulation.

[0141] Figure 3 This is a schematic diagram of a specific implementation of the vaginal microecological imbalance intervention strategy generation system based on multi-omics data provided in this application embodiment, with reference to... Figure 3 The system may include:

[0142] The acquisition module 31 is used to acquire multi-omics data of the vaginal microecology, including proteomic data, microbial community data, metabolomic data and transcriptomic data;

[0143] Processing module 32 is used to perform dimensionality reduction and data cleaning on the multi-omics data to generate a standardized dataset;

[0144] The association module 33 is used to perform feature association processing on the standardized dataset to generate a feature association matrix;

[0145] The determination module 34 is used to analyze the feature correlation matrix using causal reasoning technology to extract key influencing factors that lead to vaginal microecological imbalance and to determine the interaction relationships between the key influencing factors.

[0146] The parsing module 35 is used to perform collaborative parsing processing on the standardized dataset, the key influencing factors, and their interaction relationships to generate data parsing results;

[0147] The generation module is used to generate intervention strategies for vaginal microecological imbalance based on the data analysis results.

[0148] The vaginal microecological imbalance intervention strategy generation system based on multi-omics data in this application is used to implement the aforementioned vaginal microecological imbalance intervention strategy generation method based on multi-omics data. Therefore, the specific implementation of the vaginal microecological imbalance intervention strategy generation system based on multi-omics data can be found in the embodiment section of the vaginal microecological imbalance intervention strategy generation method based on multi-omics data above. The specific implementation can be referred to the description of the corresponding embodiment, and will not be repeated here.

[0149] This application also provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement the steps of the above-described method for generating intervention strategies for vaginal microecological imbalance based on multi-omics data.

[0150] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, implements the steps of the above-described method for generating intervention strategies for vaginal microecological imbalance based on multi-omics data.

[0151] In one exemplary embodiment, the computer storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0152] The embodiments of this application also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the embodiments of the method for generating intervention strategies for vaginal microecological imbalance based on multi-omics data.

[0153] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0154] The foregoing has provided a detailed description of the method and system for generating intervention strategies for vaginal microecological imbalance based on multi-omics data, as provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A method for generating intervention strategies for vaginal microecological imbalance based on multi-omics data, characterized in that, include: Obtain multi-omics data of vaginal microecology, including proteomic data, microbial community data, metabolomic data, and transcriptomic data; The multi-omics data are subjected to dimensionality reduction and data cleaning to generate a standardized dataset; The standardized dataset is subjected to feature association processing to generate a feature association matrix; The feature correlation matrix was analyzed using causal reasoning techniques to extract key influencing factors leading to vaginal microecological imbalance and to determine the interaction relationships among these key influencing factors. The standardized dataset, the key influencing factors, and their interactions are subjected to collaborative analysis to generate data analysis results. Based on the data analysis results, an intervention strategy for vaginal microecological imbalance is generated; The feature correlation matrix was analyzed using causal reasoning techniques to extract key influencing factors leading to vaginal microecological imbalance and to determine the interaction relationships among these key influencing factors, including: Select target first correlation degree values ​​and target second correlation degree values ​​that are higher than a preset threshold from the feature correlation matrix; The feature terms corresponding to the first correlation degree value and the second correlation degree value of each target are combined to obtain multiple feature term combinations; Causal reasoning techniques are used to analyze the relationships between combinations of features to identify the target features that cause changes. These target features are identified as key influencing factors leading to vaginal microecological imbalance, and the interaction relationships between these key influencing factors are analyzed. Based on the data analysis results, intervention strategies for vaginal microecological imbalance are generated, including: Based on the data analysis results and the characteristic performance of the corresponding omics data in the standardized dataset, the intervention priority of each key influencing factor is determined. Based on the trigger intensity values ​​of each key influencing factor and the degree of correlation between the interactions among the key influencing factors, the intervention level of the key influencing factors and the intervention direction of the key influencing factors at different intervention levels are determined. Based on the omics data characteristics corresponding to each key influencing factor, the first intervention method for key influencing factors at different intervention levels and in different intervention directions is determined. Based on the trigger intensity values ​​of each key influencing factor and the interaction relationship between each key influencing factor, the target intervention intensity corresponding to each first intervention method is set; For the first target key influencing factor that has a direct relationship with the key influencing factors, the first intervention method corresponding to the first target key influencing factor needs to be linked and adapted to obtain the second intervention method. The intervention methods, intensity, and linkage adaptation requirements of key influencing factors at all intervention levels are integrated to form an intervention strategy for vaginal microecological imbalance. The intervention methods include the first intervention method or the second intervention method. For the first key influencing factor that has a direct relationship with the key influencing factors, the first intervention method corresponding to the first key influencing factor needs to be adapted and linked to obtain the second intervention method, including: Pairing the key influencing factors of the primary objective that have a direct relationship with each other yields multiple pairings; Analyze whether the first intervention methods corresponding to the first key influencing factors of each pair of pairs interfere with each other during the execution process. For the first pairings that interfere with each other, determine the adjustment priority of the first intervention methods corresponding to the first pairings based on the degree of correlation of the direct action relationship corresponding to the first pairings and the trigger intensity value of each first key influencing factor. According to the adjustment priority and the omics data characteristics of the key influencing factors of the first target in the first pairing, the first intervention method corresponding to the first pairing with mutual interference is adjusted to obtain the second intervention method, and the first intervention method corresponding to the first pairing without mutual interference is taken as the second intervention method.

2. The method according to claim 1, characterized in that, Based on the trigger intensity values ​​of each key influencing factor and the interaction relationships between them, the target intervention intensity corresponding to each primary intervention method is set, including: Based on the trigger intensity value of each key influencing factor, the initial intervention intensity of the first intervention method corresponding to each key influencing factor is set; For the key influencing factors of the first objective that have a direct interaction relationship, the adjustment coefficients corresponding to each key influencing factor of the first objective are determined according to the degree of correlation of the interaction relationship between the key influencing factors of the first objective. Based on the initial intervention intensity and the adjustment coefficient, calculate the intermediate intervention intensity of the first intervention method corresponding to each key influencing factor of the first objective; For the key influencing factors of the first objective that have multiple direct relationships with the key influencing factors of the second objective, the intermediate intervention intensity corresponding to each key influencing factor of the second objective is adjusted according to the corresponding multiple adjustment coefficients to obtain the target intervention intensity corresponding to each first intervention method.

3. The method according to claim 1, characterized in that, The standardized dataset is subjected to feature association processing to generate a feature association matrix, including: Feature association analysis was performed on the standardized proteome data, standardized microbiome data, standardized metabolome data, and standardized transcriptome data in the standardized dataset to determine the first degree of association between each feature item. Cross-correlation analysis was performed on the standardized proteome data, standardized microbiome data, standardized metabolome data, and standardized transcriptome data to determine the second correlation value between the feature items of different standardized omics data. Based on the characteristic correspondence of each standardized omics data, the first correlation degree value and the second correlation degree value are integrated to form a feature correlation matrix.

4. The method according to claim 1, characterized in that, Causal reasoning techniques are used to analyze the relationships between combinations of features to identify target features that trigger changes. These target features are identified as key influencing factors leading to vaginal microecological imbalance. The interactions between these key influencing factors are then analyzed, including: Causal reasoning techniques are used to analyze the relationships between combinations of features to identify the first feature that triggers the change. Based on the changes of each first feature in the corresponding combination of features, a trigger strength value is assigned to each first feature. Select target feature items whose trigger intensity value is not lower than the preset intensity value from each first feature item as key influencing factors that lead to vaginal microecological imbalance; Based on the trigger intensity values ​​of each key influencing factor, determine the interaction relationship between the key influencing factors.

5. A system for generating intervention strategies for vaginal microecological imbalance based on multi-omics data, used in the method for generating intervention strategies for vaginal microecological imbalance based on multi-omics data as described in any one of claims 1 to 4, characterized in that, include: The acquisition module is used to acquire multi-omics data of the vaginal microecology, including proteomic data, microbial community data, metabolomic data, and transcriptomic data. The processing module is used to perform dimensionality reduction and data cleaning on the multi-omics data to generate a standardized dataset; The association module is used to perform feature association processing on the standardized dataset to generate a feature association matrix; The determination module is used to analyze the feature correlation matrix using causal reasoning techniques to extract key influencing factors that lead to vaginal microecological imbalance and to determine the interaction relationships between these key influencing factors. The parsing module is used to perform collaborative parsing processing on the standardized dataset, the key influencing factors, and their interactions to generate data parsing results; The generation module is used to generate intervention strategies for vaginal microecological imbalance based on the data analysis results.

6. A computing device, characterized in that, It includes a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are invoked and executed by the processing component to implement the method for generating intervention strategies for vaginal microecological imbalance based on multi-omics data as described in any one of claims 1 to 4.

7. A computer storage medium, characterized in that, The system contains a computer program that, when executed by a computer, implements the method for generating intervention strategies for vaginal microecological imbalance based on multi-omics data as described in any one of claims 1 to 4.

Images