Drug-target interaction mechanism prediction method and system based on ternary comparative expertise decision-making
By decomposing the drug-target interaction mechanism prediction task into multiple ternary comparison subtasks and adopting a category-balanced decision voting strategy, the problem of inaccurate drug-target mechanism prediction in existing technologies is solved, more efficient drug-target mechanism prediction is achieved, and the application of computer-aided tools for drug design is promoted.
Patent Information
- Application Number
- CN202411416119.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-11
- Publication Date
- 2025-10-03
- Estimated Expiration
- 2044-10-11
AI Technical Summary
Existing technologies have difficulty accurately predicting the various mechanisms of action between drugs and targets, especially under long-tail distribution data conditions. The predictions are inaccurate and it is difficult to accurately identify the specific type of action of the drug on the target.
The multi-classification task is decomposed into multiple simple ternary comparison subtasks, which are trained using an expert knowledge model. The prediction results are integrated through a class-balanced decision voting mechanism. An expert knowledge decision method based on ternary comparison is designed, which combines the encoder-decoder architecture and deep learning algorithm to predict the drug-target interaction mechanism.
It improves the prediction accuracy of drug target mechanism under long-tail distribution conditions, enhances the feature representation ability of tail categories, shortens the drug development cycle and reduces costs.
Smart Images
Figure CN119446335B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the fields of bioinformatics and long-tail learning, and relates to a drug-target interaction mechanism prediction method and system based on ternary comparative professional knowledge decision-making. Background Art
[0002] Drug development is a complex and time-consuming process. To develop drugs that are reliably effective and safe, researchers must conduct extensive experimental validation, consuming significant human, material, and financial resources. According to statistics, the average cost of developing a small molecule drug is approximately 15 years and approximately $2 billion. As a key step in drug discovery and repositioning, drug-target interaction prediction can identify new targets for existing drugs and unlock new therapeutic uses for them. In the initial stages of drug development, if the ability to accurately predict interactions between a drug and specific targets in the body can be demonstrated, researchers can focus on specific targets and avoid blindly screening through a vast pool of candidates. Therefore, accurate and rapid identification of drug-target interactions can shorten drug development cycles and reduce R&D costs. This is particularly valuable in addressing emerging diseases, as it can uncover new clinical uses for rigorously tested and validated drugs, which has significant economic and societal value.
[0003] Existing related algorithms can quickly and accurately predict the interaction between drugs and specific targets, that is, whether the drug molecules can successfully bind to certain targets in the human body. However, drugs do not work simply by binding to targets, but affect the subsequent biological activities of target receptors from different mechanisms of action (agonism, antagonism, etc.) and to different degrees (complete, partial, etc.). Therefore, the interaction between drugs and targets should be studied in depth as a more complex multi-category mechanism prediction problem, but there are currently almost no deep learning models / algorithms for predictive analysis of all known mechanisms of action. In addition, since there is too little existing mechanism of action data and the entire data set presents a long-tail distribution, the prediction performance of existing algorithms in the tail category is very challenging. The existing technology has the following shortcomings:
[0004] 1) Insufficient research on predictive mechanisms of action: Studying the binding of drugs to human targets is essentially studying the various mechanisms of action produced by drug molecules acting on targets. Most existing methods focus solely on determining whether the two can bind. However, precisely determining the specific mechanism of action between the two (e.g., agonism, antagonism, etc.) is crucial to scientifically and accurately accelerating the drug repositioning process. Therefore, accurately determining the exact mechanism of action of a drug on its target poses a significant challenge to research related to drugs and human targets.
[0005] 2) Inaccurate predictions based on long-tail distribution data: Certain drug-target mechanisms of action (such as allosteric regulation) are relatively rare, so the amount of relevant data is extremely scarce compared to other mainstream action categories (such as activation, antagonism, etc.). Therefore, it is difficult to accurately predict the mechanism of action based on small sample data in the case of class imbalance. Summary of the Invention
[0006] To solve the above problems, the present invention provides a method and system for predicting drug-target interaction mechanisms based on ternary comparative professional knowledge decision-making.
[0007] Taking into account the difficulty of learning long-tail data, the present invention models the original multi-classification task into multiple simple sub-tasks based on ternary comparisons. It proposes a sub-task training strategy based on ternary comparison expertise decision-making. Based on the prediction results of all sub-tasks, a decision voting mechanism based on class balance is adopted to obtain the final multi-classification prediction result, which is then applied to the task of predicting drug-target interaction mechanisms. The first aspect of the invention relates to:
[0008] A drug-target interaction mechanism prediction method and system based on ternary comparative professional knowledge decision-making includes the following steps:
[0009] 1. Preprocessing of drug-target mechanism of action dataset;
[0010] Based on the existing drug-target interaction dataset, we further collected and organized samples related to various mechanisms of action, such as agonism and antagonism. After retrieving all qualified samples, we need to preprocess the raw data. The specific preprocessing steps include:
[0011] Step 11: Only drug-target pairs that are relevant to humans and have complete field information are retained.
[0012] Step 12: Use bioinformatics tools such as RDKit to determine whether the SMILES of the drug molecule is illegal.
[0013] Step 13: For proteins for which sequence information does not exist, the target protein sequence is searched across platforms in the UniProt database using a unique identifier.
[0014] Step 14: Analyze the mechanism type field of the processed dataset (such as including agonists and inhibitors) and filter the head and tail classes based on the data volume.
[0015] Finally, a large-scale dataset with "drug molecule-target protein sequence-mechanism type" as the basic data elements was obtained, of which the first two were used as input to the model, and the latter was used as the true label for the predicted output.
[0016] 2. Professional knowledge model training based on ternary comparison;
[0017] Step 21: Subtask data division. The original drug target mechanism prediction task is labeled as a multi-classification task with C categories. Each subtask aims to classify only two categories, resulting in a total of C*(C-1) / 2 subtasks. Each subtask is trained by a specific expertise model to extract knowledge related to the corresponding two categories (category 1, category 2). In order to ensure the accuracy of the expertise model prediction, an additional category is introduced. It is used to represent samples that do not belong to the two selected categories. Therefore, samples related to the two categories and samples belonging to the categories are extracted from the entire data set. samples to achieve the prediction task based on ternary comparison. Ultimately, the training dataset for this subtask is the combination of the above three datasets.
[0018] Step 22: Expertise model training. The expertise model is responsible for performing a simple three-way comparison task to determine whether the mechanism of the drug-target pair belongs to category 1, category 2, or Each expertise model is based on a classic encoder-decoder architecture. The encoder module consists of two encoders, one for drug molecules and the other for target proteins, to produce representation vectors for drugs and proteins:
[0019] 1) For drug molecule m, it is usually represented as a molecular graph, and then the naive GCN model is used to iteratively update its initial biochemical feature representation Specifically, the node embedding of layer l is Updated to:
[0020]
[0021] in is the output node embedding of the (l-1)th layer; is the trainable weight of the lth layer; A is the adjacency matrix of the molecular graph; σ(·) represents the nonlinear activation layer. After L iterative calculations, the final node embedding vector can be obtained Finally, a global pooling layer is applied to obtain the representation of the entire drug molecule graph.
[0022] 2) For the target protein t, it is usually represented as an amino acid sequence of unequal length, and then a 1D CNN model is used to iteratively update its initial biochemical feature representation Specifically, the node embedding of layer l is Updated to:
[0023]
[0024] in is the output embedding of the (l-1)th layer; k (l) Represent the number of input channels, output channels and convolution kernel size of the lth layer respectively; σ(·) represents the nonlinear activation layer. After K iterations, the final protein representation can be obtained.
[0025] The decoder module takes the representation vector of the drug and protein combination from the encoder as input and uses multiple linear layers to obtain a 3D output, where each dimension represents 1 class, 2 classes, or Class, the probability score of the node is obtained through the Softmax nonlinear layer The prediction of the category probability score is optimized using the cross entropy classification loss function. The specific formula is as follows:
[0026]
[0027] Where y is the true label of the sample. If the interaction between the samples is of type 1, the value is 0; if the interaction is of type 2, the value is 1; if the interaction between the samples is neither type 1 nor type 2, the value is -1.
[0028] 3. Decision voting based on category balance;
[0029] After obtaining the predicted labels of the subtasks from all the expertise models, a decision voting strategy based on class balance is adopted to effectively integrate the prediction results of each expertise model. Specifically, C*(C-1) / 2 initial prediction results Q are obtained from the expertise models, which are defined as follows:
[0030] Q=(q 12 ,q 13 ,…,q 1c ,q 23 ,…,q 2C ,…,q C-1,C ), (4)
[0031] where q i,j represents the output of the ternary comparison task between the i-th and j-th categories, and q i,j ∈{-1,0,1}. Next, based on the three possible outputs, update the final voting vector Y∈R for C categories C As follows: 1) If q i,j is 0, indicating that category i is the output label, and the reward will be assigned to the voting score of category i, denoted as Y i ; 2) If q i,j is 1, indicating that category j is the output label, and the reward will be assigned to the voting score of category j, denoted as Y j ;3) If qi,j -1, indicating category is the output label, and the penalty will be assigned to Y i and Y j .
[0032] Finally, after traversing the initial predictions of all professional models, the voting scores of all categories are counted. Due to the imbalance between categories, a class balance weight vector H∈R based on the number of samples is used C To calculate the final score, the category with the most votes is selected as the final prediction result.
[0033] 4. Achieve the task of predicting drug-target interaction mechanisms;
[0034] All samples in the drug-target interaction mechanism dataset were divided into training, validation, and test sets at a ratio of 80%, 10%, and 10%. The data was batched and fed into the model for multiple epochs of training. An adaptive learning rate optimization algorithm was used to iteratively update the model parameters. During this process, the performance of the model was recorded at each epoch. Accuracy and F1 scores were used to evaluate the model's performance. The model parameters with the best average performance were selected to ultimately generate a drug-target interaction mechanism prediction model.
[0035] The second aspect of the present invention relates to a drug-target interaction mechanism prediction system based on ternary comparison expertise decision-making: comprising a memory and one or more processors, wherein the memory stores executable code, and when the one or more processors execute the executable code, they are used to implement the drug-target interaction mechanism prediction method based on ternary comparison expertise decision-making of the present invention.
[0036] The third aspect of the present invention relates to a computer-readable storage medium having a program stored thereon, which, when executed by a processor, implements the drug-target interaction mechanism prediction method based on ternary comparison expertise decision of the present invention.
[0037] The advantages of this invention are: decomposing the original multi-classification problem into multiple simple subtasks, and effectively solving the problem of drug target mechanism prediction with long-tail distribution through a novel professional knowledge decision-making method based on ternary comparison. Secondly, a professional knowledge training strategy based on ternary comparison is designed: first, a divide-and-conquer strategy is adopted to decompose the original task into subtasks, ensuring that the main category does not occupy the resources of the tail category and reducing the difficulty of the original task; at the same time, the category is introduced The decision boundaries of the subtasks of categories 1 and 2 are determined and supplemented by new samples to generate more robust representations for the tail categories. Then, a decision voting module based on class balance is proposed to further integrate professional knowledge. Finally, a flexible model evaluation method is adopted to comprehensively evaluate the performance of the model by taking into account multiple indicators such as accuracy and F1 value. Through this mechanism prediction model, the ability to distinguish similar mechanism types can be enhanced, and a more effective and robust feature representation is provided for the tail categories, thereby accurately and efficiently completing the task of predicting the mechanism of action of drug targets. In addition, the detailed exploration of the evolvability and generalization of this method promotes its practicality and effectiveness in real-world scenarios, thereby providing an effective computer-aided tool for drug design, achieving the effects of shortening the drug development cycle and reducing R&D costs. BRIEF DESCRIPTION OF THE DRAWINGS
[0038] Figure 1 It is a flow chart of the method of the present invention.
[0039] Figure 2 It is a specific professional knowledge model architecture diagram of the present invention. DETAILED DESCRIPTION
[0040] The technical solution of the present invention is explained and described clearly and completely below.
[0041] Example 1
[0042] This embodiment relates to a method for predicting drug-target interaction mechanisms based on ternary comparison expertise decision-making, comprising the following steps:
[0043] 1. Preprocessing of drug-target mechanism of action dataset;
[0044] Based on the existing drug-target interaction dataset, we further collected and organized samples related to various mechanisms of action, such as agonism and antagonism. After retrieving all qualified samples, we need to preprocess the raw data. The specific preprocessing steps include:
[0045] Step 11: Only drug-target pairs that are relevant to humans and have complete field information are retained.
[0046] Step 12: Use bioinformatics tools such as RDKit to determine whether the SMILES of the drug molecule is illegal.
[0047] Step 13: For proteins for which sequence information does not exist, the target protein sequence is searched across platforms in the UniProt database using a unique identifier.
[0048] Step 14: Analyze the mechanism type field of the processed dataset (such as including agonists and inhibitors) and filter the head and tail classes based on the data volume.
[0049] Finally, a large-scale dataset with "drug molecule-target protein sequence-mechanism type" as the basic data elements was obtained, of which the first two were used as input to the model, and the latter was used as the true label for the predicted output.
[0050] 2. Professional knowledge model training based on ternary comparison;
[0051] Step 21: Subtask data division. The original drug target mechanism prediction task is labeled as a multi-classification task with C categories. Each subtask aims to classify only two categories, resulting in a total of C*(C-1) / 2 subtasks. Each subtask is trained by a specific expertise model to extract knowledge related to the corresponding two categories (category 1, category 2). In order to ensure the accuracy of the expertise model prediction, an additional category is introduced. It is used to represent samples that do not belong to the two selected categories. Therefore, samples related to the two categories and samples belonging to the categories are extracted from the entire data set. samples to achieve the prediction task based on ternary comparison. Ultimately, the training dataset for this subtask is the combination of the above three datasets.
[0052] Step 22: Expertise model training. The expertise model is responsible for performing a simple three-way comparison task to determine whether the mechanism of the drug-target pair belongs to category 1, category 2, or Each expertise model is based on a classic encoder-decoder architecture. The encoder module consists of two encoders, one for drug molecules and the other for target proteins, to produce representation vectors for drugs and proteins:
[0053] 1) For drug molecule m, it is usually represented as a molecular graph, and then the naive GCN model is used to iteratively update its initial biochemical feature representation Specifically, the node embedding of layer l is Updated to:
[0054]
[0055] in is the output node embedding of the (l-1)th layer; is the trainable weight of the lth layer; A is the adjacency matrix of the molecular graph; σ(·) represents the nonlinear activation layer. After L iterative calculations, the final node embedding vector can be obtained Finally, a global pooling layer is applied to obtain the representation of the entire drug molecule graph.
[0056] 2) For the target protein t, it is usually represented as an amino acid sequence of unequal length, and then a 1D CNN model is used to iteratively update its initial biochemical feature representation Specifically, the node embedding of layer l is Updated to:
[0057]
[0058] in is the output embedding of the (l-1)th layer; k (l) Represent the number of input channels, output channels and convolution kernel size of the lth layer respectively; σ(·) represents the nonlinear activation layer. After K iterations, the final protein representation can be obtained.
[0059] The decoder module takes the representation vector of the drug and protein combination from the encoder as input and uses multiple linear layers to obtain a 3D output, where each dimension represents 1 class, 2 classes, or Class, the probability score of the node is obtained through the Softmax nonlinear layer The prediction of the category probability score is optimized using the cross entropy classification loss function. The specific formula is as follows:
[0060]
[0061] Where y is the true label of the sample. If the interaction between the samples is of type 1, the value is 0; if the interaction is of type 2, the value is 1; if the interaction between the samples is neither type 1 nor type 2, the value is -1.
[0062] 3. Decision voting based on category balance;
[0063] After obtaining the predicted labels of the subtasks from all the expertise models, a decision voting strategy based on class balance is adopted to effectively integrate the prediction results of each expertise model. Specifically, C*(C-1) / 2 initial prediction results Q are obtained from the expertise models, which are defined as follows:
[0064] Q=(q 12 ,q 13 ,…,q 1C ,q 23 ,…,q 2C ,…,q C-1,C ), (4)
[0065] where q i,j represents the output of the ternary comparison task between the i-th and j-th categories, and q i,j ∈{-1,0,1}. Next, based on the three possible outputs, update the final voting vector Y∈R for C categories C As follows: 1) If q i,jis 0, indicating that category i is the output label, and the reward will be assigned to the voting score of category i, denoted as Y i ; 2) If q i,j is 1, indicating that category j is the output label, and the reward will be assigned to the voting score of category j, denoted as Y j ;3) If q i,j -1, indicating category is the output label, and the penalty will be assigned to Y i and Y j .
[0066] Finally, after traversing the initial predictions of all professional models, the voting scores of all categories are counted. Due to the imbalance between categories, a class balance weight vector H∈R based on the number of samples is used C To calculate the final score, the category with the most votes is selected as the final prediction result.
[0067] 4. Achieve the task of predicting drug-target interaction mechanisms;
[0068] All samples in the drug-target interaction mechanism dataset were divided into training, validation, and test sets at a ratio of 80%, 10%, and 10%. The data was batched and fed into the model for multiple epochs of training. An adaptive learning rate optimization algorithm was used to iteratively update the model parameters. During this process, the performance of the model was recorded at each epoch. Accuracy and F1 scores were used to evaluate the model's performance. The model parameters with the best average performance were selected to ultimately generate a drug-target interaction mechanism prediction model.
[0069] Example 2
[0070] This embodiment relates to a drug-target interaction mechanism prediction system based on ternary comparison expertise decision-making: it includes a memory and one or more processors, wherein the memory stores executable code, and when the one or more processors execute the executable code, it is used to implement the drug-target interaction mechanism prediction method based on ternary comparison expertise decision-making according to Example 1.
[0071] Example 3
[0072] This embodiment relates to a computer-readable storage medium having a program stored thereon. When the program is executed by a processor, the drug-target interaction mechanism prediction method based on ternary comparison professional knowledge decision-making of Example 1 is implemented.
[0073] The contents described in the embodiments of this specification are merely an enumeration of the implementation forms of the inventive concept. The scope of protection of the present invention should not be regarded as limited to the specific forms described in the embodiments. The scope of protection of the present invention also extends to equivalent technical means that can be conceived by those skilled in the art based on the inventive concept.
Claims
1. A drug-target interaction mechanism prediction method based on ternary comparative expertise decision-making includes the following steps: S1. Preprocess the drug-target interaction mechanism dataset: Based on the existing drug-target interaction dataset, collect and organize samples related to various mechanisms of action such as agonism and antagonism; After retrieving all qualified samples, the raw data are preprocessed; S2. Training a professional knowledge model based on ternary comparison, including: S21: Subtask data division: The original drug target mechanism prediction task is marked as a multi-classification task with multiple categories; each sub-task aims to classify only two categories, resulting in a total of subtasks; each subtask is trained by a specific expertise model to extract knowledge related to the corresponding two categories; in order to ensure the accuracy of the expertise model prediction, an additional category is introduced , used to represent samples that do not belong to the two selected categories; therefore, samples related to the two categories and samples belonging to the categories are extracted from the entire data set. samples to achieve the prediction task based on ternary comparison; finally, the training dataset of this subtask is the combination of the above three datasets; S22: Expertise model training: The expertise model is responsible for performing a simple ternary comparison task to determine whether the mechanism of a drug-target pair belongs to category 1, category 2, or Class; Each expertise model is based on the classic encoder-decoder architecture; the encoder module includes two encoders, which process drug molecules and target proteins respectively, thereby obtaining representation vectors of drugs and proteins: 1) Drug molecules It is represented as a molecular graph, and then the naive GCN model is used to iteratively update its initial biochemical feature representation. Specifically, Node embedding of the layer Updated to: ,（1） in It is The trainable weights of the layer; is the adjacency matrix of the molecular graph; Represents a nonlinear activation layer; after After the iterative calculation, the final node embedding vector is obtained ,Finally, a global pooling layer is applied to obtain the representation of the entire drug molecule graph; 2) Target protein It is represented as an amino acid sequence of unequal length, and then a 1D CNN model is used to iteratively update its initial biochemical feature representation. ;No. Node embedding of the layer Updated to: ,（2） in Representing the The number of input channels, output channels, and convolution kernel size of the layer; Represents a nonlinear activation layer; after After iterative calculations, the final protein representation is obtained ; The decoder module takes the representation vector of the drug and protein combination from the encoder as input and uses multiple linear layers to obtain a 3D output, where each dimension represents 1 class, 2 classes, or Class, the probability score of the node is obtained through the Softmax nonlinear layer ; The prediction of the category probability score is optimized using the cross entropy classification loss function. The specific formula is as follows: （3） in is the true label of the sample. If the interaction between the samples is of type 1, the value is 0; if the interaction is of type 2, the value is 1; if the interaction between the samples is neither type 1 nor type 2, the value is -1; S3. Class-balanced decision voting: After obtaining the predicted labels for the subtasks from all expertise models, a class-balanced decision voting strategy is adopted to effectively integrate the prediction results of each expertise model. Specifically, Obtained from professional models Initial prediction results , which is defined as follows: ,（4） in Represents the and The output of the ternary comparison task of the class, and ; Next, based on the three possible outputs, update The final voting vector of each category As follows: 1) If 0, indicating the category is the output label, and the reward will be assigned to the category The voting score of 2) If 1, indicating the category is the output label, and the reward will be assigned to the category The voting score of 3) If -1, indicating category is the output label, and the penalty will be assigned to and ; Finally, after traversing the initial predictions of all professional models, the voting scores of all categories are counted; due to the imbalance between categories, a class balance weight vector based on the number of samples is used To calculate the final score, and select the category with the most votes as the final prediction result; S4. Implement the task of drug-target interaction mechanism prediction.
2. The method for predicting drug-target interaction mechanisms based on ternary comparison expertise decision-making according to claim 1, wherein: Step S1 specifically includes: S11: Only drug-target pairs that are relevant to humans and have complete field information are retained; S12: Use the bioinformatics tool RDKit to determine whether the SMILES of drug molecules are illegal; S13: For protein sequence information that does not exist, the target protein sequence is searched across platforms in the UniProt database using a unique identifier; S14: Analyze the mechanism type field of the processed data set and filter the head and tail classes based on the data volume; Ultimately, a large-scale dataset with "drug molecule-target protein sequence-mechanism type" as the basic data elements was obtained, where the first two were used as input to the model and the latter was used as the true label for the predicted output.
3. The method for predicting drug-target interaction mechanisms based on ternary comparison expertise decision-making according to claim 1, characterized in that: Step S4 specifically includes: All samples in the drug-target interaction mechanism dataset were divided into training set, validation set and test set according to the ratio of 80%, 10% and 10% respectively; the data were divided into batches and sent to the model for multiple epoch training. The adaptive learning rate optimization algorithm was used to iteratively update the model parameters. In this process, the performance of each epoch model was recorded, and the performance of the model was evaluated using the accuracy and F1 value. The model parameters with the best average performance were selected to finally obtain the drug-target interaction mechanism prediction model.
4. A drug-target interaction mechanism prediction system based on ternary comparative expertise decision-making, characterized by: The method comprises a memory and one or more processors, wherein the memory stores executable code, and when the one or more processors execute the executable code, the method is used to implement the drug-target interaction mechanism prediction method based on ternary comparison expertise decision-making according to any one of claims 1 to 3.
5. A computer-readable storage medium, characterized in that A program is stored thereon, and when the program is executed by a processor, the drug-target interaction mechanism prediction method based on ternary comparison professional knowledge decision-making according to any one of claims 1 to 3 is implemented.