A drug target prediction model and method based on graph neural network

Abstract

The present invention is applicable to the field of bioinformatics and provides a drug target prediction method based on graph neural networks, comprising the following steps: taking drug and target nodes as the core, automatically extracting their local subgraphs to capture the local contextual information of the nodes in complex biological networks; treating different types of edges as feature information and integrating them into the feature learning process, dynamically adjusting the contribution of different types of edges to node representation through the attention mechanism, introducing residual connections and normalization processing strategies to fuse node features; using a two-layer graph convolutional network to deeply mine the fused node features and calculate the potential interaction prediction score between the drug and the target. The present invention also provides a drug target prediction model based on graph neural networks. The present invention performs independent feature extraction and training for the predicted drugs and targets, effectively improving the accuracy and richness of node representation while maintaining the flexibility of the model.

Description

Technical Field

[0001] The present invention belongs to the field of bioinformatics, and in particular relates to a drug target prediction model and method based on graph neural network. Background Art

[0002] In practical applications, drug target prediction technology has demonstrated tremendous potential. For example, in the development of anticancer drugs, bioinformatics techniques have enabled the prediction of multiple new cancer targets, leading to the discovery of several compounds with potential therapeutic effects. These compounds, after further experimental verification, have demonstrated promising anticancer activity, providing new insights and approaches for anticancer drug development. Furthermore, drug target prediction technology can also be used to predict drug side effects and drug resistance, helping researchers optimize drug structures, reduce adverse drug reactions, and improve drug safety and efficacy. In the future, as technology continues to advance and improve, drug-target interaction prediction technology will become more efficient and accurate, bringing new breakthroughs to the development of new drugs. On the one hand, with the continuous development of bioinformatics, computer science, and artificial intelligence technologies, drug target prediction methods will become more diverse and intelligent. On the other hand, with the continuous accumulation and sharing of biological data, researchers will be able to use more experimental data to train and validate models, improving the reliability and accuracy of predictions.

[0003] DTI (drug-target interaction) prediction, a key task in analyzing the potential connection between drugs and targets in organisms, is usually conceived as a binary classification problem, aiming to determine whether a given drug-target pair (DTP) interacts. However, in practical applications, labeled DTP data is extremely scarce and expensive to obtain. This bottleneck severely restricts the efficiency and depth of the model's learning of DTP relationships. To overcome this limitation, heterogeneous biological data, with its rich multi-source information and multi-dimensional perspectives, has opened up new avenues for modeling the complex relationship between drugs and targets. These data not only cover the structural characteristics of drugs and the biological properties of targets, but also involve information on disease associations, drug side effects, and other aspects, providing strong support for a comprehensive understanding of DTI mechanisms.

[0004] Currently, research on DTI prediction using biological heterogeneous data is showing a diversified trend, mainly including three major directions:

[0005] The first is a similarity-based approach, which infers potential interactions by calculating the similarity between drugs and targets. Similarity-based methods have a solid theoretical foundation in the field of drug-target interaction (DTI) prediction. Their core assumption is that, in biology and pharmacology, drugs with similar functions or structures tend to bind to the same targets. Conversely, targets with similar properties may also attract drugs with the same or similar functions. This assumption has driven the development of a three-step strategy aimed at more accurately predicting drug-target interactions.

[0006] However, although these strategies have demonstrated their effectiveness in many aspects, they also expose some limitations. Specifically, these methods are unable to cope with complex data relationships. In real biological systems, drugs with very different chemical structures may bind to the same target due to specific molecular mechanisms. This non-intuitive binding pattern is often ignored by similarity-based methods. In addition, the practice of separating feature learning from prediction tasks may also result in the learned representation not being fully suitable for subsequent prediction tasks, because the goal of feature learning may deviate from the optimal solution of the prediction task. This suboptimal solution limits the accuracy and generalization ability of DTI predictions, prompting researchers to continuously explore more integrated and intelligent prediction methods.

[0007] The second approach is knowledge graph (KG)-based methods, which construct complex networks consisting of drugs, targets, and their relationships, leveraging the graph's structural information to aid prediction. Knowledge graph (KG)-based technologies are increasingly being used in the biomedical field. They construct information-rich biological heterogeneous graphs by representing biological entities (such as genes, proteins, drugs, and diseases) as nodes and the complex relationships between these entities as edges. To effectively mine knowledge and perform reasoning within this vast graph, researchers have introduced knowledge graph embedding models, such as TransE, DistMult, and ComplEx, which aim to map graph nodes and relationships into a low-dimensional vector space, simplifying the processing of complex relationships and facilitating computational feasibility. However, while these shallow embedding models have made significant progress in representation learning for biological heterogeneous graphs, they still have some shortcomings in modeling component relationships. Specifically, relationships within biological heterogeneous graphs are often highly complex and diverse, including but not limited to interactions between drugs and targets, associations between genes and diseases, and interaction networks between proteins. These relationships are not only huge in number, but also often exhibit multi-level and multi-scale characteristics, making it difficult for simple embedding models to fully capture their intrinsic structure and semantic information.

[0008] The third is the method based on graph neural network (GNN), which can automatically learn the representation of nodes in the graph and capture the deep interaction patterns between drugs and targets, thereby achieving more accurate DTI prediction. When processing complex biological heterogeneous data based on the graph neural network (GNN) method, the existing technology integrates multi-source biological data into multiple interrelated drug and target networks. These networks not only cover drug-target interactions, but may also include drug-drug similarity, target-target similarity, drug-side effects, target-disease associations and other types of information. By adopting models based on graph convolutional networks (GCN) or graph attention networks (GAT), GNN can effectively extract deep representations of drugs and targets from these complex networks. These representations are rich in topological structure and semantic information between nodes, providing strong support for subsequent DTI (drug-target interaction) predictions.

[0009] IMCHGAN and SGCL-DTI are two representative models in this field. They introduce the concept of meta-paths to mine richer drug (or target) representations from heterogeneous graphs. A meta-path is a path pattern defined on a heterogeneous graph that can reveal complex relationships between different node types. However, these two models have certain limitations when utilizing meta-paths. They each independently extract meta-paths from each individual network, ignoring the cross-information and potential synergistic effects between networks. In addition, they fail to fully utilize contextual information when processing meta-paths, namely the attribute and structural information of the nodes on the path traversed by the meta-path, which is crucial for a deeper understanding of the interaction mechanism between drugs and targets. In contrast, models such as MHGNN and MHTAN-DTI pay more attention to the contextual information of meta-paths when extracting feature representations. They not only aggregate features along the meta-paths, but also consider the characteristics of each node on the path and the interactions between nodes, thereby more comprehensively capturing the complex relationship between drugs and targets. However, in the dual-channel feature extraction process of the MHGNN model, the selection of its input meta-paths relies on manual design, which not only increases the subjectivity and complexity of the model, but also may limit the model's generalization ability on different datasets. In addition, there is still room for further improvement in the algorithm efficiency and implementation details of MHGNN to meet the needs of large-scale biological heterogeneous graph processing. Summary of the Invention

[0010] The purpose of the embodiments of the present invention is to provide a drug target prediction method based on graph neural network, aiming to solve the problems raised in the above background technology.

[0011] The embodiment of the present invention is implemented as follows: a drug target prediction method based on graph neural network, comprising the following steps:

[0012] Taking drug and target nodes as cores, local subgraphs are automatically extracted to capture the local contextual information of nodes in complex biological networks.

[0013] Different types of edges are considered as feature information and integrated into the feature learning process. The contribution of different types of edges to node representation is dynamically adjusted through the attention mechanism. Residual connections and normalization processing strategies are introduced to fuse node features.

[0014] A two-layer graph convolutional network is used to deeply mine the fused node features and calculate the potential interaction prediction score between drugs and targets.

[0015] Preferably, the step of automatically extracting local subgraphs with drug and target nodes as the core to capture local context information of the nodes in the complex biological network is specifically as follows:

[0016] Construct a biologically heterogeneous graph that includes multiple types of nodes, such as drugs, targets, diseases, and side effects, and each type of node represents different feature information;

[0017] A dual-channel learning strategy is adopted to learn the features of the two nodes, drug and target. In each channel, the drug node (D) and the target node (T) are taken as the central nodes, and subgraphs D and T are extracted from the complete heterogeneous graph as the starting points of the dual-channel representation learning.

[0018] Preferably, the step further includes: setting the initial representation of nodes such as drugs, targets, diseases and side effects to one-hot encoding form, using a linear transformation layer with a bias term, and applying it independently to each node type to map their features into a common, unified feature space.

[0019] Preferably, the step further includes: defining the heterogeneous graph as Where V represents the set of nodes, E represents the set of edges, each node v has a type φ(v), each edge e has a type Ψ(e), and the possible sets of node types and edge types are expressed as and When |Tv|=|Te|=1, the graph degenerates into a normal homogeneous graph.

[0020] Preferably, the steps of treating different types of edges as feature information and integrating them into the feature learning process, dynamically adjusting the contribution of different types of edges to node representation through the attention mechanism, introducing residual connections and normalization processing strategies, and fusing node features are specifically as follows:

[0021] Learnable edge type feature embedding: In each layer of the graph attention network, each edge type Ψ(e)∈T(e) is assigned a specific dl-dimensional embedding feature, expressed as Taking both node embeddings and edge type embeddings into account, the attention score α is calculated as:

[0022]

[0023] Where Ψ(<i, j>) represents the type of edge with nodes i and j as endpoints, W r is a learnable matrix for type embedding conversion;

[0024] Residual connection: A node residual connection is added between cross-layer node representations, and the node aggregation on the lth layer is represented as:

[0025]

[0026] in, is the attention weight about edge <i, j>, σ is an activation function (default is ELU[]);

[0027] When the dimension changes at layer l, a learnable linear transformation needs to be added:

[0028]

[0029] According to formula (2), K independent attention mechanisms are performed and their results are connected as the final representation. The formula is parsed as:

[0030]

[0031]

[0032]

[0033] In the above formula, || represents the splicing operation, Indicates that through the k-th linear transformation The calculated attention scores are averaged in the last layer (layer L), that is:

[0034]

[0035] L2 normalization: Through L2 normalization, all features are scaled to the same magnitude:

[0036]

[0037] Among them, f i represents the output feature embedding of node i, The final expression from Equation (7).

[0038] Preferably, the step of using a two-layer graph convolutional network to deeply mine the fused node features and calculate the potential interaction prediction score between the drug and the target is specifically:

[0039] Construct a DTP correlation graph to discover the potential relationships between DTPs, where each node is a DTP;

[0040] By concatenating the drug feature representation and the target feature representation, the comprehensive representation fdt of each node in the reconstructed DTP graph is obtained, that is, fdt = fd||ft, where d∈VD and t∈VT;

[0041] When processing a batch of DTPs, the representation z of each DTP calculated in the above way is combined into a representation matrix Z. In this matrix, the b-th row of Z is the eigenvector of the b-th DTP in the batch. The adjacency matrix A is constructed from the DTP representation and is expressed as:

[0042] A ij =softmax((ZZ T )) ij (9);

[0043] Among them, A ij is the element in the i-th row and j-th column of A, which represents the normalized similarity between the i-th drug and the j-th target in the same batch of data. It can be seen that the matrix A is a symmetric matrix. ij The larger the value, the stronger the association between the i-th drug and the j-th target DTP;

[0044] The DTP matrix Z is input into the two-layer GCN, and the final output y is used to predict DTI. The calculation process of y is:

[0045] y = A(ReLU(AZW1))W2 (10);

[0046] Where W1 and W2 represent the learnable parameters of the first and second layers of GCN respectively;

[0047] Binary cross entropy loss is used to evaluate the difference between the predicted DTI and the true DTI, as follows:

[0048]

[0049] Among them, B represents the number of DTPs in a batch training, y b represents the actual value of DTP. When D and P interact, y b =1, otherwise, y b =0,y′ b Represents the predicted value of DTP, which is normalized to y′b ∈[0, 1].

[0050] Another object of the present invention is to provide a drug target prediction model based on a graph neural network, which is used to implement the above-mentioned drug target prediction method based on a graph neural network, including:

[0051] A subgraph construction module is used to automatically extract local subgraphs of drug and target nodes, respectively, to capture the local contextual information of nodes in complex biological networks.

[0052] The attention fusion module is used to integrate different types of edges into the feature learning process as feature information. The attention mechanism dynamically adjusts the contribution of different types of edges to node representation, introduces residual connections and normalization processing strategies, and fuses node features.

[0053] The prediction analysis module is used to deeply mine the fused node features and calculate the potential interaction prediction score between drugs and targets.

[0054] The present invention provides a drug target prediction method based on a graph neural network. This method addresses the high sensitivity and poor flexibility of manually selected meta-paths during node aggregation. It abandons traditional meta-path dependency and instead adopts a meta-path-free node aggregation strategy. By constructing a dual-channel subgraph, it performs independent feature extraction and training for the two predicted core nodes (drugs and targets). This effectively improves the accuracy and richness of node representation while maintaining model flexibility.

[0055] An embodiment of the present invention provides a drug target prediction model based on a graph neural network. By removing the dependence on meta-paths and adopting a dual-channel sub-graph learning and attention fusion mechanism, it provides a more flexible, efficient and powerful solution for drug-target interaction prediction. It is of great significance in computer-assisted drug discovery and reuse. In particular, the system does not require manual setting of meta-paths, has higher flexibility and universality, can automatically adapt to different data sets and task requirements, reduces human intervention and design complexity, and can help many people who do not understand the relevant fields of the data sets used to get started faster. In addition, the accuracy of the embodiment of the present invention is better than that of common drug target prediction models. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 A prediction flow chart of a drug target prediction module based on a graph neural network is provided as an embodiment of the present invention;

[0057] Figure 2 A schematic diagram of drug / target feature representation learning provided by an embodiment of the present invention;

[0058] Figure 3 ROC curves of the model provided by the embodiment of the present invention and the MHGNN and LDMLP-DTI models on specific tasks:

[0059] Figure 4 PR curves of the model provided in this embodiment of the present invention and the MHGNN and LDMLP-DTI models on specific tasks. DETAILED DESCRIPTION

[0060] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0061] The specific implementation of the present invention is described in detail below with reference to specific embodiments.

[0062] like Figure 1 FIG. 1 is a flow chart of a drug target prediction module based on a graph neural network according to an embodiment of the present invention, including:

[0063] Subgraph construction module: This module automatically extracts local subgraphs centered around drug and target nodes to capture the local contextual information of nodes in complex biological networks.

[0064] Attention Fusion Module: This module treats different types of edges as feature information and integrates it into the feature learning process. It dynamically adjusts the contribution of different types of edges to node representation through the attention mechanism. It also introduces residual connections and normalization processing strategies to effectively alleviate the vanishing gradient problem in deep networks and improve the learning efficiency and stability of the model.

[0065] Prediction and Analysis Module: This module uses a two-layer graph convolutional network (GCN) to deeply mine the fused node features and calculate the potential interaction prediction score between drugs and targets. It not only simplifies the complexity of traditional path dependence, but also significantly improves the accuracy and robustness of the prediction.

[0066] Specifically, a drug target prediction method based on graph neural network includes the following steps:

[0067] (1) Drug / target subgraph construction and initialization:

[0068] Based on heterogeneous biological data from diverse public database resources, a biological heterogeneous graph is first constructed, which covers multiple types of nodes such as drugs, targets, diseases, and side effects. Each type of node represents different characteristic information. To fully capture and preserve the specific attributes of drugs and targets, a dual-channel learning strategy is adopted to learn the characteristics of these two types of nodes separately. This ensures that when processing complex biological information, it can accurately reflect the essential differences and intrinsic connections between different entities. Unlike other models, this model breaks away from the inefficiency of meta-path selection in most existing models. In each channel, the drug node (D) and the target node (T) are used as the central nodes, and subgraphs D and T are extracted from the complete heterogeneous graph respectively as the starting points for dual-channel representation learning.

[0069] Specifically, taking Subgraph-D as an example, all edges with node D as the endpoint are extracted from the initial graph and reconstructed into a new heterogeneous graph. This new heterogeneous graph is used to further learn the feature representation of node D. Since there are no TS type edges in the dataset, there are only three types of nodes in Subgraph-T, namely D, T, and I.

[0070] In order to effectively integrate and process different types of nodes such as drugs, targets, diseases, and side effects in a unified framework, the initial representation of these nodes is first set to one-hot encoding form (such as Figure 2 As shown in the figure, given that different types of nodes may have different initial feature dimensions, a linear transformation layer with a bias term (also called a fully connected layer) is used independently for each node type to map their features into a common, unified feature space. The parameters of the linear layer in this mapping process will be continuously optimized during the subsequent training of the network model to ensure that the interactions and associations between nodes can be effectively modeled and analyzed within a unified framework.

[0071] A heterogeneous graph is defined as Where V represents the set of nodes, E represents the set of edges, each node v has a type φ(v), each edge e has a type Ψ(e), and the possible sets of node types and edge types are expressed as and When |Tv|=|Te|=1, the graph degenerates into a normal homogeneous graph;

[0072] (2) Drug / target feature representation learning, such as Figure 2 As shown:

[0073] Inspired by the advantages of the Graph Attention Network (GAT) model in applying to complex heterogeneous graph neural networks, a simple and effective heterogeneous graph representation learning method is proposed. It uses the graph attention mechanism as the backbone and incorporates three elements applicable to the model of the embodiment of the present invention: learnable edge type feature embedding, residual connection and L2 normalization.

[0074] Learnable edge type feature embedding: GAT does not consider the diversity of nodes or edges and is not directly applicable to heterogeneous graphs (graphs containing multiple types of nodes and edges). To make up for this deficiency and enhance the adaptability of GAT to heterogeneous graphs, an extension method is proposed to integrate edge type information into the attention calculation process. Specifically, in each layer of the graph attention network, a specific dl-dimensional embedding feature is assigned to each edge type Ψ(e)∈T(e), expressed as These embedding vectors can capture the unique semantic or structural information contained in different edge types. When considering both node embedding and edge type embedding, the attention score α is calculated as:

[0075]

[0076] Among them, Ψ(<i,j> ) represents the type of edge with nodes i and j as endpoints, W r is a learnable matrix for type embedding conversion;

[0077] Residual connection: To alleviate problems such as gradient vanishing and gradient exploding, a node residual connection is added between cross-layer node representations, providing a path to bypass nonlinear transformations, allowing gradients to be directly passed to deeper layers. At the same time, through carefully designed pre-activation implementation, residual connections play a powerful role in GNNs.

[0078] The node aggregation on layer l is expressed as:

[0079]

[0080] in, It's about the edge<i,j> The attention weight, σ is an activation function (default is ELU[]);

[0081] When the dimension changes at layer l, a learnable linear transformation needs to be added:

[0082]

[0083] Similar to GAT, a multi-head attention mechanism is used to enhance the expressive power of the model. Specifically, K independent attention mechanisms are executed according to formula (2), and their results are connected as the final representation. The formula is parsed as follows:

[0084]

[0085]

[0086]

[0087] In the above formula, || represents the splicing operation, Indicates that through the k-th linear transformation The calculated attention score;

[0088] In the last layer (layer L), the obtained representations are averaged, that is:

[0089]

[0090] L2 normalization: Through L2 normalization, all features can be scaled to the same magnitude, making the model converge more easily and faster:

[0091]

[0092] Among them, f i represents the output feature embedding of node i, The final expression from formula (7);

[0093] (3) Prediction of drug-target interactions:

[0094] Since drugs with similar structures tend to interact with targets with similar structures, it can be inferred that DTPs with similar structural drugs or targets have stronger correlations than DTPs without similar structures. A DTP correlation graph is constructed to discover the potential relationship between DTPs, where each node is a DTP. By concatenating the drug feature representation and the target feature representation, the comprehensive representation fdt of each node in the reconstructed DTP graph is obtained, that is, fdt = fd||ft, where d∈VD and t∈VT;

[0095] When processing a batch of DTPs, the representation z of each DTP calculated in the above way is combined into a representation matrix Z. In this matrix, the b-th row of Z is the eigenvector of the b-th DTP in the batch. The adjacency matrix A is constructed from the DTP representation and is expressed as:

[0096] A ij =softmax((ZZ T )) ij (9);

[0097] Among them, A ijis the element in the i-th row and j-th column of A, which represents the normalized similarity between the i-th drug and the j-th target in the same batch of data. It can be seen that the matrix A is a symmetric matrix. ij The larger the value, the stronger the association between the i-th drug and the j-th target DTP;

[0098] The DTP matrix Z is input into the two-layer GCN, and the final output y is used to predict DTI. The calculation process of y is:

[0099] y = A(ReLU(AZW1))W2 (10);

[0100] Where W1 and W2 represent the learnable parameters of the first and second layers of GCN respectively;

[0101] Binary cross entropy (BCE) loss is used to evaluate the difference between the predicted DTI and the true DTI, as follows:

[0102]

[0103] Among them, B represents the number of DTPs in a batch training, y b represents the actual value of DTP. When D and P interact, y b =1, otherwise, y b =0,y′ b Represents the predicted value of DTP, which is normalized to y′ b ∈[0, 1].

[0104] The performance of the model proposed in the embodiment of the present invention in drug-target interaction (DTI) prediction is compared with six other models in the prior art (NeoDTI, DTI-MGNN, SGLC-DTI, HAN, and MGNN), as follows:

[0105] (1) Dataset:

[0106] The first dataset, Hetero-A, was collected through DTINet. This dataset contains four different types of biological data and six different types of relationships between them. The types and numbers of nodes are: 708 drugs (D), 1512 targets (T), 5603 diseases (I), and 4192 side effects (S). The types and numbers of edges are: 1923 D-T, 10036 D-D, 199214 D-I, 80164 D-S, 7363 T-T, and 1596745 TI.

[0107] Because insufficient positive samples can lead to false negatives in data modeling and model training, which can result in large discrepancies between predicted and actual results, we chose to further train the model using a more complete dataset to optimize model prediction accuracy. Using the MHGNN, we obtained a new dataset, Hetero-B, which is an update of dataset A. The node and edge types in dataset B are the same as those in dataset A, with only a slight change in number. The node types and numbers are: 2214 drugs (D), 1968 targets (T), 7025 diseases (I), and 3935 side effects (S). The edge types and numbers are: 8750 DTs, 1091870 DDs, 542970 DIs, 104629 D-Ss, 456592 TTs, and 2922064 T-Is.

[0108] The model of the embodiment of the present invention was evaluated on two biologically heterogeneous datasets, Hetero-A and Hetero-B.

[0109] (2) Existing technology model:

[0110] NeoDTI first applies GCN to each independent relational network to capture the local and global characteristics of each node (drug or target), as well as the complex interaction patterns between them. These learned representations are then used as features, combined with known drug-target interaction data, to train an Inductive Matrix Completion (IMC) model to predict unknown drug-target interactions (DTI) using the known representation features.

[0111] The DTI-MGNN model designs two parallel graph attention network branches, focusing on capturing interactions of varying strengths between nodes in the topological graph and feature graph, respectively. This design allows the model to discern and learn information flows from different data sources, more accurately reflecting the complex relationships between nodes. The model also introduces a graph convolutional network with a shared weight matrix to extract and integrate shared information between the two graphs, promoting a deep fusion of topological structure and semantic features.

[0112] IMCHGAN uses a two-level graph attention method to analyze the interaction between different nodes and meta-paths, and obtains the potential feature representations of drugs and targets from the DTI heterogeneous network. Finally, the inductive matrix completion (IMC) model predicts the score and obtains the interaction relationship between drugs and targets.

[0113] The core concept of the SGCL-DTI model is to optimize the feature representations of drugs and targets by leveraging the rich information in graph-structured data and to enhance the discrimination between these representations through contrastive learning. First, graph convolutional networks are used to extract feature representations of drugs and targets from meta-path-based neighborhoods. Subsequently, topological and semantic graphs are constructed, with drug-target pairs as core nodes. Using contrastive learning, the feature representations of drugs and targets are further refined and optimized by comparing and distinguishing the information contained in these two graphs.

[0114] The HAN model, based on meta-paths and an attention mechanism, employs a hierarchical structure, processing information within and between meta-paths separately to obtain feature representations of central nodes. Since the HAN model is based on node classification and clustering tasks, we designed two parallel HAN-based pipelines to learn feature representations for drugs and targets, respectively, based on the task objectives. Finally, the outputs of the two pipelines are concatenated and passed through two linear layers to obtain the final prediction result.

[0115] MAGNN is an improved model based on HAN. While HAN only aggregates information from two terminal nodes when processing information within a meta-path, MAGNN aggregates representations of all nodes within a meta-path instance. Furthermore, MAGNN establishes a DTP correlation graph, assigning different weights to different DTP associations, further optimizing the feature learning module and the DTI prediction module.

[0116] (3) Comparison results:

[0117] The performance of each model is measured using multiple evaluation metrics, including AUC, AUPR, Precision, Recall, F1, and MCC. The performance data for each model on different subsets of cross-validation are presented as mean ± standard deviation. The results are shown in Tables 1 and 2.

[0118] Table 1 Comparison results on the Hetero-A dataset

[0119]

[0120]

[0121] Table 2 Comparison results on the Hetero-B dataset

[0122]

[0123] According to Tables 1 and 2, it can be seen that the HAN and MHGNN models, especially the model of the embodiment of the present invention, have shown high performance in multiple evaluation indicators. In particular, the model of the embodiment of the present invention has achieved high levels in indicators such as AUC, AUPR, Precision and F1, demonstrating its superiority in the DTI prediction task. In addition, the model of the embodiment of the present invention also performs well in the MCC (Matthews Correlation Coefficient) indicator, which shows that the correlation between its prediction results and the actual situation is very high. In comparison, although the NeoDTI model also performs well in some indicators, it is relatively weak in comprehensive indicators such as F1 and MCC; and the SGLC-DTI model has low performance in multiple indicators, which may indicate that the model has certain limitations when handling the current DTI prediction task. In general, the embodiment of the present invention has excellent performance and has potential application and research value.

[0124] The performance of the model proposed in this embodiment of the present invention is compared with that of the MHGNN and LDMLP-DTI models on specific tasks by plotting their precision-recall curves:

[0125] Precision and recall are two key indicators for evaluating classifier performance. Precision reflects the proportion of samples predicted by the model to be positive that are actually positive, while recall reflects the proportion of samples that are actually positive that are correctly predicted by the model to be positive. Figure 3 、 Figure 4 As shown in the figure, it can be seen that the curve of the model of the embodiment of the present invention is closest to the upper left corner of the coordinate axis (that is, the ideal point where the precision is 1.0 and the recall rate is 1.0), and its corresponding area (area=0.96) is the largest, which means that the model maintains a high recall rate while also maintaining a high precision, and has the best performance among the three models.

[0126] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A drug target prediction method based on graph neural network, characterized in that: The following steps are involved: Taking drug and target nodes as cores, local subgraphs are automatically extracted to capture the local contextual information of nodes in complex biological networks. Different types of edges are considered as feature information and integrated into the feature learning process. The contribution of different types of edges to node representation is dynamically adjusted through the attention mechanism. Residual connections and normalization processing strategies are introduced to fuse node features. A two-layer graph convolutional network is used to deeply mine the fused node features and calculate the potential interaction prediction score between the drug and the target; The steps of automatically extracting local subgraphs with drug and target nodes as the core to capture the local context information of the nodes in the complex biological network are as follows: Construct a biologically heterogeneous graph that includes multiple types of nodes, including drugs, targets, diseases, and side effects, and each type of node represents different feature information; A dual-channel learning strategy is used to learn the features of the two nodes, drug and target. In each channel, the drug node and target node are used as the central nodes, and subgraphs D and T are extracted from the complete heterogeneous graph as the starting points of the dual-channel representation learning. The steps further include: defining the heterogeneous graph as , where V represents the node set, E represents the edge set, and each node v has a type , each edge e has a type , the sets of possible node types and edge types are represented as and ,when When , the graph degenerates into a common homogeneous graph; The steps of treating different types of edges as feature information and integrating them into the feature learning process, dynamically adjusting the contribution of different types of edges to node representation through the attention mechanism, introducing residual connections and normalization processing strategies, and fusing node features are as follows: Learnable edge type feature embedding: In each layer of the graph attention network, for each edge type Assign a specific dimensional embedding features, expressed as , considering both node embedding and edge type embedding, the attention score α is calculated as: （1）； in, Indicates the type of edge with nodes i and j as endpoints, W r is a learnable matrix for type embedding conversion; Residual connection: Add a node residual connection between cross-layer node representations. The node aggregation on a layer is represented as: （2）； in, It's about the edge The attention weight, is an activation function; When the dimension When the layer changes, a learnable linear transformation needs to be added: （3）； According to formula (2), K independent attention mechanisms are performed and their results are connected as the final representation. The formula is parsed as: （4）； （5）； （6）； In the above formula, Represents a splicing operation, Show through the The attention scores calculated by linear transformation are averaged in the last layer, the Lth layer, that is: （7）； L2 Normalization: Through L2 normalization, all features are scaled to the same magnitude.

2. The drug target prediction method based on graph neural network according to claim 1, characterized in that: The steps also include setting the initial representation of drug, target, disease and side effect nodes to one-hot encoding form, using a linear transformation layer with a bias term, applied independently to each node type to map their features into a common, unified feature space.

3. The drug target prediction method based on graph neural network according to claim 1, characterized in that: The steps of using a two-layer graph convolutional network to deeply mine the fused node features and calculate the potential interaction prediction score between the drug and the target are specifically as follows: Construct a DTP correlation graph to discover the potential relationships between DTPs, where each node is a DTP; By splicing the drug feature representation and the target feature representation, a comprehensive representation of each node in the reconstructed DTP graph is obtained. ,Right now ,in , ; Combine the DTPs in each batch into a representation matrix , the first The row is the first row in the batch The characteristics of DTP are obtained through the matrix The adjacency matrix of the DTP graph is constructed as follows: （9）； in, is the element in the i-th row and j-th column of A, which represents the normalized similarity between the i-th drug and the j-th target in the same batch of data. It can be seen that the matrix A is a symmetric matrix. The larger the value, the stronger the association between the i-th drug and the j-th target DTP; The DTP matrix Z is input into the two-layer GCN, and the final output y is used to predict DTI. The calculation process of y is: （10）； Where W1 and W2 represent the learnable parameters of the first and second layers of GCN respectively; Binary cross entropy loss is used to evaluate the difference between the predicted DTI and the true DTI, as follows: （11）； in, Indicates the number of DTPs in a batch training. Indicates the actual value of DTP. When D and P interact, ,otherwise, , Represents the predicted value of DTP, which is normalized to .

4. A drug target prediction model based on a graph neural network, used to implement the drug target prediction method based on a graph neural network according to any one of claims 1 to 3, characterized in that: include: A subgraph construction module is used to automatically extract local subgraphs of drug and target nodes, respectively, to capture the local contextual information of nodes in complex biological networks. The attention fusion module is used to integrate different types of edges into the feature learning process as feature information. The attention mechanism dynamically adjusts the contribution of different types of edges to node representation, introduces residual connections and normalization processing strategies, and fuses node features. The prediction analysis module is used to deeply mine the fused node features and calculate the potential interaction prediction score between drugs and targets.

Images