A drug-target binding affinity prediction method based on graph neural network

Abstract

The present invention discloses a method for predicting drug-target binding affinity based on a graph neural network, belonging to the field of bioinformatics. The method comprises the following steps: S1: sample preprocessing; S2: model construction; S3: model training; and S4: DTA prediction. The method first obtains relevant data from the Davis and KIBA datasets. Through data preprocessing, drug molecular graphs, drug Morgan fingerprints, and protein sequence information are obtained, respectively, and input into a GATv2 network, a multi-layer perceptron network, and a multi-scale convolutional neural network for feature extraction. The Morgan fingerprints and molecular graph features are then spliced ​​together using layer attention to obtain drug features. These features are then spliced ​​with protein features and input into a prediction network for drug-target binding affinity prediction.

Description

Technical Field

[0001] The present invention relates to the field of bioinformatics, and in particular to a method for predicting drug-target binding affinity based on graph neural networks. Background Art

[0002] Proteins are involved in numerous cellular processes, and many diseases are caused by abnormal protein function. Consequently, a large number of drugs have been developed to target proteins. Drug-target binding affinity refers to the strength or tightness of the binding between a drug and its target. Understanding the binding affinity between a drug and its target is crucial for evaluating drug efficacy and selecting appropriate drug candidates during drug design and discovery. Structural changes in a drug can significantly affect the strength of protein binding affinity, making it crucial to identify specific drug-protein binding partners.

[0003] While experimental methods have been applied in drug discovery and drug repurposing, identifying effective and safe drugs targeting a given protein requires pharmacologists to test thousands of chemical compounds, which is extremely time-consuming and labor-intensive. The rise of computers has brought computational methods to drug discovery, and in the biomedical field, computational methods have become a new hot topic as a complement to experimental methods. Drug-target interaction (DTI) plays a crucial role in drug discovery and drug repurposing. Drug-target binding affinity (DTA) is a regression problem that provides information about the strength of the interaction between a drug and target pair, typically expressed as metrics such as dissociation constant, inhibition constant, and half-maximal inhibitory concentration. With the development of machine learning and artificial intelligence technologies, a growing number of studies have begun exploring the use of these techniques to predict drug-target binding affinities. Machine learning algorithms can learn patterns from large amounts of known drug-target interaction data, enabling the prediction of unknown drug-target interactions. This data-driven approach generally offers advantages in predictive accuracy and efficiency.

[0004] In short, as diseases are constantly being conquered, there is an extremely high demand for the speed of drug discovery. Technical means relying on deep learning have greatly improved the efficiency and success rate in actual drug research and development. Summary of the Invention

[0005] The technical problem to be solved by the present invention is that the accuracy of drug-target binding affinity prediction can be improved. In terms of drug feature extraction, the global information of the drug is difficult to capture, and a simple two-layer graph neural network cannot obtain its specific structural information, such as ring structure; in terms of protein feature extraction, protein feature extraction is not comprehensive. A drug-target binding affinity prediction method based on graph neural network is provided.

[0006] The present invention solves the above technical problems through the following technical solutions, which include the following steps:

[0007] S1: Sample pretreatment

[0008] The Davis and KIBA datasets are divided into training and test sets according to a set ratio, and the drug protein information in the training and test sets is processed into drug molecular graphs, Morgan fingerprints, and protein sequence information;

[0009] S2: Model construction

[0010] Construct an initial drug-target binding affinity prediction model, including a drug protein feature extraction model based on graph neural networks and multi-layer perceptron networks, and a prediction model based on a prediction network;

[0011] S3: Model training

[0012] The initial drug-target binding affinity prediction model is trained using the training set, and the final drug-target binding affinity prediction model is obtained after training;

[0013] S4: DTA Forecast

[0014] The final drug-target binding affinity prediction model is tested on the test set and the binding affinity index is output.

[0015] Furthermore, in step S1, the specific processing process is as follows:

[0016] S11: Divide the Davis and KIBA datasets into training and test sets according to the set ratio;

[0017] S12: Use RDkit to preprocess the drug SMILES sequences of samples in the dataset into a graph with nodes and adjacency matrix, that is, to obtain drug molecular graph information;

[0018] S13: Generate a Morgan fingerprint with a radius of 2 and a bit number of 1024 based on the drug SMILES sequence using RDkit, thus obtaining the Morgan fingerprint information;

[0019] S14: The protein FASTA sequence of the sample in the dataset is converted into a corresponding integer sequence through the corresponding vocabulary, that is, the protein sequence information is obtained.

[0020] Furthermore, in step S2, the specific processing process of the drug protein feature extraction model is as follows:

[0021] S21: The drug molecule graph is fed into the GATv2 network to obtain 96-dimensional drug molecule graph features. The GATv2 network is a graph neural network.

[0022] S22: Send the drug Morgan fingerprint into the multi-layer perceptron network to obtain the 96-dimensional drug Morgan fingerprint feature;

[0023] S23: The drug Morgan fingerprint features and drug molecular graph features are concatenated and reduced in dimension after passing through the layer attention module to obtain a 96-dimensional drug feature.

[0024] S24: The protein integer sequence is passed through a multi-scale convolutional neural network to obtain 96-dimensional protein features.

[0025] Furthermore, in step S2, the processing of the prediction model is as follows:

[0026] S25: Splicing the drug feature and the protein feature to obtain a spliced ​​feature;

[0027] S26: The spliced ​​features are fed into the prediction network to obtain the binding affinity prediction results.

[0028] Furthermore, in step S21, the drug molecule graph is represented as G = (V, E), where V represents the node of the drug atom, and E represents the set of edges connecting the drug nodes. Each node obtains an n-dimensional representation through Deepchem, including extracted atomic features, bond features, substructure features and topological structure features.

[0029] Furthermore, in step S21, the GATv2 network adopts a dynamic attention mechanism to aggregate the corresponding graph representations obtained through the drug SMILES sequence to obtain drug molecular graph features. The aggregation method is as follows:

[0030] e(h i ,h j )=a T LeakyRelu(W0[h i ||h j ])j∈N i

[0031] Among them, e(h i ,h j ) represents the importance of neighbor node j to the original node i, N i Represented as the set of all neighbor nodes of node i is a learnable matrix.

[0032] Furthermore, in step S22, the molecular structure is traversed through the drug Morgan fingerprint, and the environmental information around each atom is recorded to obtain a 1024-dimensional vector, which is then reduced to a 96-dimensional vector through a multi-layer perceptron network, and the supplementary structural information of the drug is output, that is, the drug Morgan fingerprint feature is obtained.

[0033] Furthermore, in step S23, the drug molecular graph features and the drug Morgan fingerprint features are spliced ​​together by the layer attention module, and the splicing method is as follows:

[0034] h k =[W1h j ∥(1-W1)h i ].

[0035] Furthermore, in step S24, the multi-scale convolutional neural network includes three convolution branches. Different convolution branches have different numbers of convolution layers, and the corresponding convolution kernel sizes are also different. The processing method of the multi-scale convolutional neural network is as follows:

[0036] y=W2(m(F1(Q))∥m(F2(Q))∥m(F3(Q))

[0037] Among them, y represents protein features, Q is the input matrix obtained by protein integer sequence, and F i represents the i-th convolution branch, m represents the maximum pooling process, and W2 is a learnable matrix.

[0038] Furthermore, in step S26, the prediction network includes three linear transformation layers and a fully connected layer. A Relu activation function layer and a dropout layer are connected after each linear transformation layer. The binding index is obtained through the three linear transformation layers, and then processed through the fully connected layer to obtain the final drug protein binding affinity index.

[0039] Compared with the prior art, the present invention has the following advantages:

[0040] 1. The Morgan fingerprint information is integrated into the drug feature extraction part, which can supplement the structural information lost when the drug molecular graph passes through the graph neural network. After obtaining the molecular graph features and Morgan fingerprint features, layer attention is used to obtain their corresponding contributions, ultimately obtaining more comprehensive drug feature information;

[0041] 2. For an entire protein sequence, only a portion of it plays a role in the reaction. For example, the active site in an enzyme protein is only a portion of the region involved in the catalytic reaction. Therefore, a multi-scale convolutional neural network is used for protein feature extraction. This network consists of three convolutional branches. Different convolutional branches have different numbers of convolutional layers, which means that the sizes of their convolution kernels are different. This enables the convolutional neural network to extract local information at more scales. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 is a schematic flow chart of a method for predicting drug-target binding affinity based on a graph neural network in an embodiment of the present invention;

[0043] Figure 2 Schematic diagram of the drug-target binding affinity prediction model structure in an embodiment of the present invention;

[0044] Figure 3 is a schematic diagram of the structure of a multi-scale convolutional neural network in an embodiment of the present invention;

[0045] Figure 4 Schematic diagram of the structure of the prediction network in the prediction model in an embodiment of the present invention;

[0046] Figure 5 : This is the attention heat map of different graph neural networks in the embodiments of the present invention, where a is the GAT network and b is the GATv2 network;

[0047] Figure 6 Schematic diagram of the fitting curves of the final drug-target binding affinity prediction model in the embodiment of the present invention on different data sets, where a is the Davis data set and b is the KIBA data set. DETAILED DESCRIPTION

[0048] The following is a detailed description of an embodiment of the present invention. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process. However, the protection scope of the present invention is not limited to the following embodiment.

[0049] This embodiment provides a technical solution: a method for predicting drug-target binding affinity based on a graph neural network, comprising the following steps:

[0050] S1: Sample pretreatment

[0051] The Davis and KIBA datasets are divided into training and test sets according to a set ratio, and the drug protein information in the training and test sets is processed into drug molecular graphs, Morgan fingerprints, and protein sequence information;

[0052] S2: Model construction

[0053] Construct an initial drug-target binding affinity prediction model, including a drug protein feature extraction model based on graph neural networks and multi-layer perceptron networks, and a prediction model based on a prediction network;

[0054] S3: Model training

[0055] The initial drug-target binding affinity prediction model is trained using the training set, and the final drug-target binding affinity prediction model is obtained after training;

[0056] S4: DTA Forecast

[0057] The final drug-target binding affinity prediction model is used to predict the binding affinity index on the test set.

[0058] In this embodiment, step S1 includes the following sub-steps:

[0059] S11: The datasets used in the present invention are the Davis and KIBA datasets, where the Davis dataset includes interactions between 442 proteins and 68 drugs. The KIBA dataset initially contained 467 proteins, 52,498 drugs, and KIBA scores between proteins and drugs. The KIBA score is used to measure the biological activity of kinase inhibitors and is considered to be related to the binding affinity of the kinase. In order to make the dataset more balanced, a dataset containing 229 proteins and 2,111 drugs was finally formed after screening. All drugs in the Davis dataset interact with proteins, while there are 118,254 interactions in the KIBA dataset.

[0060] Table 1 The number and proportion of various types of data in the dataset

[0061] Davis KIBA protein 442 299 drug 68 2111 interaction 30056 118254

[0062] S12: Drug SMILES sequences were preprocessed using RDkit into a graph with nodes and an adjacency matrix, as well as a Morgan fingerprint with a radius of 2 and a number of 1024. Protein FASTA sequences were converted into integer sequences using a dictionary. SMILES are short ASCII strings used to describe chemical substances.

[0063] In this example, the specific processing of the initial drug-target binding affinity prediction model is as follows:

[0064] S21: The drug molecule graph is fed into the GATv2 network to obtain 96-dimensional drug molecule graph features. The GATv2 network is a graph neural network.

[0065] S22: Send the drug Morgan fingerprint into the multi-layer perceptron network to obtain the 96-dimensional drug Morgan fingerprint feature;

[0066] S23: The drug Morgan fingerprint features and drug molecular graph features are concatenated and reduced in dimension after passing through the layer attention module to obtain a 96-dimensional drug feature.

[0067] S24: Send the protein integer sequence to the multi-scale convolutional neural network to obtain 96-dimensional protein features;

[0068] S25: Splicing the drug feature and the protein feature to obtain a spliced ​​feature;

[0069] S26: The spliced ​​features are fed into the prediction network to obtain the binding affinity prediction results.

[0070] In this embodiment, the GATv2 network adopts a dynamic attention mechanism to aggregate the corresponding graph representations obtained through the drug SMILES sequence to obtain the drug molecule graph features, that is, graph G = (V, E), where V represents the node of the drug atom and E represents the set of edges connecting the drug nodes. The n-dimensional representation of each node obtained by Deepchem includes the extracted atomic features (atom type, charge, number of bonds, number of hydrogen bonds, etc.), bond features (bond type (single bond, double bond, triple bond), bond length, bond angle, etc.), substructure features (such as the number and size of rings, connections between rings, etc.) and topological structure features (such as the topological structure of the molecule, connection pattern between atoms, etc.). Among them, e(h i ,h j ) represents the importance of neighbor node j to the original node i, N i Represented as the set of all neighbor nodes of node i is a learnable matrix that is aggregated as follows:

[0071] e(h i ,h j )=a T LeakyRelu(W0[h i ||h j ])j∈N i

[0072] In this embodiment, the molecular structure is traversed through the drug Morgan fingerprint, and the environmental information around each atom is recorded to obtain a 1024-dimensional vector. The 1024-dimensional vector is then reduced to a 96-dimensional vector through a multi-layer perceptron network, and the supplementary structural information of the drug is output, that is, the drug Morgan fingerprint feature is obtained.

[0073] In this embodiment, layer attention is used to stitch together the drug molecular graph features and the drug Morgan fingerprint features. The stitching method is as follows:

[0074] h k =[W1h j ∥(1-W1)h i ].

[0075] In this embodiment, the multi-scale convolutional neural network includes three convolution branches, where different convolution branches have different numbers of convolution layers, which means that the sizes of their convolution kernels are different, which enables the convolutional neural network to extract local information at multiple scales. The input matrix Q passes through F i Mapping to Matrix After activation by the activation function Relu, the processing method of the multi-scale convolutional neural network is as follows:

[0076] y=W2(m(F1(Q))∥m(F2(Q))∥m(F3(Q))

[0077] Among them, F i It represents a branch in multiple branches, m represents the maximum pooling, is a learnable matrix.

[0078] In this example, after obtaining the drug and protein feature vectors, they are concatenated and fed into a prediction network. The prediction network consists of three linear transformation layers and a fully connected layer. Each linear transformation layer is followed by a Relu activation function layer and a dropout layer. The fully connected layer then processes the vectors to produce the final drug-protein binding affinity index. The prediction network maps the high-dimensional features obtained from the fusion of drug-protein features to the final output, the drug-target binding affinity index.

[0079] In this embodiment, Table 2 shows the impact of adjusting different modules on the experimental results. MCNN represents the replacement of the multi-scale convolution module, ECFP represents the addition of Morgan fingerprint information and feature extraction modules. Compared with only adding multi-scale convolution, all performances are improved in Davis and KIBA datasets. LA (layer) represents the addition of layer attention. Compared with adding ECFP and MCNN modules, all indicators are improved in both datasets, and the final results are obtained in the ablation experiment. The lower the MSE, the better, CI and The higher the better.

[0080] Table 2 Comparison of experimental results of adding different modules

[0081]

[0082] Table 3 shows the results of a model comparison of different graph neural networks, including mainstream models such as GCN, GAT, GIN, GCN-GAT, and GATv2. Experimental data shows that the GATv2 network achieves excellent results across all metrics on both datasets. Compared to other graph neural networks, only the MSE on the KIBA dataset is slightly lower than that of the GIN by 0.1%. All other metrics are the best among the graph neural network comparisons. This is due to the GATv2 network's use of dynamic attention. The Davis dataset has fewer samples and an unbalanced label distribution, while the KIBA dataset has a high concentration of single-label samples. This makes the GATv2 network better at distinguishing similarities and differences between samples.

[0083] Table 3 Comparison of experimental results of different graph neural networks

[0084]

[0085] Table 4 shows the experimental results of different networks on the Davis and KIBA datasets, including traditional machine learning algorithms and the latest deep learning algorithms. Traditional machine learning methods include KronRLS and SimBoost, and deep learning methods include DeepDTA, GraphDTA, MATT DTI, DeepGLSTM, and DGDTA. As shown in Table 4, the drug-target binding affinity prediction method involved in the present invention outperforms other models in multiple indicators in the test set. The CI in the two datasets was 0.903 and 0.904, respectively, which were the best among all models.

[0086] Table 4 Experimental results of different networks on Davis and KIBA datasets

[0087]

[0088]

[0089] In summary, the drug-target binding affinity prediction method based on graph neural network in the above embodiment can supplement the structural information of drug characteristics by adding the Morgan fingerprint information of the drug to the drug protein sequence data. For example, the graph neural network cannot know the specific structural information, such as the ring structure, etc.; an optimization scheme is proposed to aggregate the drug molecular graph features and the drug Morgan fingerprint features by adding LA attention; since the protein participates in the drug effect only at a certain point, the feature extraction of the protein is also improved in the present invention, and multi-scale convolution is used to obtain local features of different scales to simulate drug-target binding; finally, the effectiveness of the improvement is proved through various comparative experiments, and the best results are obtained in multiple evaluation indicators, proving that it has the best effect.

[0090] Although the embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

Claims

1. A drug-target binding affinity prediction method based on graph neural network, characterized in that: The following steps are involved: S1: Sample pretreatment The Davis and KIBA datasets are divided into training and test sets according to a set ratio, and the drug protein information in the training and test sets is processed into drug molecular graphs, Morgan fingerprints, and protein sequence information; S2: Model construction Construct an initial drug-target binding affinity prediction model, including a drug protein feature extraction model based on graph neural networks and multi-layer perceptron networks, and a prediction model based on a prediction network; In step S2, the specific processing process of the drug protein feature extraction model is as follows: S21: The drug molecule graph is fed into the GATv2 network to obtain 96-dimensional drug molecule graph features. The GATv2 network is a graph neural network. S22: Send the drug Morgan fingerprint into the multi-layer perceptron network to obtain the 96-dimensional drug Morgan fingerprint feature; S23: The drug Morgan fingerprint features and drug molecular graph features are concatenated and reduced in dimension after passing through the layer attention module to obtain a 96-dimensional drug feature. S24: The protein integer sequence is passed through a multi-scale convolutional neural network to obtain a 96-dimensional protein feature; In step S2, the processing of the prediction model is as follows: S25: Splicing the drug feature and the protein feature to obtain a spliced ​​feature; S26: sending the spliced ​​features into the prediction network to obtain the binding affinity prediction results; In step S26, the prediction network includes three linear transformation layers and one fully connected layer. A Relu activation function layer and a dropout layer are connected after each linear transformation layer. The binding index is obtained through the three linear transformation layers, and then processed through the fully connected layer to obtain the final drug-protein binding affinity index. S3: Model training The initial drug-target binding affinity prediction model is trained using the training set, and the final drug-target binding affinity prediction model is obtained after training; S4: DTA Forecast The final drug-target binding affinity prediction model is tested on the test set and the binding affinity index is output.

2. The method for predicting drug-target binding affinity based on graph neural network according to claim 1, characterized in that: In step S1, the specific processing process is as follows: S11: Divide the Davis and KIBA datasets into training and test sets according to the set ratio; S12: Use RDkit to preprocess the drug SMILES sequences of samples in the dataset into a graph with nodes and adjacency matrix, that is, to obtain drug molecular graph information; S13: Generate a Morgan fingerprint with a radius of 2 and a bit number of 1024 based on the drug SMILES sequence using RDkit, thus obtaining the Morgan fingerprint information; S14: The protein FASTA sequence of the sample in the dataset is converted into a corresponding integer sequence through the corresponding vocabulary, that is, the protein sequence information is obtained.

3. The method for predicting drug-target binding affinity based on graph neural network according to claim 1, characterized in that: In step S21, the drug molecule graph is represented as G = (V, E), where V represents the node of the drug atom, and E represents the set of edges connecting the drug nodes. Each node obtains an n-dimensional representation through Deepchem, including extracted atomic features, bond features, substructure features and topological structure features.

4. The method for predicting drug-target binding affinity based on graph neural network according to claim 1, characterized in that: In step S21, the GATv2 network adopts a dynamic attention mechanism to aggregate the corresponding graph representations obtained through the drug SMILES sequence to obtain drug molecular graph features. The aggregation method is as follows: e(h i ,h j )=a T LeakyRelu(W0[h i ||h j ])j∈N i Among them, e(h i ,h j ) represents the importance of neighbor node j to the original node i, N i Represented as the set of all neighbor nodes of node i is a learnable matrix.

5. The method for predicting drug-target binding affinity based on graph neural network according to claim 1, characterized in that: In step S22, the molecular structure is traversed through the drug Morgan fingerprint, and the environmental information around each atom is recorded to obtain a 1024-dimensional vector. The 1024-dimensional vector is then reduced to a 96-dimensional vector through a multi-layer perceptron network, and the supplementary structural information of the drug is output, that is, the drug Morgan fingerprint feature is obtained.

6. The method for predicting drug-target binding affinity based on graph neural network according to claim 1, characterized in that: In step S23, the drug molecular graph features and the drug Morgan fingerprint features are spliced ​​together using the layer attention module. The splicing method is as follows: h k =[W1h j ||(1-W1)h i ]。 7. The method for predicting drug-target binding affinity based on graph neural network according to claim 1, characterized in that: In step S24, the multi-scale convolutional neural network includes three convolution branches. Different convolution branches have different numbers of convolution layers, and the corresponding convolution kernel sizes are also different. The processing method of the multi-scale convolutional neural network is as follows: y=W2(m(F1(Q))||m(F2(Q))||m(F3(Q)) Among them, y represents protein features, Q is the input matrix obtained by protein integer sequence, and F i represents the i-th convolution branch, m represents the maximum pooling process, and W2 is a learnable matrix.

Images