Biological Experiment Safety Risk Assessment and Early Warning System

Abstract

The present application provides a biological experiment safety risk assessment and early warning system, which involves the field of data processing technology, including: a biological attribute feature detection module detects and obtains the biological attribute features of unknown organisms; a first associated organism determination module inputs the biological attribute features into a concentrated knowledge base for similarity traversal comparison, and determines multiple first associated organisms; a high-frequency associated organism determination module calls a comprehensive knowledge base to determine multiple high-frequency associated organisms; an unknown risk coefficient output module retrieves and screens multiple key risk feature sets based on the biological Internet, and outputs unknown risk coefficients; a potential risk level assessment module obtains laboratory protection features, retrieves and analyzes sample experimental event sets, determines potential risk levels, and generates risk warning signals. The present application solves the technical problems of the prior art that the comprehensiveness, accuracy, and reliability of the assessment are poor due to insufficient historical data, and achieves more accurate and comprehensive biological risk assessment and early warning.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and specifically to a biological experiment safety risk assessment and early warning system. Background Art

[0002] Risk assessment and early warning in biological experiments are important links to ensure laboratory safety. Traditional biological experiment risk assessment systems usually rely on known biological characteristics and historical data for analysis. These systems mainly detect some of the biological properties, such as amino acid sequences or nucleotide sequences, and compare them with existing knowledge bases to identify possible related organisms and assess related risks. Such systems are effective in dealing with common biological hazards and can provide certain risk assessments and early warnings. However, for emerging or rare biological hazards, the existing database lacks sufficient historical data to support accurate risk assessment. The scarcity of such data directly affects the comprehensiveness, accuracy and reliability of risk assessment, making the assessment results insufficient as an effective basis for decision-making. Especially in the context of the rapid development of biotechnology today, the continuous emergence of new biological hazards has put higher demands on the existing assessment system. Due to the lack of sufficient information, laboratories may not be able to take effective preventive measures when handling these unknown biological materials, thereby increasing the risks to laboratory workers and the environment. Summary of the invention

[0003] This application provides a biological experiment safety risk assessment and early warning system, which solves the technical problems of existing assessment methods in the face of unknown or rare organisms, such as the lack of sufficient historical data, resulting in poor comprehensiveness, accuracy and reliability of risk assessment, thereby achieving more accurate and comprehensive biological risk assessment and early warning, and improving the safety of laboratory operations.

[0004] In view of the above problems, the present application provides a biological experiment safety risk assessment and early warning system, the system comprising: a biological attribute feature detection module, the biological attribute feature detection module is used to detect and obtain biological attribute features of unknown organisms, wherein the biological attribute features at least include amino acid sequences, nucleotide sequences and metabolic features; a first associated organism determination module, the first associated organism determination module is used to input the amino acid sequence, nucleotide sequence and metabolic features into a concentrated knowledge base, perform similarity traversal comparison according to a predetermined comparison strategy, select concentrated organisms that meet a first similarity as first associated organisms, and determine multiple first associated organisms; a high-frequency associated organism determination module, the high-frequency associated organism determination module is used to call a comprehensive knowledge base based on the multiple first associated organisms, map and obtain multiple associated organism sets, select associated organisms that meet a second similarity Set it as a high-frequency associated organism, determine multiple high-frequency associated organisms, wherein the second similarity is greater than the first similarity, and the high-frequency associated organisms have an association mark; an unknown risk coefficient output module, the unknown risk coefficient output module is used to use the multiple high-frequency associated organisms as constraints, obtain multiple sample risk feature sets based on biological Internet retrieval, screen and determine multiple key risk feature sets, perform comprehensive risk assessment based on multiple associations and multiple key risk feature sets, and output an unknown risk coefficient; a potential risk level assessment module, the potential risk level assessment module is used to obtain laboratory protection features, use the unknown risk coefficient and the laboratory protection features as constraints, retrieve and obtain a sample experimental event set, determine the potential risk level based on the sample experimental event set analysis, and generate a risk warning signal if the potential risk level is greater than a predetermined risk threshold.

[0005] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0006] The biological attribute feature detection module detects and obtains the biological attribute features of unknown organisms, provides basic data for subsequent comparison and risk assessment, and ensures the comprehensiveness of the assessment. The first associated organism determination module inputs the biological attribute features into the concentrated knowledge base, performs similar traversal comparison according to the predetermined comparison strategy, selects concentrated organisms that meet the first similarity as the first associated organism, determines multiple first associated organisms, and quickly identifies known organisms with similar features to unknown organisms, providing a basis for subsequent high-frequency associated organism determination. The high-frequency associated organism determination module calls the comprehensive knowledge base based on the multiple first associated organisms, maps and obtains multiple sets of associated organisms, selects associated organisms that meet the second similarity as high-frequency associated organisms, determines multiple high-frequency associated organisms, and compares a wider range of knowledge bases to improve the accuracy and comprehensiveness of the comparison, screen out organisms that are highly similar to the unknown biological features, and provide a more reliable basis for risk assessment. The unknown risk coefficient output module uses the multiple high-frequency associated organisms as constraints, obtains multiple sample risk feature sets based on biological Internet retrieval, screens and determines multiple key risk feature sets, conducts comprehensive risk assessment based on multiple correlations and multiple key risk feature sets, outputs unknown risk coefficients, comprehensively assesses the potential risks of unknown organisms, and provides a quantitative risk coefficient to ensure the accuracy and comprehensiveness of the assessment results. The potential risk level assessment module obtains laboratory protection features, uses the unknown risk coefficients and the laboratory protection features as constraints, retrieves and obtains sample experimental event sets, and determines the potential risk level based on the analysis of the sample experimental event set. If the potential risk level is greater than the predetermined risk threshold, a risk warning signal is generated to ensure that the laboratory can take timely measures to prevent potential hazards.

[0007] In summary, this application integrates multiple modules such as biological attribute feature detection, related biological determination, risk coefficient output and potential risk level assessment to build a more efficient and intelligent biological experiment safety risk assessment and early warning system. This system uses comprehensive feature detection, multi-level knowledge base comparison and dynamic Internet retrieval, and combines laboratory protection features. Even in the absence of sufficient historical data, it can conduct detailed and accurate risk assessments on unknown organisms and provide dynamic risk warnings in a timely manner, significantly improving the comprehensiveness, accuracy and reliability of biological laboratory risk assessments and early warnings, and ensuring the safety of laboratory operations.

[0008] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1A schematic diagram of the structure of the biological experiment safety risk assessment and early warning system provided in the embodiment of the present application;

[0010] Figure 2 A schematic diagram of a process for determining multiple differential biological attribute features in a biological experiment safety risk assessment and early warning system provided in an embodiment of the present application;

[0011] Figure 3 A schematic diagram of the process of screening key features of the first sample risk feature set in the biological experiment safety risk assessment and early warning system provided in an embodiment of the present application.

[0012] Explanation of reference numerals: biological attribute feature detection module 10 , first associated biological determination module 20 , high-frequency associated biological determination module 30 , unknown risk coefficient output module 40 , potential risk level assessment module 50 . DETAILED DESCRIPTION

[0013] The embodiments of the present application provide a biological experiment safety risk assessment and early warning system to solve the technical problems of existing assessment methods in the face of unknown or rare organisms, such as the lack of sufficient historical data, which leads to poor comprehensiveness, accuracy and reliability of risk assessment, thereby achieving more accurate and comprehensive biological risk assessment and early warning, and improving the safety of laboratory operations.

[0014] like Figure 1 As shown, the embodiment of the present application provides a biological experiment safety risk assessment and early warning system, the system comprising:

[0015] The biological attribute feature detection module 10 is used to detect and obtain the biological attribute features of an unknown organism, wherein the biological attribute features at least include amino acid sequences, nucleotide sequences and metabolic features.

[0016] Specifically, biological attribute characteristics refer to the inherent characteristics of organisms, which are used to describe and distinguish different organisms, including but not limited to amino acid sequences, nucleotide sequences and metabolic characteristics. Amino acid sequence is the basic building block of protein, which is a chain molecule formed by connecting multiple amino acids in a specific order. The amino acid sequence determines the structure and function of the protein. The nucleotide sequence is the basic building block of DNA or RNA, which is a chain molecule formed by nucleotides arranged in a specific order. The nucleotide sequence determines the storage and transmission of genetic information. Metabolic characteristics are the overall characteristics of a series of chemical reactions in organisms, which are used to describe the metabolic activities of organisms, including metabolic pathways, metabolites, etc., and can reflect the physiological state and environmental conditions of organisms.

[0017] The biological attribute feature detection module 10 first extracts DNA, RNA and protein samples from the unknown organism. This process can be completed by laboratory cultivation, molecular biology technology or on-site sampling. For example, DNA samples for nucleotide sequence analysis can be obtained through cell lysis, purification and other steps.

[0018] Then, through a series of advanced experimental techniques, the key features of unknown organisms can be fully detected. High-throughput sequencing technology can be used to determine the nucleotide sequence of DNA or RNA samples; mass spectrometry analysis can be used to determine the amino acid sequence of protein samples; and chromatography, mass spectrometry or biosensors can be used to analyze metabolites in organisms. For example, Illumina sequencing can quickly and efficiently generate a large amount of sequence data, and fully understand the genome information of organisms. Liquid chromatography and mass spectrometry can accurately determine the amino acid sequence of proteins and understand the structure and function of proteins. Gas chromatography and mass spectrometry can quantitatively analyze the types and concentrations of metabolites and understand the metabolic pathways and activities of organisms.

[0019] The biological attribute feature detection module 10 obtains comprehensive and detailed biological attribute feature information through the above process, providing a solid data foundation for subsequent comparison and risk assessment.

[0020] The first associated organism determination module 20 is used to input the amino acid sequence, nucleotide sequence and metabolic characteristics into a concentrated knowledge base, perform similarity traversal comparison according to a predetermined comparison strategy, select concentrated organisms that meet the first similarity as first associated organisms, and determine multiple first associated organisms.

[0021] Specifically, the condensed knowledge base is a biological information database that has been strictly screened and optimized, storing a large amount of feature data of known organisms for rapid comparison and analysis. The predetermined comparison strategy is a strategy for comparing biological features based on specific algorithms or rules to determine similarity. Similarity is an indicator that measures the degree of similarity between two biological features. Similarity traversal comparison is a process of globally comparing input features based on similarity to find the most similar biological features. The first similarity is a preset threshold used to determine the minimum requirement for the degree of similarity between unknown biological data and data in the knowledge base. The first associated organism refers to a known organism with similar features to the unknown organism that is screened out based on similarity during the preliminary comparison process.

[0022] The first associated organism determination module 20 uses bioinformatics tools and a concentrated knowledge base to compare data, performs similarity traversal comparison according to a predetermined comparison strategy, and selects organisms that meet the first similarity as the first associated organisms, thereby providing an important basis for subsequent risk assessment. This process ensures the accuracy and comprehensiveness of the initial screening and improves the system's ability to identify unknown organisms.

[0023] The specific process of determining multiple first associated organisms is as follows: First, the amino acid sequence, nucleotide sequence and metabolic characteristics of the unknown organism obtained by the biological attribute feature detection module 10 are input into the condensed knowledge base. Use bioinformatics tools to automatically search for known organisms similar to the input unknown biological attribute characteristics in the condensed knowledge base according to a predetermined comparison strategy. For example, use BLAST for sequence alignment; use Bioconductor or Galaxy for biological data analysis, and KEGG or Reactome database for metabolic pathway analysis, etc. After the alignment is completed, calculate the similarity between each candidate known organism and the unknown organism. These similarity values ​​are evaluated according to a preset similarity threshold to determine which organisms can be identified as first associated organisms. If the similarity of multiple known organisms meets the conditions, the module will determine all of these organisms as first associated organisms. This helps to expand the reference range of risk assessment, especially when the unknown organism has a high similarity with multiple known organisms.

[0024] Exemplarily, BLAST is used for sequence alignment, and the amino acid sequence of the unknown protein is input into the BLAST database as the query sequence. BLAST will return a list of known protein sequences and give similarity scores between these sequences and the query sequence. By setting a threshold, i.e., the first similarity, known proteins with statistically significant similarity to the unknown protein can be screened out, and these proteins can be identified as the first associated organisms. Similarly, for nucleotide sequences and metabolic features, corresponding bioinformatics tools can be used for alignment and similarity assessment.

[0025] A high-frequency associated organism determination module 30 is used to call a comprehensive knowledge base based on the multiple first associated organisms, map and obtain multiple sets of associated organisms, select associated organisms that meet the second similarity as high-frequency associated organisms, and determine multiple high-frequency associated organisms, wherein the second similarity is greater than the first similarity, and the high-frequency associated organisms have an association mark.

[0026] Specifically, the comprehensive knowledge base is a database containing extensive biological information, covering more types and quantities of biometric data for in-depth comparison and analysis. The associated biological set is a set of known organisms with a certain degree of similarity to the unknown biological characteristics determined through the comparison process. The second similarity is a similarity standard higher than the first similarity and is used for stricter screening. High-frequency associated organisms refer to known organisms that are highly similar to the unknown biological characteristics and have a higher degree of association, which are screened out through the second similarity. The association mark is a mark that indicates the degree of similarity between the high-frequency associated organism and the unknown biological characteristics.

[0027] First, the high-frequency associated organism determination module 30 will call the information of multiple first associated organisms determined previously. This information comes from preliminary bioinformatics analysis, in which the amino acid sequence, nucleotide sequence and metabolic characteristics of the unknown organism are compared with the known organisms in the concentrated knowledge base. Use bioinformatics tools to compare the first associated organisms, map and obtain multiple associated organism sets in the comprehensive knowledge base. These sets contain known organisms that may be related to unknown organisms. Then the module evaluates the similarity between each organism in the associated organism set and the unknown organism according to the preset second similarity threshold. Only when the similarity meets or exceeds the second similarity threshold will the related organism be selected as a high-frequency associated organism. The selected high-frequency associated organisms will be given an association mark, which reflects the strength of association between these organisms and the unknown organism. Exemplarily, it is assumed that the first associated organism is a certain bacterium A, and its amino acid sequence similarity is 92%. Call the comprehensive knowledge base, use BLAST to compare the sequence of bacterium A, find another bacterium B, with a sequence similarity of 96%, and select bacterium B as a high-frequency associated organism.

[0028] Finally, the module outputs multiple high-frequency associated organisms and their association markers, which can be used for further risk assessment, classification research or other bioinformatics analysis. The above process further ensures the accuracy and comprehensiveness of the comparison results, improves the system's ability to identify unknown organisms and the reliability of risk assessment.

[0029] The unknown risk coefficient output module 40 is used to obtain multiple sample risk feature sets based on biological Internet retrieval with the multiple high-frequency associated organisms as constraints, screen and determine multiple key risk feature sets, conduct a comprehensive risk assessment based on multiple correlation degrees and multiple key risk feature sets, and output an unknown risk coefficient.

[0030] Specifically, the biological Internet is a metaphorical concept that refers to a virtual network formed by the interconnection of bioinformatics tools and databases, which allows the sharing, retrieval and analysis of biological data. The sample risk feature set refers to a series of risk-related features extracted from a specific biological sample. The key risk feature set is a subset screened from multiple sample risk feature sets, and these features play a decisive role in assessing the risk of biological samples. The correlation degree is an indicator to measure the degree of similarity between the characteristics of high-frequency associated organisms and unknown organisms. In bioinformatics, the correlation degree usually refers to the degree of genetic, biochemical or ecological similarity or connection between two or more biological entities. The unknown risk coefficient is a quantitative indicator used to assess the potential risk level of unknown organisms, which is calculated based on the comprehensive calculation of high-frequency associated organisms and their characteristics.

[0031] The unknown risk coefficient output module 40 first obtains multiple sample risk feature sets related to high-frequency associated organisms based on the previously determined multiple high-frequency associated organisms through the biological Internet search function. The retrieval process can be completed using search engines, scientific databases such as NCBI, PubMed, etc., and bioinformatics tools such as BLAST, HMMER, etc. to obtain sample risk feature data related to high-frequency associated organisms. For example, PubMed is used to search for research papers and risk assessment reports related to specific high-frequency associated organisms. Next, the retrieved relevant data is sorted and summarized to form a sample risk feature set. These feature sets may include information such as toxicity, pathogenicity, and transmission pathways of known organisms. Then, the retrieved sample risk feature set is analyzed, and a feature subset that is critical to the risk assessment of unknown organisms is screened out through high-frequency clustering. This process can identify key features through statistical analysis and machine learning algorithms. For example, the risk features with the greatest impact and the most common are screened out through R language, Python's pandas library, etc. Finally, a comprehensive risk assessment is performed based on multiple correlations of high-frequency associated organisms and the screened key risk feature set. Use mathematical models or machine learning algorithms, such as regression analysis, decision trees, etc. to evaluate, calculate and output unknown risk coefficients. These coefficients can be used to guide biosafety management and the formulation of risk prevention measures. For example, the final risk coefficient can be obtained by combining multiple correlations and key risk features through the weighted average method.

[0032] The unknown risk coefficient output module 40 can output the risk coefficient of unknown organisms based on the information of high-frequency related organisms by searching the biological Internet and comprehensive risk assessment, thereby ensuring the comprehensiveness and accuracy of the assessment results and improving the ability to identify and respond to potential risks of unknown organisms. At the same time, it provides specific and operational data support for laboratory safety management.

[0033] A potential risk level assessment module 50 is used to obtain laboratory protection characteristics, retrieve and obtain a sample experimental event set based on the unknown risk coefficient and the laboratory protection characteristics as constraints, and determine the potential risk level based on the analysis of the sample experimental event set. If the potential risk level is greater than a predetermined risk threshold, a risk warning signal is generated.

[0034] Specifically, laboratory protection characteristics refer to the various safety measures and capabilities that the laboratory has to protect against potential risks, including physical protection equipment, protection levels, emergency response capabilities, etc. The sample experimental event set is a data set of relevant experimental cases obtained from past experimental event records and databases, which is used to analyze and evaluate risk levels. The potential risk level is an indicator of the potential degree of danger of a biological experiment based on a comprehensive evaluation. The predetermined risk threshold is a standard value used to determine whether a warning needs to be issued for the potential risk level. The risk warning signal is an alarm issued when the risk level assessment reaches or exceeds the preset threshold, which is used to prompt laboratory personnel to take additional safety measures or suspend experimental operations.

[0035] The potential risk level assessment module 50 first collects the laboratory's protection feature information, which may come from the laboratory's safety manual, on-site inspection or pre-experimental safety assessment. Then, using the unknown risk coefficient and the laboratory protection feature as constraints, the relevant sample experimental event set is retrieved through the biological Internet. These event sets may include data on similar experiments in history, experimental data on known pathogens, etc. For example, for a laboratory with a biosafety level of 2, the module will retrieve related experimental events, such as infection events and leakage events that occurred in the past. Then, the module uses statistical analysis tools, such as R language, Python's pandas library, etc. to perform data analysis, and analyzes the retrieved sample experimental event set through statistical analysis methods, such as risk matrix, regression analysis or Monte Carlo simulation, and evaluates the potential risk level of the experimental operation in combination with the laboratory protection features. For example, using the regression analysis method, the effect of a specific protection level in dealing with similar risks is evaluated, thereby deriving the potential risk level. The calculated potential risk level is compared with a predetermined risk threshold. If the potential risk level determined by the analysis exceeds the predetermined risk threshold, the module will automatically generate a risk warning signal. These signals can notify laboratory personnel through a laboratory information management system or a mobile device.

[0036] The potential risk level assessment module 50 can assess the potential risks of unknown organisms in the laboratory environment in combination with the specific conditions of the laboratory, and generate risk warning signals when necessary, thereby ensuring the safety of laboratory staff and the environment.

[0037] Furthermore, the system described in the embodiment of the present application is also used to construct a comprehensive knowledge base and a condensed knowledge base, including:

[0038] An initial knowledge base is constructed based on biological big data, wherein the initial knowledge base includes multiple known biological attribute features; a differential analysis is performed on the multiple known biological attribute features to determine multiple differential biological attribute features, and a concentrated knowledge base is constructed based on the multiple differential biological attribute features; a similarity clustering is performed on the multiple known biological attribute features according to the multiple differential biological attribute features to determine multiple mapped biological attribute feature sets, wherein the differential biological attribute features and the mapped biological attribute feature sets have a corresponding relationship; based on the corresponding relationship, a comprehensive knowledge base is constructed according to the multiple differential biological attribute features and the multiple mapped biological attribute feature sets.

[0039] Specifically, biological big data refers to a large amount of data generated in biological research, including genomic data, proteomic data, metabolomic data, etc. The initial knowledge base is a database containing a large number of known biological attribute features based on biological big data. Differential biological attribute features are features that are significantly different between organisms determined through differential analysis. The mapped biological attribute feature set is a set of biological attribute features with specific relationships formed through similarity clustering.

[0040] The system improves the accuracy of biological attribute feature comparison and risk assessment by building a comprehensive knowledge base and a condensed knowledge base, using biological big data and differentiated analysis methods. The following are the specific implementation steps: First, the system uses data collection tools and bioinformatics databases based on biological big data to collect and organize a large number of known biological attribute features, including gene sequences, protein structures, biological activities, etc., to form an initial knowledge base containing a wide range of biological information.

[0041] Next, statistical methods and data mining tools are used to perform differential analysis on multiple known biological attribute features in the initial knowledge base to identify features that can reflect differences between organisms. These features will be used to build a concentrated knowledge base that focuses on the key differences between organisms to facilitate more accurate biological classification and comparative analysis. For example, t-tests or ANOVA are used to analyze the amino acid sequences and nucleotide sequences of different organisms to find significantly different features.

[0042] Then, based on the differential biological attribute features, the system similarly clusters multiple known biological attribute features and determines multiple sets of mapped biological attribute features, which can help reveal the relationship or functional similarity between organisms. Cluster analysis can be done using machine learning algorithms, such as K-means clustering, hierarchical clustering, etc. For example, the K-means algorithm is used to cluster biological attribute features with similar amino acid sequences into several groups. Based on the correspondence between differential biological attribute features and mapped biological attribute feature sets, a detailed and comprehensive knowledge base is constructed. This knowledge base contains not only the information of a single organism, but also the relationships and networks between organisms, supporting in-depth and comprehensive comparison and analysis.

[0043] Exemplarily, when constructing the initial knowledge base, the BioPython library of Python is used to extract data from the GenBank database to collect amino acid sequence and nucleotide sequence data of thousands of known bacteria. The statistical analysis function of the R language is used to perform a t-test on these data to determine the significantly different amino acid sequences between different bacteria as differential biological attribute features. Based on the differential biological attribute features, the K-means clustering algorithm is used to cluster these features to form several mapping biological attribute feature sets. The differential biological attribute features are matched with the mapping biological attribute feature sets to form a comprehensive database containing detailed and extensive biological attribute features. The SQL database management system is used to store and manage these data.

[0044] By systematically building a comprehensive knowledge base and a condensed knowledge base, using biological big data, differential analysis and similarity clustering methods, the accuracy and efficiency of biological attribute feature comparison and risk assessment can be improved. This process ensures the comprehensiveness and simplicity of the database, supporting more accurate and efficient bioinformatics analysis and risk assessment.

[0045] Preferably, Figure 2 As shown, the determination of multiple difference biological attribute features described in the embodiment of the present application also includes:

[0046] A first attribute feature is randomly selected from the multiple known biological attribute features, wherein the first attribute feature includes a first amino acid sequence, a first nucleotide sequence and a first metabolic feature; a similarity comparison is performed between the first amino acid sequence and other amino acid sequences in the multiple known biological attribute features, and the frequency of the similarity being less than a predetermined amino acid similarity threshold is counted and set as a first difference; if the first difference is greater than the predetermined threshold, a similarity comparison is performed between the first nucleotide sequence and other nucleotide sequences in the multiple known biological attribute features, and the frequency of the similarity being less than a predetermined nucleotide similarity threshold is counted and set as a second difference; if the second difference is greater than the predetermined threshold, a similarity comparison is performed between the first metabolic feature and other metabolic features in the multiple known biological attribute features to obtain a third difference; if the third difference is greater than the predetermined threshold, the first attribute feature is set as a differential biological attribute feature to obtain multiple differential biological attribute features.

[0047] Specifically, the first attribute feature is a set of features randomly selected from known biological attribute features, including a first amino acid sequence, a first nucleotide sequence, and a first metabolic feature. The predetermined threshold is a set standard value of similarity or difference, which is used to screen differential biological attribute features. The predetermined amino acid similarity threshold is a preset amino acid sequence similarity threshold, which is used to determine whether two amino acid sequences are similar. The predetermined nucleotide similarity threshold is a preset nucleotide sequence similarity threshold, which is used to determine whether two nucleotide sequences are similar.

[0048] The difference is a measure of the similarity between a feature and other features. The higher the difference, the lower the similarity between the feature and other features. The first difference is the frequency of the statistical similarity less than the predetermined amino acid similarity threshold after the first amino acid sequence is compared with other amino acid sequences. The second difference is the frequency of the statistical similarity less than the predetermined nucleotide similarity threshold after the first nucleotide sequence is compared with other nucleotide sequences. The third difference is the frequency of the statistical similarity less than the predetermined metabolic feature similarity threshold after the first metabolic feature is compared with other metabolic features.

[0049] First, the system randomly selects a feature from multiple known biological attribute features as the first attribute feature. The first attribute feature includes the first amino acid sequence, the first nucleotide sequence, and the first metabolic feature. Use a comparison tool (such as BLAST) to perform a similarity comparison between the first amino acid sequence and the amino acid sequences in other known biological attribute features. The frequency of similarity less than the predetermined amino acid similarity threshold is counted, and this frequency is set as the first difference. If the first difference is greater than the predetermined threshold, proceed to the next step; otherwise, end the analysis of the current feature and randomly select a new first attribute feature.

[0050] Use the comparison tool to compare the first nucleotide sequence with the nucleotide sequences in other known biological attribute features. Count the frequency of similarity less than the predetermined nucleotide similarity threshold, and set this frequency as the second difference. If the second difference is greater than the predetermined threshold, proceed to the next step; otherwise, end the analysis of the current feature and randomly select a new first attribute feature.

[0051] Use a comparison tool (such as MetaboAnalyst) to perform a similarity comparison between the first metabolic feature and the metabolic features in other known biological attribute features, obtain the frequency of similarity less than the predetermined metabolic feature similarity threshold, and set this frequency as the third difference. If the third difference is greater than the predetermined threshold, the first attribute feature is set as the differential biological attribute feature and recorded; otherwise, end the analysis of the current feature and randomly select a new first attribute feature. Repeat the above steps until multiple differential biological attribute features are determined.

[0052] Exemplarily, the predetermined threshold is set to 25, and an amino acid sequence of a bacterium is randomly selected from the initial knowledge base. Through BLAST comparison, it is found that the frequency of the sequence being less than 80% similar to other sequences is 50 times, then the first difference is 50>25, and the nucleotide sequence is continued to be compared. The nucleotide sequence of the bacterium is compared by BLASTN, and the frequency of the similarity being less than 85% is found to be 40 times, then the second difference is 40>25, and the metabolic characteristics are continued to be compared. MetaboAnalyst comparison is performed on the metabolic characteristics of the bacterium, and the frequency of the similarity being less than 70% is found to be 30 times, then the third difference is 30>25. All attribute features of the bacterium are set as differential biological attribute features.

[0053] The above process identified multiple differential biological attribute features through random selection and step-by-step similarity comparison, ensuring that the screened features are significantly different at all levels, which is helpful for subsequent knowledge base construction and bioinformatics analysis, and improves the system's comparison and risk assessment capabilities.

[0054] Furthermore, the system described in the embodiment of the present application is also used to configure a predetermined comparison strategy, including:

[0055] Configure multiple similarity comparison operators, use the amino acid sequence and the multiple similarity comparison operators as constraints, retrieve sample comparison data, analyze and determine multiple first comparison accuracies, and configure a first weight ratio based on the multiple first comparison accuracies; determine the second weight ratio of the nucleotide sequence and the third weight ratio of the metabolic feature in turn; and construct the predetermined comparison strategy based on the multiple similarity comparison operators, the first weight ratio, the second weight ratio, and the third weight ratio.

[0056] Specifically, similarity comparison operators refer to algorithms or tools used to measure the similarity between biological attribute features. For example, Needleman-Wunsch algorithm, Smith-Waterman algorithm, ClustalW, BLAST, etc. Sample comparison data is data obtained after comparison based on similarity comparison operators, which is used to analyze and determine the comparison accuracy. Comparison accuracy is an indicator to measure the accuracy of the comparison results, usually expressed as a percentage or score. The first weight ratio, the second weight ratio and the third weight ratio refer to the proportion of the comparison operators corresponding to amino acid sequences, nucleotide sequences and metabolic characteristics in the overall score in the comprehensive evaluation, respectively, reflecting the relative importance of each biological attribute feature.

[0057] First, the system selects and configures multiple similarity comparison operators for comparing biological attribute features. These operators will be used for different types of biological sequence comparisons. For example, BLAST is selected for amino acid sequence comparison, Needleman-Wunsch is used for nucleotide sequence comparison, and MetaboAnalyst is used for metabolic feature comparison.

[0058] Next, the sample alignment is performed with the amino acid sequence and multiple similarity alignment operators as constraints to obtain the alignment result data. The retrieved sample alignment data is analyzed to determine the alignment accuracy of each similarity alignment operator, that is, the reliability and accuracy of the alignment result. This process can be achieved by calculating the alignment accuracy of each alignment operator in different samples. For example, the system can perform BLAST alignment on the amino acid sequences of multiple samples and then calculate the alignment accuracy of each sample.

[0059] Then, based on the multiple first comparison accuracies, the first weight ratio of the amino acid sequence comparison is configured. This weight ratio reflects the importance of the amino acid sequence in the overall comparison strategy. In the same way, the nucleotide sequence and metabolic characteristics are compared in turn, the comparison accuracy is analyzed, and the second and third weight ratios are configured.

[0060] Finally, based on multiple similarity comparison operators, the first weight ratio, the second weight ratio and the third weight ratio, a predetermined comparison strategy is comprehensively constructed. This strategy contains multiple comparison operators and their weight ratios to guide the subsequent similarity comparison process. When implementing the predetermined comparison strategy, bioinformatics software and databases can be used to store and retrieve biological sequence data, and statistical analysis tools can be used to calculate the comparison accuracy and weight ratio.

[0061] Exemplarily, there are three samples, which are respectively recorded as sample A, sample B and sample C. BLAST is selected for amino acid sequence alignment, Needleman-Wunsch is used for nucleotide sequence alignment, and MetaboAnalyst is used for metabolic feature alignment.

[0062] Calculate multiple first comparison accuracies: for sample A, the amino acid sequence comparison accuracy is 85%; for sample B, the amino acid sequence comparison accuracy is 75%; for sample C, the amino acid sequence comparison accuracy is 90%. Calculate the amino acid sequence weight ratio, that is, the first weight ratio: .

[0063] Calculate multiple second comparison accuracies: for sample A, the nucleotide sequence comparison accuracy is 70%; for sample B, the nucleotide sequence comparison accuracy is 60%; for sample C, the nucleotide sequence comparison accuracy is 80%. Calculate the nucleotide sequence weight ratio, that is, the second weight ratio:

[0064] .

[0065] Calculate multiple third comparison accuracies: for sample A, the metabolic feature comparison accuracy is 65%; for sample B, the metabolic feature comparison accuracy is 75%; for sample C, the metabolic feature comparison accuracy is 60%. Calculate the metabolic feature weight ratio, that is, the third weight ratio:

[0066] .

[0067] Finally, the weights are normalized so that their sum is 1, and the proportion of each weight in the total weight is calculated to obtain the normalized weight: ; ; Based on multiple similarity comparison operators, the first weight ratio, the second weight ratio and the third weight ratio, a predetermined comparison strategy is comprehensively constructed.

[0068] The above process improves the accuracy and reliability of the comparison by performing precision analysis and weight configuration through comparison tools and data analysis tools.

[0069] Furthermore, the unknown risk coefficient output module 40 of the embodiment of the present application determines a plurality of key risk feature sets, further comprising:

[0070] Taking the multiple high-frequency associated organisms as constraints, multiple sample risk feature sets are obtained based on biological Internet retrieval, wherein the sample risk features include sample risk probability, sample transmission speed and sample severity; a first sample risk feature set is randomly selected, key feature screening is performed on the first sample risk feature set, a first key risk feature set is determined, and added to the multiple key risk feature sets.

[0071] Specifically, the sample risk features include sample risk probability, sample propagation speed and sample severity, where the sample risk probability refers to the possibility that a certain biological feature will cause a risk event to occur. The sample propagation speed is the speed at which a certain biological feature propagates, usually expressed as a propagation rate or doubling time. The sample severity refers to the severity of a risk event caused by a certain biological feature, usually expressed as an impact range or impact depth. The first sample risk feature set is a feature set randomly selected from the acquired sample risk feature set for key feature screening. The first key risk feature set is the key risk features screened out from the first sample risk feature set and added to the final multiple key risk feature sets.

[0072] The unknown risk coefficient output module 40 uses high-frequency associated organisms as search conditions to obtain sample risk feature data through the biological Internet. The retrieved sample risk features include sample risk probability, sample transmission speed and sample severity, and the retrieved risk feature data are sorted and summarized to form multiple sample risk feature sets. These feature sets may come from existing bioinformatics databases, such as Ensembl, UniProt, etc., or special biosafety databases.

[0073] Next, the module randomly selects a feature set from the acquired sample risk feature set as the first sample risk feature set. Statistical analysis tools such as Python's pandas library and R language use statistical analysis methods, such as regression analysis or principal component analysis, to analyze the first sample risk feature set and screen out the features most relevant to the risk assessment of unknown organisms. The screened key features are summarized to form the first key risk feature set. Then, the first key risk feature set is added to the final multiple key risk feature sets, and the above steps are repeated until all key risk feature sets are determined.

[0074] This module is based on high-frequency associated organisms, uses biological Internet retrieval to obtain sample risk characteristic data, and uses statistical analysis tools for data analysis and screening, effectively identifying multiple key risk characteristic sets, ensuring the accuracy and relevance of key characteristics, thereby improving the comprehensiveness and reliability of unknown biological risk assessment.

[0075] Preferably, Figure 3 As shown, the unknown risk coefficient output module 40 of the embodiment of the present application performs key feature screening on the first sample risk feature set, and further includes:

[0076] A risk feature is randomly selected from the first sample risk feature set as the first risk feature, wherein the first risk feature includes a first risk probability, a first propagation speed and a first severity; a deviation calculation is performed on the first risk probability and other risk probabilities in the first sample risk feature set, and the number of times the statistical difference is less than a predetermined probability deviation threshold is set as a first criticality; if the first criticality is greater than the predetermined critical threshold, a deviation calculation is performed on the first propagation speed and other propagation speeds in the first sample risk feature set to obtain a second criticality; if the second criticality is greater than the predetermined critical threshold, a deviation calculation is performed on the first severity and other severities in the first sample risk feature set to obtain a third criticality; if the third criticality is greater than the predetermined critical threshold, the first risk feature is set as the first key risk feature and added to the first key risk feature set.

[0077] Specifically, the unknown risk coefficient output module 40 randomly selects a risk feature from the first sample risk feature set as the first risk feature. The first risk feature includes a first risk probability, a first propagation speed, and a first severity.

[0078] Deviation calculation is performed on the first risk probability and other risk probabilities in the first sample risk feature set. The difference between the two risk probabilities is calculated, and the number of times the difference is less than the predetermined probability deviation threshold is counted, which is set as the first criticality. For example, the predetermined probability deviation threshold is 0.1, and the number of times the difference is less than 0.1 is calculated to be 30 times, then the first criticality is 30. Among them, the criticality is a measure of the significance of a feature as a key feature in the feature set. The higher the criticality, the more important the feature is in the feature set. The predetermined probability deviation threshold is a standard value used to determine whether the risk probability deviation is significant. The predetermined key threshold is a standard value used to determine whether a feature is a key feature. If the first criticality is greater than the predetermined key threshold, proceed to the next step; otherwise, end the analysis of the current feature and randomly select a new first risk feature.

[0079] Deviation calculation is performed on the first propagation speed and other propagation speeds in the first sample risk feature set. The difference between the two propagation speeds is calculated, and the number of times the difference is less than the predetermined propagation speed deviation threshold is counted, which is set as the second criticality. If the second criticality is greater than the predetermined criticality threshold, the next step is continued; otherwise, the analysis of the current feature is terminated, and a new first risk feature is randomly selected again.

[0080] Deviation calculation is performed on the first severity and other severities in the first sample risk feature set. The difference between the two severities is calculated, and the number of times the difference is less than the predetermined severity deviation threshold is counted, and set as the third criticality. If the third criticality is greater than the predetermined critical threshold, the first risk feature is set as the first critical risk feature and added to the first critical risk feature set; otherwise, the analysis of the current feature is terminated, and a new first risk feature is randomly selected again.

[0081] Exemplarily, the predetermined probability deviation threshold is set to 0.1, the predetermined propagation speed deviation threshold is set to 0.05, the predetermined severity deviation threshold is set to 0.1, and the predetermined critical threshold is set to 20. The risk probability of the first risk feature randomly selected is 0.75, the first propagation speed is 0.6 / day, and the first severity is medium. The number of times the difference between the first risk probability and the risk probability of other samples is less than 0.1 is calculated to be 35 times, that is, the first criticality is 35>20, and the deviation of the propagation speed is continued to be calculated. The number of times the difference between the first propagation speed and the propagation speed of other samples is less than 0.05 is calculated to be 22 times, that is, the second criticality is 22>20, and the deviation of the severity is continued to be calculated. The number of times the difference between the first severity and the severity of other samples is less than 0.1 is calculated to be 30 times, that is, the third criticality is 30>20, and the risk feature is set as the first critical risk feature and added to the first critical risk feature set.

[0082] By performing step-by-step deviation calculations and key feature screening on the first sample risk feature set, the module identifies multiple key risk feature sets, ensuring the accuracy and relevance of key features, thereby improving the comprehensiveness and reliability of unknown biological risk assessment.

[0083] Furthermore, the unknown risk coefficient output module 40 of the embodiment of the present application outputs the unknown risk coefficient, and further includes:

[0084] The multiple key risk feature sets are dimensionlessly processed to determine multiple risk feature sets; a comprehensive risk assessment is performed on the multiple risk feature sets through a biological expert system to determine multiple risk coefficient sets, and multiple risk coefficient means are calculated; based on the coefficient of variation method, credible weights are configured according to the multiple correlations, and the multiple risk coefficient means are weighted and fused according to the multiple credible weights to output the unknown risk coefficient.

[0085] Specifically, dimensionless processing is to normalize or standardize the data to make it dimensionless so that data of different characteristics can be compared on the same scale. The biological expert system is a system based on artificial intelligence and expert knowledge base, which is used to conduct comprehensive risk assessment of biological characteristics. The risk coefficient set is a set of risk coefficients obtained through comprehensive assessment, which is used to indicate the risk level of each risk characteristic. The risk coefficient mean is the average value of the risk coefficient set, which indicates the overall risk level of each characteristic. The coefficient of variation method is a statistical method used to measure the degree of dispersion of data distribution, usually expressed as the ratio of standard deviation to mean. The credible weight is an importance index assigned to each characteristic in the evaluation of multiple risk characteristics, which reflects the credibility or influence of each characteristic in the overall risk assessment.

[0086] First, the unknown risk coefficient output module 40 performs dimensionless processing on multiple key risk feature sets so that they have no units, which is convenient for comparison and analysis. For example, the data of risk probability, transmission speed and severity are scaled to the range of [0,1] using a normalization method. Next, a biological expert system is used to comprehensively evaluate the risk feature set after dimensionless processing to determine multiple risk coefficient sets. The expert system evaluates each feature set based on a known biological risk feature library and outputs the corresponding risk coefficient. This process can use a rule-based system such as CLIPS or DROOLS. Then, the risk coefficient set is statistically analyzed to calculate the mean risk coefficient of each feature set, which represents the average risk level of all risk features. The coefficient of variation method is used to calculate the coefficient of variation of each risk feature set, that is, the ratio of the standard deviation to the mean. The credible weight is configured according to the correlation between the coefficient of variation and multiple risk feature sets. The smaller the coefficient of variation, the smaller the relative variation of the data, the higher the credibility, and the larger the weight, which is usually the inverse of the coefficient of variation. The mean of the risk coefficient is weighted and fused according to the credible weight to calculate the final unknown risk coefficient. Weighted fusion can use the weighted average method. The above process can effectively determine the unknown risk factor, thereby improving the reliability and accuracy of risk assessment.

[0087] Furthermore, the potential risk level assessment module 50 of the embodiment of the present application determines the potential risk level, and further includes:

[0088] The number of abnormal events in the sample experimental event set is counted to determine the proportion of abnormal events; the proportion of abnormal events is input into the abnormal proportion-risk level data table, and the potential risk level is matched and output, wherein the abnormal proportion-risk level data table is determined based on the biological expert system.

[0089] Specifically, the potential risk level assessment module 50 identifies and counts the number of abnormal events from the sample experimental event set. Among them, abnormal events refer to those events that deviate from the normal range. Next, the module calculates the proportion of the number of abnormal events in the sample experimental event set, that is, the abnormal event ratio, expressed as a percentage. Then, the module inputs the abnormal event ratio into the abnormal ratio-risk level data table, and matches and outputs the potential risk level. The abnormal ratio-risk level data table is a reference table that maps the abnormal event ratio to the potential risk level. This data table is determined by the biological expert system based on historical data and expert knowledge. Finally, the module determines the potential risk level of the sample experimental event based on the matching results and generates a corresponding risk assessment report. The potential risk level is a level that represents the potential degree of danger of the experimental event. Through the above process, the potential risk level assessment module 50 can effectively determine the potential risk level of the experimental event and accurately assess the potential risk of unknown organisms in the laboratory environment.

[0090] In summary, the biological experiment safety risk assessment and early warning system provided in the embodiments of the present application has the following technical effects:

[0091] The biological attribute feature detection module 10 detects and obtains the biological attribute features of unknown organisms, provides basic data for subsequent comparison and risk assessment, and ensures the comprehensiveness of the assessment. The first associated organism determination module 20 inputs the biological attribute features into the concentrated knowledge base, performs similar traversal comparison according to the predetermined comparison strategy, selects concentrated organisms that meet the first similarity as the first associated organism, determines multiple first associated organisms, and quickly identifies known organisms similar to the unknown biological features, providing a basis for the subsequent high-frequency associated organism determination. The high-frequency associated organism determination module 30 calls the comprehensive knowledge base based on the multiple first associated organisms, maps and obtains multiple associated organism sets, selects associated organisms that meet the second similarity as high-frequency associated organisms, determines multiple high-frequency associated organisms, and compares a wider range of knowledge bases to improve the accuracy and comprehensiveness of the comparison, screens out organisms that are highly similar to the unknown biological features, and provides a more reliable basis for risk assessment. The unknown risk coefficient output module 40 uses the multiple high-frequency associated organisms as constraints, retrieves multiple sample risk feature sets based on the biological Internet, screens and determines multiple key risk feature sets, conducts a comprehensive risk assessment based on multiple correlations and multiple key risk feature sets, outputs unknown risk coefficients, comprehensively assesses the potential risks of unknown organisms, and provides a quantitative risk coefficient to ensure the accuracy and comprehensiveness of the assessment results. The potential risk level assessment module 50 obtains laboratory protection features, retrieves and obtains sample experimental event sets based on the unknown risk coefficients and the laboratory protection features as constraints, and determines the potential risk level based on the analysis of the sample experimental event set. If the potential risk level is greater than the predetermined risk threshold, a risk warning signal is generated to ensure that the laboratory can take timely measures to prevent potential hazards.

[0092] Overall, the embodiment of the present application integrates multiple modules such as biological attribute feature detection, related biological determination, risk coefficient output and potential risk level assessment to construct a more efficient and intelligent biological experiment safety risk assessment and early warning system. This system uses comprehensive feature detection, multi-level knowledge base comparison and dynamic Internet retrieval, and combines laboratory protection features. Even in the absence of sufficient historical data, it can conduct detailed and accurate risk assessments on unknown organisms and provide dynamic risk warnings in a timely manner, significantly improving the comprehensiveness, accuracy and reliability of biological laboratory risk assessments and early warnings, and ensuring the safety of laboratory operations.

[0093] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. Biological experiment safety risk assessment and early warning system, characterized by: The system comprises: A biological attribute feature detection module, wherein the biological attribute feature detection module is used to detect and obtain biological attribute features of an unknown organism, wherein the biological attribute features at least include amino acid sequences, nucleotide sequences and metabolic features; A first associated organism determination module, the first associated organism determination module is used to input the amino acid sequence, nucleotide sequence and metabolic characteristics into a concentrated knowledge base, perform similarity traversal comparison according to a predetermined comparison strategy, select concentrated organisms that meet a first similarity as first associated organisms, and determine multiple first associated organisms; A high-frequency associated organism determination module, the high-frequency associated organism determination module is used to call a comprehensive knowledge base based on the multiple first associated organisms, map and obtain multiple associated organism sets, select associated organisms that meet a second similarity as high-frequency associated organisms, and determine multiple high-frequency associated organisms, wherein the second similarity is greater than the first similarity, and the high-frequency associated organisms have an association degree mark; An unknown risk coefficient output module, the unknown risk coefficient output module is used to obtain multiple sample risk feature sets based on biological Internet search with the multiple high-frequency associated organisms as constraints, screen and determine multiple key risk feature sets, perform comprehensive risk assessment based on multiple correlation degrees and multiple key risk feature sets, and output unknown risk coefficients; A potential risk level assessment module, the potential risk level assessment module is used to obtain laboratory protection characteristics, retrieve and obtain a sample experimental event set based on the unknown risk coefficient and the laboratory protection characteristics as constraints, and determine the potential risk level based on the sample experimental event set analysis. If the potential risk level is greater than a predetermined risk threshold, a risk warning signal is generated; The system is also used to construct a comprehensive knowledge base and a condensed knowledge base, including: Building an initial knowledge base based on biological big data, wherein the initial knowledge base includes multiple known biological attribute features; Performing differential analysis on the plurality of known biological attribute features to determine a plurality of differential biological attribute features, and constructing a concentrated knowledge base based on the plurality of differential biological attribute features; Perform similarity clustering on a plurality of known biological attribute features according to a plurality of differential biological attribute features to determine a plurality of mapped biological attribute feature sets, wherein the differential biological attribute features and the mapped biological attribute feature sets have a corresponding relationship; Based on the corresponding relationship, construct a comprehensive knowledge base according to the plurality of differential biological attribute features and the plurality of mapped biological attribute feature sets; Wherein, the determining of a plurality of different biological attribute characteristics comprises: Randomly selecting a first attribute feature from the plurality of known biological attribute features, wherein the first attribute feature includes a first amino acid sequence, a first nucleotide sequence, and a first metabolic feature; Performing a similarity comparison between the first amino acid sequence and other amino acid sequences in the plurality of known biological attribute features, and counting the frequency of similarities being less than a predetermined amino acid similarity threshold, and setting the frequency as a first difference; If the first difference is greater than a predetermined threshold, then a similarity comparison is performed on the first nucleotide sequence and other nucleotide sequences in the plurality of known biological attribute features, and the frequency of similarities less than a predetermined nucleotide similarity threshold is counted and set as a second difference; If the second difference is greater than a predetermined threshold, performing a similarity comparison on the first metabolic feature and other metabolic features in the plurality of known biological attribute features to obtain a third difference; If the third difference is greater than a predetermined threshold, the first attribute feature is set as a difference biological attribute feature to obtain a plurality of difference biological attribute features; The system is also used to configure a predetermined comparison strategy, including: Configure multiple similarity comparison operators, use the amino acid sequence and the multiple similarity comparison operators as constraints, retrieve sample comparison data, analyze and determine multiple first comparison accuracies, and configure a first weight ratio based on the multiple first comparison accuracies; sequentially determining the second weight ratio of the nucleotide sequence and the third weight ratio of the metabolic characteristics; Constructing the predetermined comparison strategy based on the multiple similarity comparison operators, the first weight ratio, the second weight ratio and the third weight ratio; The unknown risk coefficient output module determines a plurality of key risk feature sets, including: Taking the multiple high-frequency associated organisms as constraints, multiple sample risk feature sets are obtained based on biological Internet search, wherein the sample risk features include sample risk probability, sample transmission speed and sample severity; Randomly selecting a first sample risk feature set, performing key feature screening on the first sample risk feature set, determining a first key risk feature set, and adding the first key risk feature set to the multiple key risk feature sets; The unknown risk coefficient output module performs key feature screening on the first sample risk feature set, including: Randomly selecting a risk feature from the first sample risk feature set as a first risk feature, wherein the first risk feature includes a first risk probability, a first transmission speed, and a first severity; Calculate the deviation between the first risk probability and other risk probabilities in the first sample risk feature set, and count the number of times the difference is less than a predetermined probability deviation threshold, and set the result as a first criticality; If the first criticality is greater than a predetermined critical threshold, performing deviation calculation on the first propagation speed and other propagation speeds in the first sample risk feature set to obtain a second criticality; If the second criticality is greater than a predetermined critical threshold, performing deviation calculation on the first severity and other severities in the first sample risk feature set to obtain a third criticality; If the third criticality is greater than a predetermined critical threshold, setting the first risk feature as a first critical risk feature and adding it to a first critical risk feature set; The unknown risk coefficient output module outputs the unknown risk coefficient, including: Performing dimensionless processing on the multiple key risk feature sets to determine multiple risk feature sets; Performing a comprehensive risk assessment on the multiple risk feature sets through a biological expert system, determining multiple risk coefficient sets, and calculating multiple risk coefficient means; Based on the coefficient of variation method, credible weights are configured according to the multiple associations, and the multiple risk coefficient means are weighted and fused according to the multiple credible weights to output the unknown risk coefficient; The potential risk level assessment module determines the potential risk level, including: Counting the number of abnormal events in the sample experimental event set to determine the proportion of abnormal events; The abnormal event ratio is input into the abnormality ratio-risk level data table, and the potential risk level is matched and output, wherein the abnormality ratio-risk level data table is determined based on the biological expert system.

Images