Toxic protein library construction method and apparatus, device and medium

Abstract

The present application relates to the field of bioinformatics. Disclosed are a toxic protein library construction method and apparatus, a device and a medium. The method comprises: acquiring a consensus peptide structure corresponding to a confirmed toxic protein sequence; on the basis of functional performance characteristic data and fluctuation trend characteristic data of the consensus peptide structure, constructing a target toxic protein core peptide set capable of covering a key functional domain of the confirmed toxic protein sequence; using the target toxic protein core peptide set to perform functional screening on a protein sequence to be recognized in the functional characteristic domain, so as to obtain an intermediate protein sequence for which it is necessary to verify whether a toxicity characteristic is satisfied; performing toxicity screening on the intermediate protein sequence in the toxicity characteristic domain; determining as a target toxic protein sequence the intermediate protein sequence that meets the toxicity characteristic; and finally, using the target toxic protein sequence to construct a toxic protein database. Thus, by means of a two-tier screening means in which functional screening and toxicity screening are combined, the toxic protein library construction method is achieved which can not only discover novel toxic proteins but also detect low-abundance toxic proteins, and has a wide coverage range.

Description

Methods, apparatus, equipment and media for constructing libraries of toxic proteins Technical Field

[0001] This application relates to the field of bioinformatics, and in particular to a method, apparatus, equipment and medium for constructing a library of toxic proteins. Background Technology

[0002] The construction of toxic protein databases has significant applications in bioinformatics and toxicology research, particularly in drug design, disease mechanism analysis, and ecological risk assessment. However, the methods for constructing toxic protein databases in related technologies still have certain limitations. Summary of the Invention

[0003] This application aims to at least partially solve one of the technical problems in related technologies. To this end, this application proposes a method, apparatus, equipment, and medium for constructing a library of toxic proteins. The main technical solutions adopted in this application include:

[0004] In a first aspect, embodiments of this application provide a method for constructing a toxic protein library. This method includes: obtaining common peptide structures corresponding to confirmed toxic protein sequences; and constructing a core peptide set of target toxic proteins that covers the key functional domains of confirmed toxic protein sequences based on functional performance characteristic data and fluctuation trend characteristic data of the common peptide structures; using the core peptide set of target toxic proteins to perform functional screening on the target protein sequences in their functional characteristic domains to obtain intermediate protein sequences that need to be verified for toxicity characteristics; performing toxicity screening on the intermediate protein sequences in their toxicity characteristic domains; and identifying the intermediate protein sequences that meet the toxicity characteristics as target toxic protein sequences; wherein the intermediate protein sequences that meet the toxicity characteristics are obtained by toxicity identification of the intermediate protein sequences using a toxic protein identification model; and constructing a toxic protein database using the target toxic protein sequences.

[0005] Optionally, based on the functional performance characteristics and fluctuation trend characteristics of the shared peptide structures, a set of target toxic protein core peptides capable of covering the key functional domains of the confirmed toxic protein sequence is constructed, including: aggregating the shared peptide structures into core peptides of the toxic protein family; clustering based on the functional performance characteristics and fluctuation trend characteristics of the core peptides of the toxic protein family to obtain multiple initial sets of toxic protein core peptides; and matching the confirmed toxic protein sequence with each initial set of toxic protein core peptides to identify the target toxic protein core peptide set among the multiple initial sets of toxic protein core peptides.

[0006] Optionally, the confirmed toxic protein sequence is matched with each initial toxic protein core peptide set to identify the target toxic protein core peptide set among multiple initial toxic protein core peptide sets. This includes: performing functional enrichment analysis on each initial toxic protein core peptide set to obtain the functional characteristics of each initial toxic protein core peptide set; calculating the matching degree between the confirmed toxic protein sequence and each initial toxic protein core peptide set based on the functional characteristics of each initial toxic protein core peptide set; and determining the initial toxic protein core peptide set corresponding to the highest matching degree as the target toxic protein core peptide set.

[0007] Secondly, embodiments of this application provide a toxic protein library construction device, which includes: a target toxic protein core peptide set construction module, used to obtain the common peptide structure corresponding to the confirmed toxic protein sequence, and construct a target toxic protein core peptide set that can cover the key functional domains of the confirmed toxic protein sequence based on the functional performance characteristic data and fluctuation trend characteristic data of the common peptide structure; a target toxic protein sequence determination module, used to use the target toxic protein core peptide set to perform functional screening on the functional feature domain of the protein sequence to be identified, to obtain intermediate protein sequences that need to be verified to meet toxicity characteristics, and to perform toxicity screening on the toxicity feature domain of the intermediate protein sequences, and to determine the intermediate protein sequences that meet the toxicity characteristics as target toxic protein sequences; wherein, the intermediate protein sequences that meet the toxicity characteristics are obtained by toxic protein identification model to identify the toxicity of the intermediate protein sequences; and a toxic protein database construction module, used to construct a toxic protein database using the target toxic protein sequences.

[0008] Thirdly, this application also provides a computer device including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of any of the above methods.

[0009] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of any of the methods described above.

[0010] Fifthly, the present invention provides a computer program product, including a computer program that, when executed by a processor, implements the steps of any of the above methods.

[0011] In the above embodiments, the common peptide structure corresponding to the confirmed toxic protein sequence is obtained. Based on the functional performance characteristics and fluctuation trend characteristics of the common peptide structure, a core peptide set of target toxic proteins covering the key functional domains of the confirmed toxic protein sequence is constructed. The core peptide set of target toxic proteins is used to perform functional screening on the protein sequence to be identified in its functional characteristic domains to obtain intermediate protein sequences that need to be verified for toxicity characteristics. These intermediate protein sequences are then screened for toxicity in their toxicity characteristic domains, and the intermediate protein sequences that meet the toxicity characteristics are identified as target toxic protein sequences. Finally, a toxic protein database is constructed using the target toxic protein sequences. Thus, a two-level screening method combining functional screening and toxicity screening achieves a toxic protein library construction method with broad coverage that can discover new toxic proteins and detect low-abundance toxic proteins. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the specific embodiments of this application or the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0013] Figure 1a is a flowchart of a method for constructing a toxic protein library according to an embodiment of this application;

[0014] Figure 1b is a schematic diagram of the interface of a toxic protein database provided according to an embodiment of this application;

[0015] Figure 2 is a flowchart of a method for identifying a set of core peptides of a target toxic protein according to an embodiment of this application;

[0016] Figure 3a is a flowchart of a method for identifying the core peptide set of a target toxic protein according to yet another embodiment of this application;

[0017] Figure 3b is an analysis curve of the enrichment degree of each toxic protein core peptide set according to an embodiment of this application;

[0018] Figure 4 is a flowchart of a method for determining a toxic protein identification model according to an embodiment of this application;

[0019] Figure 5 is a flowchart of a method for obtaining an initial protein sequence set according to an embodiment of this application;

[0020] Figure 6 is a structural block diagram of a toxic protein library construction device according to an embodiment of this application;

[0021] Figure 7 is an internal structural diagram of a computer device according to an embodiment of this application. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] The construction of toxic protein databases has significant applications in bioinformatics and toxicology research, particularly in drug design, disease mechanism analysis, and ecological risk assessment. However, several technical bottlenecks remain in the methods for constructing toxic protein libraries. Commonly used methods include the following:

[0024] 1. Experimental Sequencing: This is the most direct method for obtaining protein sequences, typically relying on mass spectrometry for high-precision sequencing. Mass spectrometry sequencing provides precise structural information of toxic proteins, enabling the good identification of protein sequences, structures, and subsequent translational modifications. However, experimental sequencing demands sophisticated equipment and skilled technicians, is expensive, and typically requires a considerable amount of time for each sample. Therefore, experimental sequencing is mainly suitable for studying specific high-abundance toxic proteins, while its application is limited for low-abundance or complex-sourced samples. Furthermore, mass spectrometry is sensitive to sample extraction purity and abundance, making it difficult to capture toxic proteins with low abundance or those difficult to extract, resulting in the absence of certain toxic proteins in databases.

[0025] 2. Homology Alignment: This method predicts potential sequences of toxic proteins by comparing sequences with existing protein sequence databases. For sequences similar to known toxic proteins, homology alignment can quickly infer unknown sequences. However, this method is ineffective for toxic proteins with significant evolutionary differences or those not yet covered in the database. Especially for newly discovered toxic proteins, accurate predictions through homology alignment are difficult if there are no known sequences for reference. Furthermore, as the amount of data increases, the alignment process may consume more computational resources, reducing the efficiency of database updates.

[0026] 3. Public Database Integration Method: This method integrates protein sequence data from different sources, such as UniProt and PDB, providing researchers with a wealth of protein information. While integration offers abundant toxic protein sequence information, the diverse data sources may lead to redundancy. For example, multiple databases may contain the same or highly similar sequences, increasing the system's storage burden and impacting retrieval efficiency. Furthermore, differences in sequence formats, annotations, and naming conventions across different databases can easily result in data inconsistencies during integration, increasing the difficulty of use.

[0027] Therefore, it is evident that the methods for constructing toxic protein libraries in related technologies face challenges in several aspects, as follows:

[0028] (1) Data redundancy: In the integration methods of public databases, data from different sources often contain a large number of repetitive toxic protein sequences. This not only wastes storage resources but also reduces the retrieval efficiency of the database in practical applications. For example, when retrieving the sequence of a certain toxic protein, researchers may need to screen out different versions of the same sequence, thereby increasing the data processing burden. Data redundancy may also lead to distorted judgment of sequence variations, affecting downstream toxicology research and drug development processes.

[0029] (2) Data consistency issues: In the integration methods of public databases, data from different sources often contain a large number of repetitive toxic protein sequences. This not only wastes storage resources but also reduces the retrieval efficiency of the database in practical applications. For example, when retrieving the sequence of a certain toxic protein, researchers may need to screen out different versions of the same sequence, thereby increasing the data processing burden. Data redundancy may also lead to distorted judgment of sequence variations, affecting downstream toxicology research and drug development processes.

[0030] (3) Limitations of novel toxic protein discovery: The reliance on homology matching methods leads to significant limitations in the discovery of novel toxic proteins. Many toxic proteins are difficult to predict accurately using existing databases due to the lack of evolutionarily similar sequences, which is particularly prominent in the study of environmental toxins and pathogen toxins. For drug target research of toxic proteins, the lack of accurate novel toxic protein sequence data will limit new drug design and development efforts.

[0031] (4) High experimental costs: Although experimental sequencing methods are accurate, their high cost and time consumption pose a bottleneck for building large-scale toxic protein databases. Especially when dealing with complex biological samples, the demand for experimental equipment and the cost of consumables can significantly increase the project budget. In addition, the collection of toxic proteins carries certain risks, especially when detecting highly toxic samples, and the high operating and protective costs further increase the difficulty.

[0032] (5) Difficulty in detecting low-abundance toxic proteins: Many important toxic proteins exist at low abundance in samples, making experimental sequencing methods ineffective in detecting them, resulting in a lack of such information in databases. However, these low-abundance toxic proteins often possess high toxicity and may be key to the research and development of novel antidotes or toxin neutralizers. The limitations of experimental sequencing methods make the detection and data collection of such toxic proteins difficult.

[0033] According to the embodiments of this application, a method, apparatus, device and medium for constructing a library of toxic proteins are provided. It should be noted that the steps shown in the flowcharts in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowcharts, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0034] This embodiment provides a method for constructing a toxic protein library. Figure 1a is a flowchart of a toxic protein library construction method according to an embodiment of this application. As shown in Figure 1a, the process includes the following steps:

[0035] S110. Obtain the common peptide structure corresponding to the confirmed toxin protein sequence, and based on the functional performance characteristics and fluctuation trend characteristics of the common peptide structure, construct a set of target toxin protein core peptides that can cover the key functional domains of the confirmed toxin protein sequence.

[0036] Among them, confirmed toxic protein sequences can be used to describe protein sequences that have been experimentally verified or confirmed to be toxic by authoritative databases. Specifically, confirmed toxic protein sequences can be obtained by screening multiple databases such as the Toxin and Toxin Target Database (T3DB) or UNIPROT.

[0037] Understandably, due to the sequence structure characteristics of proteins, it is difficult to find two proteins with completely identical sequences. Therefore, during detection and identification, certain common peptide structures are used as an important basis for protein identification and classification. Common peptide structures refer to peptide segments that are prevalent in a group of protein sequences and have similar structures and functions. For example, for toxic protein sequences, certain key regions of the toxic protein's function can be used as common peptide structures to be extracted and identified.

[0038] Furthermore, after obtaining the common peptide structure corresponding to the confirmed toxin protein sequence, it is necessary to construct a set of target toxin protein core peptides that can cover the key functional domains of the confirmed toxin protein sequence based on the functional performance characteristic data and fluctuation trend characteristic data contained in the common peptide structure.

[0039] Functional performance characteristic data can be used to describe the functional performance of proteins in biological processes, reflecting their activity, mechanism of action, and interactions with other molecules. For example, the characteristic data on the role, activity, and disease relationship of the shared peptide structure corresponding to the confirmed toxic protein sequence in vivo can be used as functional performance characteristic data.

[0040] Fluctuation trend data can be used to describe the trend of protein sequence fluctuations under different conditions or at different time points. For example, the change in the expression level of the common peptide structure corresponding to the confirmed toxic protein sequence or the sequence variation trend can be used as fluctuation trend data.

[0041] It is understandable that the core peptide set of the target toxic protein can be constructed based on the functional performance characteristics and fluctuation trend characteristics contained in the common peptide structure.

[0042] The target toxic protein core peptide set can refer to a group of toxic protein core peptides that share a common peptide structure and exhibit similar functional performance and a co-fluctuation trend in their fluctuation structure. For example, the target toxic protein core peptide set can be a Toxin Protein Family Featured Peptides Module (TFFP module).

[0043] Specifically, during aggregation, key functional domains that can cover the confirmed toxic protein sequence are constructed based on functional performance characteristic data and fluctuation trend characteristic data contained in the shared peptide structure. These key functional domains can be specific regions within the protein responsible for specific biological functions. For example, in constructing the core peptide set of the target toxic protein, the key functional domain is a specific region directly related to toxic effects.

[0044] S120. Using the core peptide set of the target toxic protein, the protein sequence to be identified is functionally screened in the functional feature domain to obtain the intermediate protein sequence that needs to be verified to meet the toxicity characteristics. The intermediate protein sequence is then screened for toxicity in the toxicity feature domain, and the intermediate protein sequence that meets the toxicity characteristics is identified as the target toxic protein sequence.

[0045] The protein sequence to be identified, intermediate protein sequence, and target toxic protein sequence can all refer to specific protein sequences. Specifically, the protein sequence to be identified can refer to a protein sequence whose toxicity has not yet been confirmed. For example, all protein sequences in the entire UniProt library can be used as the protein sequence to be identified.

[0046] Intermediate protein sequences can refer to protein sequences obtained after preliminary screening during the toxic protein recognition process. For example, intermediate protein sequences can be protein sequences obtained after screening for functional feature domains, which are similar to known toxic proteins in certain functional features, but whose toxicity has not yet been definitively confirmed.

[0047] The target toxic protein sequence refers to a protein sequence that, after toxicity screening and verification, exhibits characteristics similar to known toxic proteins and is thus confirmed to possess toxic properties. Intermediate protein sequences that meet the toxicity criteria are obtained through toxicity identification using a toxic protein recognition model.

[0048] Specifically, the process begins by using a set of core peptides for the target toxic protein, established based on the confirmed toxic protein sequence, to perform preliminary functional screening of the protein sequences to be identified. This aims to identify protein sequences that are functionally similar to known toxic proteins. For example, certain core peptide structures from the target toxic protein core peptide set (i.e., the TFFP module) that can serve as identification features can be compared with all protein sequences to be identified in the database to screen for candidate toxic protein sequences, i.e., intermediate protein sequences, with high affinity. This screening not only expands the coverage of the toxic protein database but also identifies potential toxic protein sequences, providing a reliable data foundation for subsequent toxic protein detection and functional studies.

[0049] Furthermore, after obtaining the intermediate protein sequences that need to be verified for toxicity characteristics, a pre-trained machine learning model can be used to perform secondary verification on these candidate toxic proteins to obtain the target toxic protein sequences confirmed to have toxicity characteristics. Specifically, the intermediate protein sequences are used as input features of the machine learning model, and high-precision algorithms such as random forests are used for secondary classification. The model then performs in-depth analysis of the characteristics of the candidate toxic proteins to identify and confirm their toxic properties. Finally, the intermediate protein sequences confirmed by the toxicity screening and toxic protein identification model, if they meet the toxicity characteristics, can be identified as the target toxic protein sequences.

[0050] S130. Construct a toxic protein database using the target toxic protein sequence.

[0051] Among them, the toxic protein database can be a resource database used to store and provide a large amount of sequence structure data and related information of known toxic proteins.

[0052] Specifically, the toxin protein sequences selected through secondary screening are all highly reliable target toxin protein sequences that conform to the characteristics and functions of toxin proteins. Furthermore, a database architecture capable of storing, managing, and retrieving toxin protein sequences can be designed, and the screened target toxin protein sequences can be stored in the database so that researchers can easily access and use this data. For example, Figure 1b shows the construction of the toxin protein database TOXINIA using the target toxin protein sequences obtained through secondary screening. As shown in Figure 1b, conventional BLAST searches can be performed on the toxin protein sequences, and further filtering based on different similarities can be performed for more refined searches of toxin protein information.

[0053] In the above embodiments, the common peptide structure corresponding to the confirmed toxic protein sequence is obtained. Based on the functional performance characteristics and fluctuation trend characteristics of the common peptide structure, a core peptide set of target toxic proteins covering the key functional domains of the confirmed toxic protein sequence is constructed. The target toxic protein core peptide set is used to perform functional screening on the protein sequence to be identified in its functional characteristic domains to obtain intermediate protein sequences that need to be verified for toxicity characteristics. These intermediate protein sequences are then screened for toxicity in their toxicity characteristic domains, and the intermediate protein sequences that meet the toxicity characteristics are identified as target toxic protein sequences. Finally, a toxic protein database is constructed using the target toxic protein sequences. Thus, a two-level screening method combining functional screening and toxicity screening achieves a toxic protein library construction method with broad coverage that can discover new toxic proteins and detect low-abundance toxic proteins.

[0054] In some implementations, based on the functional performance characteristics and fluctuation trend characteristics of the shared peptide structure, a core peptide set of the target toxic protein that covers the key functional domains of the confirmed toxic protein sequence is constructed, as shown in Figure 2, including:

[0055] S210, polymerize the shared peptide structure into the core peptide of the toxic protein family.

[0056] The core peptides of the toxin protein family refer to representative and conserved peptide segments within the toxin protein family. These peptides are ubiquitous in toxin proteins and are closely related to their toxic functions. Specifically, bioinformatics tools and algorithms can be used to identify common peptide structures in toxin protein sequences. These common peptide structures extracted from confirmed toxin protein sequences are then further aggregated into core peptides of the toxin protein family (TFFP). These core peptides contain fragments with strong toxin protein specificity, ensuring that the subsequently established set of target toxin protein core peptides has a strong ability to identify toxin protein specificity.

[0057] S220. Clustering was performed based on the functional performance characteristics and fluctuation trend characteristics of the core peptides of the toxic protein family to obtain multiple initial sets of toxic protein core peptides.

[0058] The initial set of toxic protein core peptides can also refer to a group of toxic protein core peptides that share a common peptide structure and exhibit similar functional performance and a co-fluctuation trend in their fluctuation structure. It is important to understand that the initial set of toxic protein core peptides is a preliminary set formed by clustering functional performance and fluctuation trend data of toxic protein family core peptides, and may contain multiple toxic protein core peptides with similar functions or structural characteristics.

[0059] Specifically, clustering can be based on the co-oscillation principle of the feature plane to identify TFFPs with highly correlated fluctuation trends and merge them to achieve the clustering process. It is understandable that the co-oscillation principle of the feature plane indicates that under the same biological function or specific role, the expression patterns of different feature modules usually exhibit consistent fluctuation trends; that is, under similar conditions, their expression feature change trends are similar and have strong synergy. Therefore, based on this, a similarity matrix can be constructed among multiple TFFPs to calculate the correlation and consistency between each TFFP. Specifically, firstly, the expression pattern of each TFFP is analyzed, its fluctuation on the feature plane is quantified, and a similarity matrix is ​​constructed to quantify the correlation between different TFFPs. Further, for TFFPs showing high correlation in the similarity matrix, they can be identified as modules with the same or similar fluctuation trends, and these TFFPs with consistent expression patterns and similar fluctuation trends are clustered into an initial set of toxic protein core peptides.

[0060] S230. Match the confirmed toxic protein sequence with each initial toxic protein core peptide set to identify the target toxic protein core peptide set among multiple initial toxic protein core peptide sets.

[0061] Specifically, the matching degree between the confirmed toxic protein sequence and each initial toxic protein core peptide set can be calculated to identify the best matching toxic protein core peptide set for each toxic protein, and this set is used as the target toxic protein core peptide set. These target toxic protein core peptide sets cover the key functional domains of the toxic protein, ensuring the high functionality and specificity of the finally constructed toxic protein database.

[0062] Optionally, the target toxic protein core peptide set can also be identified by calculating the correlation weights between different initial toxic protein core peptide sets and combining them with their functional characteristics. This allows for the identification of modules with similar functions and related fluctuation trends in the same dimension, which are then clustered into larger functional TFFP modules. In other words, it merges initial toxic protein core peptide sets with completely overlapping functions or highly consistent fluctuation trends, thereby achieving effective module redundancy reduction and enhancing the stability and reliability of TFFP modules.

[0063] Furthermore, based on the distribution of toxicity and TFFP quantity in each initial toxic protein core peptide set, toxic protein core peptide sets meeting high stability requirements were screened. Among these, highly stable toxic protein core peptide sets satisfy the following conditions: they can enrich at least two or more toxicity information molecules and contain at least two or more toxic protein family core peptides. Toxic protein core peptide sets meeting these conditions can be considered stable toxic protein core peptide sets. Moreover, because they maintain consistent performance across different characteristic peptides, they can be used as target toxic protein core peptide sets.

[0064] In the above embodiments, by confirming the toxic protein sequence and performing stepwise analysis, the core peptide set of the target toxic protein is ultimately determined. This helps to accurately identify and classify toxic proteins, providing a scientific basis for building a high-quality toxic protein database. Simultaneously, this method can also improve the accuracy of toxic protein identification, reduce data redundancy, and provide more reliable data support for toxicological research and drug development.

[0065] In some implementations, the confirmed toxic protein sequence is matched with each initial toxic protein core peptide set to identify the target toxic protein core peptide set among multiple initial toxic protein core peptide sets, as shown in Figure 3a, including:

[0066] S310. Perform functional enrichment analysis on each initial toxic protein core peptide set to obtain the functional characteristics of each initial toxic protein core peptide set.

[0067] Functional enrichment analysis is a bioinformatics method used to determine whether certain biological functions or processes are over- or under-represented in a group of proteins, helping to identify biological pathways or functional categories that are significantly active under specific conditions—in other words, functional annotation. Functional characteristics refer to the biological functional attributes of proteins, describing and classifying certain features of a protein's biological role. For example, functional characteristics can refer to the biological processes, molecular functions, or cellular components involved in the core peptide of a toxic protein.

[0068] Specifically, bioinformatics tools and databases are used to analyze each initial toxic protein core peptide set to identify the common functions of proteins in these sets. Further, by comparing the functional distribution of these sets with that of background proteins (such as confirmatory toxic proteins), it is determined which functions are enriched in that set. For example, the frequency and enrichment of a certain functional feature in each initial toxic protein core peptide set are calculated. When the enrichment of that functional feature in a certain initial toxic protein core peptide set reaches above the 90th percentile, that initial toxic protein core peptide set is identified as having significantly enriched that functional feature. Similarly, based on most of the enriched functional features, including molecular functions (such as protein binding, catalytic activity, etc.) and biological processes (such as cellular processes, metabolic processes, etc.), functional enrichment analysis is performed on each initial toxic protein core peptide set to assign functional annotations, thus obtaining the functional features of each initial toxic protein core peptide set. For example, if the adhesion of toxic proteins is taken as a functional feature, please refer to Figure 3b, where the horizontal axis represents the ID number of different toxic protein core peptide sets, and the vertical axis represents the distribution of toxic protein enrichment. The figure shows the adhesion of each toxic protein core peptide set to the toxic protein, and shows the distribution of different toxic protein core peptide sets in terms of toxic protein enrichment.

[0069] S320. Based on the functional characteristics of each initial toxic protein core peptide set, calculate the matching degree between the confirmed toxic protein sequence and each initial toxic protein core peptide set.

[0070] Specifically, the functional annotation results (i.e., functional features) of each initial toxic protein core peptide set obtained from functional enrichment analysis are used to assess the functional similarity between each initial toxic protein core peptide set and known confirmed toxic protein sequences. Furthermore, by calculating the matching degree of functional features, the degree of similarity between confirmed toxic protein sequences and initial toxic protein core peptide sets can be quantified.

[0071] For example, firstly, functional enrichment analysis is used to determine functional features from the confirmed toxic protein sequence and each initial toxic protein core peptide set. These features include the protein's biological function, molecular function (such as protein binding, catalytic activity, etc.), and biological processes (such as cellular processes, metabolic processes, etc.). These functional features are then used to functionally annotate each initial toxic protein core peptide set. Secondly, taking GPCR_Rhodpsn_7TM as an example, to calculate the matching degree of this functional feature, it needs to be converted into a numerical feature vector. That is, the frequency of occurrence of this functional feature or other relevant numerical indicators in the confirmed toxic protein sequence and each initial toxic protein core peptide set are calculated for quantitative analysis. Further, similarity calculation methods such as cosine similarity or Jaccard similarity coefficient are used to compare the functional similarity of this feature vector between the confirmed toxic protein sequence and each initial toxic protein core peptide set, obtaining several matching scores. These matching scores can then be used to assess the functional similarity between the confirmed toxic protein sequence and each initial toxic protein core peptide set on the functional feature GPCR_Rhodpsn_7TM. Similarly, multiple functional features can be converted into numerical feature vectors, with each feature corresponding to a functional dimension. Thus, for each initial set of toxic protein core peptides, a multi-dimensional numerical feature vector can be obtained, which contains multiple functional dimensions and features, for subsequent calculation and matching.

[0072] S330. Determine the initial set of toxic protein core peptides corresponding to the highest matching degree as the target set of toxic protein core peptides.

[0073] Specifically, after quantifying the matching degree between the confirmed toxic protein sequence and the initial toxic protein core peptide set, the initial toxic protein core peptide set most similar to the functional characteristics of the confirmed toxic protein sequence can be selected. That is, the initial toxic protein core peptide set corresponding to the highest matching degree is determined as the target toxic protein core peptide set, ensuring that this target toxic protein core peptide set is most likely to contain key functional domains similar to the confirmed toxic protein. For example, for each functional feature identified by functional enrichment analysis, the matching degree between the confirmed toxic protein sequence and each initial toxic protein core peptide set is calculated to identify the best-matching toxic protein core peptide set for the confirmed toxic protein. After screening, 274 mature toxic protein core peptide sets were finally identified, which can be used as the target toxic protein core peptide sets. These target toxic protein core peptide sets serve as the core classification structure of the toxic protein database, covering the key functional domains of toxic proteins and ensuring the high functionality and specificity of the toxic protein database.

[0074] For example, please refer to Table 1, which shows the number of functional annotations and corresponding functional domain features of each toxic protein core peptide set.

[0075] Table 1. Functional annotations and statistics of the core peptide set of the target toxic protein.

[0076] Understandably, the proteins enriched in each TFFP module typically share the same or similar functional domains, such as common and biologically important functional features found in toxic proteins like GPCR_Rhodpsn_7TM, P-loop_NTPase, and Prot_kinase_dom. This demonstrates that the core peptide set of the target toxic protein can effectively focus on specific functional groups, achieving highly specific functional annotation. Furthermore, during database construction, the core peptide set of the target toxic protein can accurately identify and classify toxic proteins, effectively reducing noisy data, improving annotation accuracy, and thus enhancing the overall quality of the database.

[0077] Similarly, Table 2 shows the total number of adhesion proteins, the category of functional annotation, the number of proteins that meet the annotation, and their percentage in each toxic protein core peptide set.

[0078] Table 2. Statistical and functional annotation analysis of the core peptide set of the target toxic protein.

[0079] As can be seen from the table, the vast majority of toxic proteins exhibit high consistency in the functional annotations of their corresponding target toxic protein core peptide sets, with 80%-100% of proteins matching the annotation, and even reaching 100% in some target toxic protein core peptide sets. This result demonstrates the high stability of target toxic protein core peptide sets in classifying and annotating toxic proteins. That is, target toxic protein core peptide sets can effectively enrich toxic proteins within the same functional group, thus ensuring the accuracy of the database's functional annotations. This is particularly evident in modules with a large total number of adhesion proteins (such as TFFP Module 23 and TFFP Module 9), indicating that these toxic protein core peptide sets still maintain high functional consistency and possess excellent stability.

[0080] In the above embodiments, functional enrichment analysis is performed on each initial set of toxic protein core peptides to assign functional annotations, thus obtaining the functional characteristics of each initial set of toxic protein core peptides. This ensures that each set of toxic protein core peptides has a clear biological meaning and distinguishing ability during the discrimination process, thereby enhancing the accuracy and interpretability of the discrimination. Simultaneously, through functional enrichment analysis and matching degree calculation, toxic proteins can be identified and classified more accurately, providing more reliable data support for toxicological research and drug development.

[0081] In some implementations, the common peptide structure is constructed by analyzing the confirmatory toxin sequence using multiple sequence alignment methods to extract the common peptide structure from the confirmatory toxin sequence.

[0082] Specifically, based on confirmed toxin protein sequences collected from databases, preliminary preprocessing operations are performed, including methods such as removing non-coding regions or adjusting sequence lengths to facilitate comparison. Further, multiple sequence alignment is used to analyze the toxin protein sequences, identifying similar regions and differences between sequences and displaying them in the form of an alignment matrix. Finally, representative conserved peptide segments are extracted as shared peptide structures for subsequent aggregation into Toxin Protein Family Featured Peptides (TFFPs), ensuring their specific recognition ability for toxin proteins.

[0083] In the above embodiments, multiple sequence alignment is used to extract common peptides representing key features of toxic proteins from the alignment results, so that they can be subsequently aggregated into core peptides of the toxic protein family. This provides important feature information for constructing a toxic protein database and helps to improve the accuracy and practicality of the database.

[0084] In some implementations, please refer to Figure 4, the toxic protein recognition model is obtained in the following manner:

[0085] S410. Obtain the initial set of toxic protein sequences and the initial set of non-toxic protein sequences.

[0086] The initial toxic protein sequence set can refer to a set of known toxic protein sequences obtained from a full-database protein database that are functionally similar to the confirmed toxic protein sequences and exhibit a co-fluctuation trend in their fluctuation structure. The initial non-toxic protein sequence set can refer to a set of known non-toxic protein sequences obtained from a full-database protein database that correspond to the initial toxic protein sequence set. Similarly, the initial toxic protein sequence set and the initial non-toxic protein sequence set can be obtained by filtering from multiple databases such as the Toxin and Toxin Target Database (T3DB) or the UNIPROT database.

[0087] S420. The target toxic protein core peptide set is matched with the initial toxic protein sequence set and the initial non-toxic protein sequence set to obtain the matched toxic protein sequence set and the matched non-toxic protein sequence set.

[0088] Understandably, if the initial set of toxic protein sequences is used as the positive sample for training the toxic protein recognition model, and the initial set of non-toxic protein sequences is used as the negative sample for training the toxic protein recognition model, in order to ensure that they are highly representative and accurate, it is also necessary to use the target toxic protein core peptide set to match the initial toxic protein sequence set and the initial non-toxic protein sequence set respectively. That is, the target toxic protein core peptide set is used as a feature to compare with each protein sequence in the initial toxic protein sequence set and the initial non-toxic protein sequence set in order to identify and extract the matching toxic protein sequence.

[0089] For example, a matching threshold can be set to determine the degree of similarity between a non-toxic protein sequence and the core peptide set of the target toxic protein, thus classifying it as a potential toxic protein sequence. Further, based on this matching threshold, protein sequences that show significant similarity to the core peptide set of the target toxic protein are screened from the initial toxic protein sequence set and the initial non-toxic protein sequence set, forming a matched toxic protein sequence set, which includes protein sequences considered to have potential toxic characteristics. Simultaneously, protein sequences that do not meet the matching criteria are grouped into a matched non-toxic protein sequence set, ultimately resulting in a matched toxic protein sequence set and a matched non-toxic protein sequence set.

[0090] The set of matched toxic protein sequences refers to the set of sequences obtained after matching and screening the target toxic protein core peptide set with the initial toxic protein sequence set, which includes toxic proteins that have significant matching with the target toxic protein core peptide and possess toxic characteristics. The set of matched non-toxic protein sequences, on the other hand, includes the original initial non-toxic protein sequence set and the non-toxic proteins that were screened from the original initial toxic protein sequence set after matching with the target toxic protein core peptide set and do not possess toxic characteristics.

[0091] S430. Based on the set of matched toxic protein sequences and the set of matched non-toxic protein sequences, the model is trained and tested to obtain the toxic protein recognition model.

[0092] Understandably, after matching and screening with the core peptide set of the target toxic protein, ensuring its reliability and accuracy, the matched toxic protein sequence set and the matched non-toxic protein sequence set can be used to construct training and testing sets for model training and testing. Furthermore, machine learning algorithms are used to train the model on the training set, and the model's recognition accuracy is evaluated on the testing set. Ultimately, an accurate and stable toxic protein recognition model is obtained, precisely identifying target toxic protein sequences that conform to the characteristics and functions of toxic proteins, and a toxic protein database is established based on this model.

[0093] Optionally, a random forest algorithm can be used, employing the set of matched toxic protein sequences as positive samples and the set of matched non-toxic protein sequences as negative samples to train the toxic protein identification model. During training, the toxic protein identification model learns the features that distinguish between toxic and non-toxic proteins. After training, a test set can be used to evaluate the toxic protein identification model to ensure its accuracy and generalization ability.

[0094] In the above embodiments, a machine model capable of accurately and stably identifying toxic proteins was constructed. This model can rapidly process and screen large amounts of sequence data, thereby reducing the experimental investment and time costs required for library construction. Furthermore, since the input to this toxic protein identification model is further matched with the core peptide set of the target toxic protein, it can achieve more sensitive identification of low-abundance toxic proteins. This method compensates for the limitations of experimental detection, making the toxic protein information in the constructed toxic protein database more comprehensive and accurate, facilitating subsequent research and applications. In addition, training and testing using model fine-tuning and transfer learning methods can further improve the toxic protein identification model's predictive ability for new toxic proteins, especially those lacking homologous sequences. This significantly improves the sensitivity of the toxic protein database to new toxic proteins, supporting the discovery of new targets in toxicology and drug development.

[0095] In some implementations, obtaining an initial set of toxic protein sequences and an initial set of non-toxic protein sequences, as shown in Figure 5, includes:

[0096] S510. Collect toxic protein data and non-toxic protein data from multiple protein databases.

[0097] Toxic protein data refers to protein data that shares a certain degree of similarity with confirmed toxic protein sequences. Specifically, this data can originate from specialized toxic protein databases or be collected through literature research, and is used to identify and study toxic proteins. For example, protein data with a certain degree of similarity to confirmed toxic protein sequences can be collected from databases such as the Toxin and Toxin Target Database (T3DB) or UNIPROT. Similarly, non-toxic protein data refers to data containing known non-toxic protein sequences obtained from a comprehensive protein database. Specifically, this data can be used to compare with toxic protein data to help establish characteristics that distinguish toxic and non-toxic proteins. Likewise, non-toxic protein sequences can be screened from databases such as UniProt and PDB as non-toxic protein data.

[0098] S520. Use homology modeling to remove redundant sequences from the toxic protein data to obtain cleaned toxic protein data.

[0099] Homology modeling is a bioinformatics method used to predict the three-dimensional structure of proteins by identifying their similarity based on the structural information of known homologous proteins (proteins with similar sequences and structures). Specifically, for the collected toxic protein data, duplicate and invalid sequences can first be removed to ensure the uniqueness of the data. Furthermore, homology screening is used to eliminate highly similar toxic protein sequences to maintain data diversity, ultimately resulting in cleaned toxic protein data.

[0100] S530. Based on the cleaned toxic protein data and non-toxic protein data, perform format conversion to obtain the initial toxic protein sequence set and the initial non-toxic protein sequence set.

[0101] Specifically, the cleaned toxic protein data and non-toxic protein data can be converted into a unified format to facilitate subsequent analysis and model training. After the format conversion is completed, the initial set of toxic protein sequences and the initial set of non-toxic protein sequences are obtained.

[0102] In the above implementation, the collected toxic protein sequence data undergoes deduplication and redundancy removal processes to effectively clean up duplicate or invalid sequences, and homology modeling is used to remove toxic proteins that are highly similar to existing sequences. This method ensures that the content of the toxic protein database is more concise and efficient, significantly reducing data redundancy issues, thereby improving the storage and computational efficiency of the toxic protein database. Furthermore, by using a unified data format and annotation standards, toxic protein data from different databases are standardized to form a highly consistent basic dataset. This consistency processing can effectively reduce problems such as incomplete annotations or inconsistent naming, ensuring the accuracy of the toxic protein database in large-scale applications and data mining.

[0103] In some embodiments, the method further includes: if the toxic protein identification model identifies an intermediate protein sequence that does not conform to the toxicity characteristics, the intermediate protein sequence that does not conform to the toxicity characteristics is identified as a non-toxic protein sequence.

[0104] Specifically, firstly, the target toxic protein core peptide set is used to functionally screen the protein sequence to be identified in its functional feature domain to obtain intermediate protein sequences that need to be verified to meet toxicity characteristics. Secondly, the intermediate protein sequences are screened for toxicity in their toxicity feature domain. If the toxic protein identification model identifies an intermediate protein sequence that does not meet the toxicity characteristics, that is, if an intermediate protein sequence does not meet the preset toxicity characteristic threshold or classification pattern under the evaluation of the toxic protein identification model, then this intermediate protein sequence will be considered to not meet the toxicity characteristics. Furthermore, this intermediate protein sequence will also be recorded and classified as a non-toxic protein sequence, and added to the non-toxic protein sequence set to provide a more accurate data foundation for the subsequent training, testing, or application of the toxic protein identification model.

[0105] Optionally, if the above situation occurs when applying the toxic protein identification model, the toxic protein database can be updated based on the identification results of the toxic protein identification model, and sequences identified as non-toxic proteins can be removed from the candidate list of potential toxic proteins to ensure the accuracy of the toxic protein database.

[0106] In the above implementation, this feedback update helps researchers screen potential non-toxic proteins from a large number of protein sequences, thereby more accurately identifying and classifying toxic proteins and providing more reliable data support for toxicology research and drug development.

[0107] This specification also provides a method for constructing a library of toxic proteins, which includes the following steps:

[0108] S602. Collect confirmed toxic protein sequences from multiple protein databases.

[0109] S604. Multiple sequence alignment is used to analyze the confirmed toxin protein sequence in order to extract the common peptide structure in the confirmed toxin protein sequence.

[0110] S606, polymerize the shared peptide structure into the core peptide of the toxic protein family.

[0111] S608. Clustering was performed based on the functional performance characteristics and fluctuation trend characteristics of the core peptides of the toxic protein family to obtain multiple initial sets of toxic protein core peptides.

[0112] S610. Perform functional enrichment analysis on each initial toxic protein core peptide set to obtain the functional characteristics of each initial toxic protein core peptide set.

[0113] S612. Based on the functional characteristics of each initial toxic protein core peptide set, calculate the matching degree between the confirmed toxic protein sequence and each initial toxic protein core peptide set.

[0114] S614. Determine the initial set of toxic protein core peptides corresponding to the highest matching degree as the target set of toxic protein core peptides.

[0115] S616. Collect toxic protein data and non-toxic protein data from multiple protein databases.

[0116] S618. Use homology modeling to remove redundant sequences from the toxic protein data to obtain cleaned toxic protein data.

[0117] S620. Based on the cleaned toxic protein data and non-toxic protein data, perform format conversion to obtain the initial toxic protein sequence set and the initial non-toxic protein sequence set.

[0118] S622. The target toxic protein core peptide set is matched with the initial toxic protein sequence set and the initial non-toxic protein sequence set to obtain the matched toxic protein sequence set and the matched non-toxic protein sequence set.

[0119] S624. Based on the set of matched toxic protein sequences and the set of matched non-toxic protein sequences, the model is trained and tested to obtain the toxic protein recognition model.

[0120] S626. Collect protein sequences to be identified from multiple protein databases.

[0121] S628. Using the core peptide set of the target toxic protein, the protein sequence to be identified is functionally screened in the functional feature domain to obtain the intermediate protein sequence that needs to be verified to meet the toxicity characteristics.

[0122] S630. Use a toxic protein identification model to screen intermediate protein sequences for toxicity in the toxicity feature domain, and identify intermediate protein sequences that meet the toxicity features as target toxic protein sequences.

[0123] S632. If the toxic protein identification model identifies an intermediate protein sequence that does not conform to the toxicity characteristics, the intermediate protein sequence that does not conform to the toxicity characteristics is identified as a non-toxic protein sequence. This sequence is then added to the set of matched non-toxic protein sequences for use in model training and testing by the toxic protein identification model.

[0124] S634. Construct a toxic protein database using the target toxic protein sequence.

[0125] It should be understood that although the steps in the flowchart above are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowchart above may include multiple steps or stages, which are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps.

[0126] This specification also provides a toxic protein library construction device 600, as shown in Figure 6, including: a target toxic protein core peptide set construction module 610, a target toxic protein sequence determination module 620, and a toxic protein database construction module 630, wherein:

[0127] The target toxic protein core peptide set construction module 610 is used to obtain the common peptide structure corresponding to the confirmed toxic protein sequence, and construct the target toxic protein core peptide set that can cover the key functional domains of the confirmed toxic protein sequence based on the functional performance characteristic data and fluctuation trend characteristic data of the common peptide structure.

[0128] The target toxic protein sequence determination module 620 is used to perform functional screening of the target toxic protein sequence in the functional feature domain using the core peptide set of the target toxic protein, to obtain intermediate protein sequences that need to be verified to meet the toxicity characteristics, and to perform toxicity screening of the intermediate protein sequences in the toxicity feature domain, and to determine the intermediate protein sequences that meet the toxicity characteristics as the target toxic protein sequences; wherein, the intermediate protein sequences that meet the toxicity characteristics are obtained by toxicity identification of the intermediate protein sequences through a toxic protein identification model.

[0129] Toxic protein database construction module 630 is used to construct a toxic protein database using target toxic protein sequences.

[0130] In some embodiments, the target toxic protein core peptide set construction module 610 is further configured to construct a target toxic protein core peptide set that covers the key functional domains of the confirmed toxic protein sequence based on the functional performance characteristic data and fluctuation trend characteristic data of the common peptide structure. This includes: aggregating the common peptide structure into toxic protein family core peptides; clustering based on the functional performance characteristic data and fluctuation trend characteristic data of the toxic protein family core peptides to obtain multiple initial toxic protein core peptide sets; and matching the confirmed toxic protein sequence with each initial toxic protein core peptide set to identify the target toxic protein core peptide set among the multiple initial toxic protein core peptide sets.

[0131] In some embodiments, the target toxic protein core peptide set construction module 610 is further configured to match the confirmed toxic protein sequence with each initial toxic protein core peptide set to identify the target toxic protein core peptide set among multiple initial toxic protein core peptide sets, including: performing functional enrichment analysis on each initial toxic protein core peptide set to obtain the functional characteristics of each initial toxic protein core peptide set; calculating the matching degree between the confirmed toxic protein sequence and each initial toxic protein core peptide set based on the functional characteristics of each initial toxic protein core peptide set; and determining the initial toxic protein core peptide set corresponding to the highest matching degree as the target toxic protein core peptide set.

[0132] In some embodiments, the target toxic protein core peptide assembly construction module 610 is also used to construct a common peptide structure by analyzing the confirmed toxic protein sequence using multiple sequence alignment methods to extract the common peptide structure in the confirmed toxic protein sequence.

[0133] In some embodiments, the target toxic protein sequence determination module 620 is further configured to obtain a toxic protein recognition model by: acquiring an initial set of toxic protein sequences and an initial set of non-toxic protein sequences; matching the target toxic protein core peptide set with the initial set of toxic protein sequences and the initial set of non-toxic protein sequences respectively to obtain a set of matched toxic protein sequences and a set of matched non-toxic protein sequences; and training and testing the model based on the set of matched toxic protein sequences and the set of matched non-toxic protein sequences to obtain a toxic protein recognition model.

[0134] In some embodiments, the target toxic protein sequence determination module 620 is also used to obtain an initial set of toxic protein sequences and an initial set of non-toxic protein sequences. This includes: collecting toxic protein data and non-toxic protein data from multiple protein databases; removing redundant sequences from the toxic protein data using homology modeling to obtain cleaned toxic protein data; and performing format conversion based on the cleaned toxic protein data and non-toxic protein data to obtain the initial set of toxic protein sequences and the initial set of non-toxic protein sequences.

[0135] In some embodiments, a toxic protein library construction device 600 is also used to identify intermediate protein sequences that do not conform to toxicity characteristics as non-toxic protein sequences if the toxic protein identification model identifies such sequences.

[0136] For specific limitations regarding the toxin library construction device, please refer to the limitations of the toxin library construction method described above, which will not be repeated here. Each module in the aforementioned toxin library construction device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0137] In this embodiment, a toxic protein library construction device is presented in the form of a functional unit. Here, a unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that execute one or more software or fixed programs, and / or other devices that can provide the above functions.

[0138] This application embodiment also provides a computer device, which can be a terminal, and its internal structure diagram is shown in Figure 7. The computer device includes a processor, memory, communication interface, display screen, and input device connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The communication interface is used for wired or wireless communication with external terminals. Wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for constructing a toxic protein library. The display screen can be a liquid crystal display or an e-ink display. The input device can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0139] Those skilled in the art will understand that the structure shown in Figure 7 is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0140] This application also provides a computer-readable storage medium. The methods described in this application can be implemented in hardware or firmware, or implemented as recordable on a storage medium, or implemented as computer code downloaded over a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and subsequently stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the methods shown in the above embodiments are implemented.

[0141] This application provides a computer program product including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the method of any embodiment of this application.

[0142] The toxin library construction method, apparatus, device, and medium described in the above embodiments can be implemented by a computer chip or physical entity, or by a product with a certain function. A typical implementation device is a computer. Specifically, the computer can be, for example, a personal computer, laptop computer, cellular phone, camera phone, smartphone, personal digital assistant, media player, navigation device, email device, game console, tablet computer, wearable device, or any combination of these devices.

[0143] For ease of description, the above devices are described separately by function as various units. Of course, in implementing this application, the functions of each unit can be implemented in one or more software and / or hardware.

[0144] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0145] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more flowchart illustrations and / or one or more block diagrams.

[0146] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0147] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide steps for implementing the functions specified in one or more flowcharts and / or one or more block diagrams.

[0148] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0149] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0150] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0151] The various embodiments in this specification are described in a progressive manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. Since they are fundamentally similar to the method embodiments, the descriptions are relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0152] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

[0153] Although embodiments of this application have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of this application, and all such modifications and variations fall within the scope defined by the appended claims.

Claims

1. A method of library building of a toxic protein, characterized by, The method includes: Obtain the common peptide structure corresponding to the confirmed toxin protein sequence, and based on the functional performance characteristics and fluctuation trend characteristics of the common peptide structure, construct a set of target toxin protein core peptides that can cover the key functional domains of the confirmed toxin protein sequence. The target toxic protein core peptide set is used to perform functional screening on the functional feature domain of the protein sequence to be identified, thereby obtaining intermediate protein sequences that need to be verified to meet the toxicity characteristics. The intermediate protein sequences are then screened for toxicity in the toxicity feature domain, and the intermediate protein sequences that meet the toxicity characteristics are identified as the target toxic protein sequences. The intermediate protein sequences that meet the toxicity characteristics are obtained by toxicity identification of the intermediate protein sequences through a toxic protein identification model. A toxic protein database was constructed using the target toxic protein sequence.

2. The method of claim 1, wherein, Based on the functional performance characteristic data and fluctuation trend characteristic data of the shared peptide structure, a set of target toxic protein core peptides capable of covering the key functional domains of the confirmed toxic protein sequence is constructed, including: The shared peptide structure was polymerized into a core peptide of the toxic protein family; Clustering was performed on the functional performance characteristics and fluctuation trend characteristics of the core peptides of the toxic protein family to obtain multiple initial sets of toxic protein core peptides. The confirmed toxic protein sequence is matched with each of the initial toxic protein core peptide sets to identify the target toxic protein core peptide set among the plurality of initial toxic protein core peptide sets.

3. The method of claim 2, wherein, The step of matching the confirmed toxic protein sequence with each of the initial toxic protein core peptide sets to identify the target toxic protein core peptide set among the plurality of initial toxic protein core peptide sets includes: Functional enrichment analysis was performed on each of the initial toxic protein core peptide sets to obtain the functional characteristics of each of the initial toxic protein core peptide sets. Based on the functional characteristics of each initial toxic protein core peptide set, the matching degree between the confirmed toxic protein sequence and each initial toxic protein core peptide set is calculated; The initial set of toxic protein core peptides corresponding to the highest matching degree is determined as the target set of toxic protein core peptides.

4. The method of claim 1, wherein, The shared peptide structure was constructed in the following manner: The confirmed toxin protein sequence was analyzed using multiple sequence alignment to extract the common peptide structure in the confirmed toxin protein sequence.

5. The method of claim 1, wherein, The toxic protein recognition model was obtained through the following method: Obtain the initial set of toxic protein sequences and the initial set of non-toxic protein sequences; The target toxic protein core peptide set is matched with the initial toxic protein sequence set and the initial non-toxic protein sequence set to obtain the matched toxic protein sequence set and the matched non-toxic protein sequence set. The toxic protein identification model is obtained by training and testing the model based on the set of matched toxic protein sequences and the set of matched non-toxic protein sequences.

6. The method of claim 5, wherein, The process of obtaining the initial set of toxic protein sequences and the initial set of non-toxic protein sequences includes: Data on toxic and non-toxic proteins were collected from multiple protein databases. Redundant sequences were removed from the toxic protein data using homology modeling to obtain cleaned toxic protein data; Based on the cleaned toxic protein data and the non-toxic protein data, the format is converted to obtain the initial toxic protein sequence set and the initial non-toxic protein sequence set.

7. The method of claim 1, wherein, The method further includes: If the toxic protein identification model identifies an intermediate protein sequence that does not meet the toxicity characteristics, the intermediate protein sequence that does not meet the toxicity characteristics is identified as a non-toxic protein sequence.

8. A device for library building of toxic proteins, characterized in that, The device includes: The target toxic protein core peptide set construction module is used to obtain the common peptide structure corresponding to the confirmed toxic protein sequence, and construct the target toxic protein core peptide set that can cover the key functional domains of the confirmed toxic protein sequence based on the functional performance characteristic data and fluctuation trend characteristic data of the common peptide structure. The target toxic protein sequence determination module is used to perform functional screening of the target toxic protein core peptide set on the functional feature domain of the protein sequence to be identified, to obtain intermediate protein sequences that need to be verified to meet the toxicity characteristics, and to perform toxicity screening on the intermediate protein sequences on the toxicity feature domain, and to determine the intermediate protein sequences that meet the toxicity characteristics as the target toxic protein sequences; wherein, the intermediate protein sequences that meet the toxicity characteristics are obtained by toxicity identification of the intermediate protein sequences through a toxic protein identification model. A toxic protein database construction module is used to construct a toxic protein database using the target toxic protein sequence.

9. A computer device, comprising: include: The system includes a memory and a processor, which are interconnected. The memory stores computer instructions, and the processor executes the computer instructions to perform a method for constructing a toxic protein library according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing a computer to execute a method for constructing a library of toxic proteins according to any one of claims 1 to 7.

Images