Drug-target interaction prediction method and device based on deep neural network

Abstract

The invention provides a drug-target interaction prediction method and device based on a deep neural network, and relates to the technical field of drug-target interaction prediction. The method comprises the following steps: calculating similarities among different elements in a drug-related network and a protein-related network to obtain an initial drug similarity matrix and an initial protein similarity matrix; preprocessing the initial drug similarity matrix and the initial protein similarity matrix to obtain a target drug similarity matrix and a target protein similarity matrix; splicing the target drug similarity matrix and the target protein similarity matrix to obtain a DPP; determining a feature adjacency matrix of the DPP and a topological adjacency matrix of the DPP according to the DPP; obtaining a feature matrix according to the initial drug similarity matrix and the initial protein similarity matrix; and obtaining a prediction result according to the feature adjacency matrix, the topological adjacency matrix and the feature matrix. According to the method, the network learning effect can be improved, and the drug-target interaction can be better predicted.

Description

Technical Field

[0001] The present invention relates to the technical field of drug-target interaction prediction, and in particular to a drug-target interaction prediction method and device based on a deep neural network. Background Art

[0002] Determining the interactions between known drugs and their targets is an important part of medicine, as it helps in drug discovery and repositioning. However, identifying drug-target interactions (DTIs) through in vitro and in vivo experiments is time-consuming and expensive. Therefore, more and more researchers have begun to explore computational methods for DTI prediction, which can not only reduce losses in the drug discovery process, but also provide guidance for research on drug repositioning, drug resistance, and side effect prediction.

[0003] In recent years, there have been some good results related to drug-target binding affinity prediction, such as DeepDTA (Ozturk et al. 2018), DeepAffinity (Karimi et al. 2019), GraphDTA (Nguyen et al. 2021), DeepDTAF (Wang et al. 2021), etc. Among them, DeepDTA uses two deep convolutional neural networks (CNNs) to encode drug and protein sequences respectively, and achieved good results at the time; DeepAffinity uses CNN to encode protein sequences, uses an unsupervised sequence to sequence model (Seq2Seq; Sutskever et al. 2014) to encode drugs, and then trains a regression model to predict drug-target binding affinity; GraphDTA uses graph neural networks (GNN) to encode drugs, and uses CNN to encode proteins, and then connects the encoded features to predict drug-target binding affinity; DeepDTAF uses protein-binding pockets to directly process target binding, and then uses the dilated convolution method to process multi-scale DTP features, and then connects features of different scales to predict drug-target binding affinity.

[0004] Chinese patent CN202210418625.4 discloses a drug target-affinity prediction method based on representation learning and graph neural network, which uses CNN and GNN to extract features from the two-dimensional structure of drugs / proteins, and GNN to extract features from the three-dimensional structure of drugs / proteins. After fusing the extracted feature information, the drug target affinity value is predicted through the fully connected layer. Chinese patent CN202110028123.6 discloses a drug-target binding affinity prediction method based on graph convolutional neural network, which first uses Bi-GRU network to extract the contextual association relationship features of drugs, and then uses graph convolutional neural networks (GCN), graph attention neural network (GAT), graph isomorphism network (GIN) and graph convolution attention joint network (GCN-GAT) to jointly model drug molecules. Then the long short-term memory model (LSTM) is used to extract the feature system of the protein. Finally, the drug-target binding affinity is predicted through the fully connected layer. Chinese patent CN116825183A discloses a drug-target binding affinity prediction method based on TransVAE. First, drug information, protein information, and drug-target binding affinity information are extracted from the database. After data preprocessing and noise reduction, the drug feature matrix and protein feature matrix are obtained respectively. Then, a TransVAE-based drug-target binding affinity prediction model (the model is mainly divided into a data encoding module, a feature fusion module, and a prediction module) is used, including a variational autoencoder, an attention network (Transformer), and attention pooling to obtain the fused features; finally, the fused features are input into the prediction module for drug-target binding affinity prediction.

[0005] However, the aforementioned research methods may cause loss of graph structure information due to layer-by-layer stacking, have low computational efficiency, and have limited processing capabilities for large-scale data, resulting in insufficient generalization capabilities of the model, which may greatly affect the final performance. Summary of the invention

[0006] The purpose of the present invention is to provide a drug-target interaction prediction method and device based on deep neural network to address the deficiencies in the above-mentioned prior art, so as to improve the learning effect of deep neural network and better predict drug-target interaction.

[0007] To achieve the above purpose, the technical solution adopted in the embodiment of the present application is as follows:

[0008] In a first aspect, an embodiment of the present application provides a drug-target interaction prediction method based on a deep neural network, comprising: calculating the similarity between different elements in a drug-related network and a protein-related network to obtain an initial drug similarity matrix and an initial protein similarity matrix; preprocessing the initial drug similarity matrix and the initial protein similarity matrix to obtain a target drug similarity matrix and a target protein similarity matrix; concatenating the target drug similarity matrix and the target protein similarity matrix to obtain a DPP; extracting features from the DPP using a KNN algorithm to obtain a feature adjacency matrix of the DPP; processing the DPP according to preset rules to obtain a topological adjacency matrix of the DPP; and using a graph embedding model to perform feature extraction on the DPP. The block performs feature extraction on the initial drug similarity matrix and the initial protein similarity matrix to obtain multi-view information; uses a feature extraction module to extract features from the multi-view information to obtain a feature matrix; obtains a feature graph and a topological graph according to the feature adjacency matrix, the topological adjacency matrix, and the feature matrix; inputs the feature graph and the topological graph into a graph attention network for feature extraction to obtain a first target feature and a second target feature; inputs the feature graph and the topological graph into a graph convolutional network for feature extraction to obtain a third target feature; inputs the first target feature, the second target feature, and the third target feature into an attention network for feature fusion to obtain a fused feature; inputs the fused feature into a fully connected layer to obtain a prediction result.

[0009] In one embodiment, the preprocessing of the initial drug similarity matrix and the initial protein similarity matrix to obtain a target drug similarity matrix and a target protein similarity matrix includes: using the RWR algorithm to process the initial drug similarity matrix and the initial protein similarity matrix to obtain an intermediate drug similarity matrix and an intermediate protein similarity matrix; using the DAE algorithm to denoise and reduce the dimension of the intermediate drug similarity matrix and the intermediate protein similarity matrix to obtain the target drug similarity matrix and the target protein similarity matrix.

[0010] In one embodiment, the preset rule is: if two DPPs have the same drug or protein, then the two DPPs are considered to have the same topological edge.

[0011] In one embodiment, the use of a graph embedding module to perform feature extraction on the initial drug similarity matrix and the initial protein similarity matrix to obtain multi-view information includes: using the graph embedding module to perform feature extraction on the initial drug similarity matrix and the initial protein similarity matrix to obtain an embedding matrix; the graph embedding module includes a convolutional network, and the convolutional network is composed of multiple convolutional layers; learning corresponding weighted parameters for the embedding matrix to obtain the multi-view information.

[0012] In one embodiment, the graph embedding module is used to perform feature extraction on the initial drug similarity matrix and the initial protein similarity matrix to obtain an embedding matrix, including: initializing the initial drug similarity matrix and the initial protein similarity matrix using random initialization weights and random initialization biases to obtain an initialization matrix; updating the initialization matrix using a SUM function to obtain an update matrix; and performing normalization, activation, and regularization processing on the update matrix in sequence to obtain the embedding matrix.

[0013] In one embodiment, obtaining a feature graph and a topological graph based on the feature adjacency matrix, the topological adjacency matrix, and the feature matrix includes: combining the feature adjacency matrix and the feature matrix to obtain the feature graph; combining the topological adjacency matrix and the feature matrix to obtain the topological graph.

[0014] In one embodiment, the calculating the similarity between different elements in the drug-related network and the protein-related network to obtain the initial drug similarity matrix and the initial protein similarity matrix includes: calculating the Jaccard similarity coefficient between different elements in the drug-related network and the protein-related network to obtain the initial drug similarity matrix and the initial protein similarity matrix.

[0015] In one embodiment, the drug-related network includes a drug-drug relationship network, a drug-related disease network, a drug and side effect network, and a drug-chemical structure similarity network; the drug-drug relationship network is used to represent the interaction between different drugs; the drug-related disease network is used to represent the relationship between drugs and diseases; the drug and side effect network is used to represent the relationship between drugs and side effects; the drug-chemical structure similarity network is used to represent the relationship between the chemical structures of different drugs; the protein-related network includes a protein-related disease network, a protein-protein relationship network, and a protein sequence similarity network; the protein-related disease network is used to represent the relationship between proteins and diseases; the protein-protein relationship network is used to represent the interaction between different proteins; the protein sequence similarity network is used to represent the relationship between different protein sequences.

[0016] In one embodiment, the feature extraction module includes an ECR network; the ECR network includes an ECR input layer, an ECR convolution layer, an ECR pooling layer, a CECJC layer, and an ECR output layer connected in sequence.

[0017] In the second aspect, the embodiment of the present application also provides a drug-target interaction prediction device based on a deep neural network, including: a preprocessing module, configured to calculate the similarity between different elements in a drug-related network and a protein-related network to obtain an initial drug similarity matrix and an initial protein similarity matrix; preprocess the initial drug similarity matrix and the initial protein similarity matrix to obtain a target drug similarity matrix and a target protein similarity matrix; concatenate the target drug similarity matrix and the target protein similarity matrix to obtain a DPP; configured to use a KNN algorithm to extract features from the DPP to obtain a feature adjacency matrix of the DPP; process the DPP according to preset rules to obtain a topological adjacency matrix of the DPP; a graph embedding module, configured to use graph embedding The input module performs feature extraction on the initial drug similarity matrix and the initial protein similarity matrix to obtain multi-view information; the feature extraction module is configured to use the feature extraction module to extract features on the multi-view information to obtain a feature matrix; the prediction module is configured to obtain a feature graph and a topological graph according to the feature adjacency matrix, the topological adjacency matrix and the feature matrix; the feature graph and the topological graph are respectively input into the graph attention network for feature extraction to obtain a first target feature and a second target feature; the feature graph and the topological graph are input into the graph convolution network for feature extraction to obtain a third target feature; the first target feature, the second target feature and the third target feature are input into the attention network for feature fusion to obtain a fused feature; the fused feature is input into the fully connected layer to obtain a prediction result.

[0018] In one embodiment, the preprocessing module is configured to: use the RWR algorithm to process the initial drug similarity matrix and the initial protein similarity matrix to obtain an intermediate drug similarity matrix and an intermediate protein similarity matrix; use the DAE algorithm to denoise and reduce the dimension of the intermediate drug similarity matrix and the intermediate protein similarity matrix to obtain the target drug similarity matrix and the target protein similarity matrix.

[0019] In one embodiment, the preset rule is: if two DPPs have the same drug or protein, then the two DPPs are considered to have the same topological edge.

[0020] In one embodiment, the graph embedding module is configured as follows: using the graph embedding module to perform feature extraction on the initial drug similarity matrix and the initial protein similarity matrix to obtain an embedding matrix; the graph embedding module includes a convolutional network, which is composed of multiple convolutional layers; learning corresponding weighted parameters for the embedding matrix to obtain the multi-view information.

[0021] In one embodiment, the graph embedding module is configured to: initialize the initial drug similarity matrix and the initial protein similarity matrix using random initialization weights and random initialization biases to obtain an initialization matrix; update the initialization matrix using a SUM function to obtain an update matrix; and perform normalization, activation, and regularization processing on the update matrix in sequence to obtain the embedding matrix.

[0022] In one embodiment, the prediction module is configured to: obtain a feature map and a topological map based on the feature adjacency matrix, the topological adjacency matrix, and the feature matrix, including: combining the feature adjacency matrix and the feature matrix to obtain the feature map; combining the topological adjacency matrix and the feature matrix to obtain the topological map.

[0023] In one embodiment, the preprocessing module is configured to: calculate the Jaccard similarity coefficients between different elements in the drug-related network and the protein-related network to obtain the initial drug similarity matrix and the initial protein similarity matrix.

[0024] In one embodiment, the drug-related network includes a drug-drug relationship network, a drug-related disease network, a drug and side effect network, and a drug-chemical structure similarity network; the drug-drug relationship network is used to represent the interaction between different drugs; the drug-related disease network is used to represent the relationship between drugs and diseases; the drug and side effect network is used to represent the relationship between drugs and side effects; the drug-chemical structure similarity network is used to represent the relationship between the chemical structures of different drugs; the protein-related network includes a protein-related disease network, a protein-protein relationship network, and a protein sequence similarity network; the protein-related disease network is used to represent the relationship between proteins and diseases; the protein-protein relationship network is used to represent the interaction between different proteins; the protein sequence similarity network is used to represent the relationship between different protein sequences.

[0025] In one embodiment, the feature extraction module includes an ECR network; the ECR network includes an ECR input layer, an ECR convolution layer, an ECR pooling layer, a CECJC layer, and an ECR output layer connected in sequence.

[0026] In a third aspect, an embodiment of the present application provides a computer device, comprising: a processor, a storage medium and a bus, wherein the storage medium stores program instructions executable by the processor, and when the computer device is running, the processor communicates with the storage medium through the bus, and the processor executes the program instructions to perform the steps of any one of the above methods.

[0027] In a fourth aspect, an embodiment of the present application provides a non-volatile computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of any one of the above methods are performed.

[0028] The beneficial effects of the present application are as follows: based on the characteristics of drugs and proteins, drug-related networks (including drug-drug relationship networks, drug-related disease networks, drug and side effect networks, and drug-chemical structure similarity networks) and protein-related networks (including protein-related disease networks, protein-protein relationship networks, and protein sequence similarity networks) are established, and convolutional neural networks are used to directly learn the feature vectors of the above seven networks, map the original high-dimensional data to a low-dimensional space to achieve data dimensionality reduction, and obtain a topological adjacency matrix and a feature adjacency matrix, which are of positive significance for finding the main dimensions of data distribution; multi-view dilated convolution, ResNet, and K nearest neighbor algorithm (KNN) in the graph embedding module ... NearestNeighbors) generates multi-view information. The complementary information and relationships in multi-view information can greatly improve the learning effect of deep neural networks, making it easier to obtain key features that affect the prediction of drug-target interactions. The ECR network in the feature extraction module is used to extract features from multi-view information to obtain more comprehensive feature information (feature matrix). A graph convolutional network and a graph attention network are designed to fuse the topological adjacency matrix, the feature adjacency matrix, and the feature matrix, and then the prediction results of drug-target interactions are obtained through the fully connected layer based on the fused features, which improves the prediction effect of drug-target interactions. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0030] Figure 1 A schematic diagram of a process for predicting drug-target interactions based on a deep neural network provided in an embodiment of the present application;

[0031] Figure 2A schematic diagram of a process for predicting drug-target interactions based on a deep neural network provided in an embodiment of the present application;

[0032] Figure 3 A schematic diagram of a process for predicting drug-target interactions based on a deep neural network provided in an embodiment of the present application;

[0033] Figure 4 A schematic diagram of a process for predicting drug-target interactions based on a deep neural network provided in an embodiment of the present application;

[0034] Figure 5 A schematic diagram of a drug-target interaction prediction method based on a deep neural network provided in an embodiment of the present application;

[0035] Figure 6 A schematic diagram of the structure of a drug-target interaction prediction device based on a deep neural network provided in an embodiment of the present application;

[0036] Figure 7 A schematic diagram of the structure of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments.

[0038] Therefore, the following detailed description of the embodiments of the present application provided in the accompanying drawings is not intended to limit the scope of the present application for which protection is sought, but merely represents selected embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0039] In the description of the present application, it should be noted that if the terms "upper", "lower", etc. appear, the orientation or position relationship indicated is based on the orientation or position relationship shown in the drawings, or is the orientation or position relationship in which the product of the application is usually placed when used. It is only for the convenience of describing the present application and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application.

[0040] In addition, the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0041] It should be noted that, in the absence of conflict, the features in the embodiments of the present application may be combined with each other.

[0042] Figure 1 A schematic diagram of a method for predicting drug-target interactions based on a deep neural network provided in an embodiment of the present application; Figure 1 As shown, the method includes:

[0043] Step 110, calculate the similarity between different elements in the drug-related network and the protein-related network to obtain an initial drug similarity matrix and an initial protein similarity matrix; preprocess the initial drug similarity matrix and the initial protein similarity matrix to obtain a target drug similarity matrix and a target protein similarity matrix; concatenate the target drug similarity matrix and the target protein similarity matrix to obtain a DPP; use a KNN algorithm to extract features from the DPP to obtain a feature adjacency matrix of the DPP; process the DPP according to preset rules to obtain a topological adjacency matrix of the DPP.

[0044] Among them, drug-related networks include drug-drug relationship network, drug-related disease network, drug-side effect network, and drug-chemical structure similarity network; drug-drug relationship network is used to represent the interaction between different drugs; drug-related disease network is used to represent the relationship between drugs and diseases; drug-side effect network is used to represent the relationship between drugs and side effects; drug-chemical structure similarity network is used to represent the relationship between the chemical structures of different drugs;

[0045] Protein-related networks include protein-related disease networks, protein-protein relationship networks, and protein sequence similarity networks; the protein-related disease network is used to represent the relationship between proteins and diseases; the protein-protein relationship network is used to represent the interactions between different proteins; and the protein sequence similarity network is used to represent the relationship between different protein sequences.

[0046] The initial drug similarity matrix and the initial protein similarity matrix can be obtained by calculating the Jaccard similarity coefficient between different elements in the drug-related network and the protein-related network.

[0047] Exemplarily, the Jaccard similarity coefficient calculation formula is shown in the following formula (1):

[0048]

[0049] For the drug-drug relationship network, M i and M j Represent drug i and drug j respectively; J(M i ,M j ) represents the interaction between drug i and drug j, which includes the enhancement, weakening or adverse reaction of drug effects; |M i ∩M j | represents the size of the intersection of two sets; |M i ∪M j | represents the size of the union of two sets.

[0050] For the drug-related disease network, M i and M j Represent drug i and disease j respectively; J(M i ,M j ) represents the relationship between drug i and disease j; |M i ∩M j | represents the size of the intersection of two sets; |M i ∪M j | represents the size of the union of two sets.

[0051] For the drug and side effect network, M i and M j Represent drug i and side effect j respectively; J(M i ,M j ) represents the relationship between drug i and side effect j; |M i ∩M j | represents the size of the intersection of two sets; |M i ∪M j | represents the size of the union of two sets.

[0052] For drug-chemical structure similarity networks, M i and M j Represent the chemical structure of drug i and drug j respectively; J(M i ,M j ) represents the relationship between the chemical structure of drug i and the chemical structure of drug j; |M i ∩M j| represents the size of the intersection of two sets; |M i ∪M j | represents the size of the union of two sets.

[0053] For protein-related disease networks, M i and M j represents protein i and disease j respectively; J(M i ,M j ) represents the relationship between protein i and disease j; |M i ∩M j | represents the size of the intersection of two sets; |M i ∪M j | represents the size of the union of two sets.

[0054] For protein-protein relationship networks, M i and M j Represent protein i and protein j respectively; J(M i ,M j ) represents the interaction relationship between protein i and protein j, which includes direct physical contact, functional association, and interaction revealed by genetic evidence; |M i ∩M j | represents the size of the intersection of two sets; |M i ∪M j | represents the size of the union of two sets.

[0055] For protein sequence similarity networks, M i and M j Represent the sequence of protein i and protein j respectively; J(M i ,M j ) represents the relationship between the sequence of protein i and the sequence of protein j; |M i ∩M j | represents the size of the intersection of two sets; |M i ∪M j | represents the size of the union of two sets.

[0056] Among them, the initial drug similarity matrix obtained in step 110 includes the initial similarity matrices corresponding to the four networks of drug-drug relationship network, drug-related disease network, drug and side effect network, and drug-chemical structure similarity network; the initial protein similarity matrix includes the initial similarity matrices corresponding to the three networks of protein-related disease network, protein-protein relationship network, and protein sequence similarity network; therefore, step 110 actually obtains seven initial similarity matrices.

[0057] Preprocessing includes integration, denoising, and dimensionality reduction of the above seven similarity matrices.

[0058] The target drug similarity matrix is ​​a similarity matrix for drugs after preprocessing; the target protein similarity matrix is ​​a similarity matrix for proteins after preprocessing. After preprocessing, the target drug similarity matrix and the target protein similarity matrix are both single matrices.

[0059] DPP is a drug-protein pair composed of a drug and a protein. The connection between drug p and protein q in DPP node i is shown in the following formula (2):

[0060]

[0061] The KNN algorithm is shown in the following formula (3):

[0062]

[0063] Among them, A and B are rows in the target drug similarity matrix and target protein similarity matrix in DPP, respectively, and both are vectors; A i and B i are the i-th element of vector A and vector B respectively; n is the number of elements in vector A and vector B.

[0064] The preset rule is: if two DPPs have the same drug or protein, then the two DPPs are considered to have the same topological edge.

[0065] Step 120: Use a graph embedding module to perform feature extraction on the initial drug similarity matrix and the initial protein similarity matrix to obtain multi-view information.

[0066] In actual operation, this step can be achieved through the following process: first, use the convolutional network in the graph embedding module to extract features from the initial drug similarity matrix and the initial protein similarity matrix to obtain multiple embedding matrices; second, learn weighting parameters for the multiple embedding matrices to obtain multiple weighted embedding matrices (i.e., the above-mentioned multi-view information). Here, the convolutional network can be single-layer or multi-layer, which is not limited here.

[0067] Step 130: Use a feature extraction module to extract features from the multi-view information to obtain a feature matrix.

[0068] Among them, the feature extraction module includes an ECR network; the ECR network includes an ECR input layer, an ECR convolution layer, an ECR pooling layer, a CECJC layer, and an ECR output layer connected in sequence. The ECR network combines the advantages of spatial dilation convolution and jump connection, and uses three adaptive circle dilation convolution kernels to capture neighborhood information of three different scales: local, medium, and long distance, thereby achieving comprehensive coverage of multi-scale information. This design not only avoids the computational complexity problem caused by excessive parameters, but also enhances the model's ability to capture key detail information through specially designed jump connection layers. The ECR network is specifically shown in the following formulas (4), (5), and (6):

[0069] x conv =ReLU(Conv2d(x,kernelsize,padding,dilation)) (4)

[0070] x pooling =Interpolate(ReLU(Conv2d(AdaptiveAvPool2d(x,1)))) (5)

[0071] x CECJC =Dropout(ReLU(Conv2d(Concat(x conv ,x pooling ,x),1))) (6)

[0072] In formula (4), x conv is the convolutional layer formula, x is the original input, kernelsize is the convolution kernel size, padding is the padding, dilation is the dilation rate, Conv2d represents the ordinary two-dimensional convolution, and ReLU represents the activation function; in formula (5), x pooling is the pooling layer formula, AdaptiveAvPool2d represents adaptive average pooling, and Interpolate represents sampling using interpolation. In formula (6), x CECJC It is the formula of CECJC layer, Concat is the merging of channels, and Dropout means using Dropout to prevent overfitting.

[0073] Through the ECR network, we learn the contextual representation and semantic representation of DPP from drug feature information and protein feature information. Specifically,

[0074] Step 140, according to the feature adjacency matrix, the topological adjacency matrix, and the feature matrix, obtain a feature map and a topological map; input the feature map and the topological map into the graph attention network for feature extraction, respectively, to obtain a first target feature and a second target feature; input the feature map and the topological map into the graph convolutional network for feature extraction, to obtain a third target feature; input the first target feature, the second target feature, and the third target feature into the attention network for feature fusion, to obtain a fused feature; input the fused feature into the fully connected layer to obtain a prediction result.

[0075] The feature map and the topological map may be formed by combining different matrices. Specifically, the feature adjacency matrix and the feature matrix may be combined to obtain the feature map; and the topological adjacency matrix and the feature matrix may be combined to obtain the topological map.

[0076] Graph Attention Networks (GAT) is a variant of Graph Neural Networks (GNN) that introduces an attention mechanism to perform weighted aggregation of node neighbors, thereby achieving adaptive allocation of weights to different neighbors. This mechanism enables GAT to better capture the relationship between nodes and improve the expressiveness of the model. The core idea of ​​GAT is that the hidden state of each node can be calculated by paying attention to its neighbor nodes. This self-attention mechanism allows the model to assign different weights to each neighbor node, instead of treating all neighbors equally like traditional graph convolutional networks (GCN).

[0077] The relevant calculations of the graph attention network are as follows:

[0078] First, given N node features, the set of all node features is F is the number of feature dimensions for each node; h i is the feature of node i; then the attention coefficient e of node i relative to node j ij As shown in the following formula (7):

[0079]

[0080] in, W are shared learnable parameters.

[0081] Secondly, the attention coefficient is normalized as shown in the following formula (8):

[0082]

[0083] Among them, LeakyReLU (Leaky Rectified Linear Unit) is an activation function commonly used in neural networks in deep learning. It is a variant of the ReLU (Rectified Linear Unit) activation function, which aims to solve the "dead neurons" problem that may occur during the training process of ReLU.

[0084] Again, LeakyReLU (i.e., f(x)) is as shown in the following formula (9):

[0085]

[0086] Here, a is a constant, usually 0.01, which is used to control the slope of the negative input.

[0087] Next, the node is updated, as shown in the following formula (10):

[0088]

[0089] in, is the new feature of node i after integrating the neighborhood information; σ is the activation function; W is a shared learnable parameter.

[0090] Finally, a multi-head attention mechanism is used to learn different meanings of DPP nodes from different representation subspaces to obtain the K-head set of node i As shown in the following formula (11):

[0091]

[0092] From the feature map G of DPP f =(A f ,X DPP ) to get the K-head set of node i As shown in the following formula (12):

[0093]

[0094] Graph Convolutional Networks (GCN) is a neural network structure used to process graph data. By performing convolution operations on graphs, it can effectively capture the complex relationships between nodes and the structural characteristics of graphs. The core idea of ​​GCN is to update the representation of nodes by aggregating the feature information of each node and its neighboring nodes. This method imitates the process of extracting local features through filters in traditional convolutional neural networks (CNN), but is suitable for irregular data structures such as graphs.

[0095] The relevant calculations of the graph convolutional network are as follows:

[0096] First, let Z ct and Z cf They represent the common representations learned from the topological graph and the feature graph respectively; the topological space and feature space output by the Ith layer of the GCN model are shown in the following formulas (13) and (14), respectively:

[0097]

[0098] Where A is the adjacency matrix of the input graph; D is the degree matrix of the input graph; is the weight matrix of layer I of the input graph.

[0099] Next, define the input of the first layer as (·) is an alternative representation, which can be the above or The output of the GCN model learned in topological space and feature space is the jointly embedded Z ct and Z cf ;

[0100] Finally, the two representations are averaged to obtain the common representation Z of the DPP node C As shown in the following formula (15):

[0101]

[0102] For a DPP node i, the three representations of node i can be used and Z iC Indicates that, let vector a i Representing their attention vectors, the attention function Att(Zt, Zf, Zc) can be obtained as follows (16):

[0103]

[0104] in, They are The corresponding attention score.

[0105] Among them, the attention network is a type of deep learning model whose core idea is to imitate the human attention mechanism, that is, to focus on the most relevant part of the current task when processing information. This mechanism allows the model to dynamically assign different processing weights to different parts of the input data, thereby improving the model's ability to process complex data.

[0106] The fully connected layers are the last few layers (usually one or several layers) in a neural network, which convert the feature map of the previous layer into a fixed-length feature vector through matrix multiplication and perform classification or regression through an activation function. In classification tasks, the activation function converts the feature vector into a probability distribution, representing the predicted probability of each category.

[0107] The drug-target interaction prediction method based on deep neural network provided in the embodiment of the present application first calculates the similarity between different elements in the drug-related network and the protein-related network to obtain an initial drug similarity matrix and an initial protein similarity matrix; secondly, pre-processes the initial drug similarity matrix and the initial protein similarity matrix to obtain a target drug similarity matrix and a target protein similarity matrix; splices the target drug similarity matrix and the target protein similarity matrix to obtain a DPP; thirdly, uses the KNN algorithm to extract features from the DPP to obtain a feature adjacency matrix of the DPP; processes the DPP according to preset rules to obtain a topological adjacency matrix of the DPP; then, uses a graph embedding algorithm to extract features from the DPP to obtain a feature adjacency matrix of the DPP; The input module performs feature extraction on the initial drug similarity matrix and the initial protein similarity matrix to obtain multi-view information; then, the feature extraction module is used to extract features from the multi-view information to obtain a feature matrix; then, according to the feature adjacency matrix, the topological adjacency matrix, and the feature matrix, a feature graph and a topological graph are obtained; the feature graph and the topological graph are respectively input into the graph attention network for feature extraction to obtain the first target feature and the second target feature; the feature graph and the topological graph are input into the graph convolutional network for feature extraction to obtain the third target feature; finally, the first target feature, the second target feature, and the third target feature are input into the attention network for feature fusion to obtain the fused feature; the fused feature is input into the fully connected layer to obtain the prediction result. In this way, based on the characteristics of drugs and proteins, drug-related networks (including drug-drug relationship network, drug-related disease network, drug and side effect network, drug-chemical structure similarity network) and protein-related networks (including protein-related disease network, protein-protein relationship network, protein sequence similarity network) were established. The convolutional neural network was used to directly learn the feature vectors of the above seven networks, and the original high-dimensional data was mapped to the low-dimensional space to achieve data dimensionality reduction, and the topological adjacency matrix and feature adjacency matrix were obtained, which is of positive significance for finding the main dimensions of data distribution; the multi-view dilated convolution, ResNet and K Nearest Neighbor algorithm (K Nearest Neighbor) in the graph embedding module were ... Neighbors) generate multi-view information. The complementary information and relationships in the multi-view information can greatly improve the learning effect of the deep neural network, making it easier to obtain key features that affect the prediction of drug-target interaction. The ECR network in the feature extraction module is used to extract features from multi-view information to obtain more comprehensive feature information (feature matrix). A graph convolutional network and a graph attention network are designed to fuse the topological adjacency matrix, the feature adjacency matrix, and the feature matrix, and then the prediction results of the drug-target interaction are obtained through the fully connected layer based on the fused features, which improves the prediction effect of the drug-target interaction.

[0108] The following is a continued explanation of the process of preprocessing the initial drug similarity matrix and the initial protein similarity matrix in the above embodiment to obtain the target drug similarity matrix and the target protein similarity matrix by combining examples. Figure 2 The following is a schematic diagram of a method for predicting drug-target interactions based on a deep neural network provided in an embodiment of the present application. Figure 2 As shown, the above step 120 may include the following steps 210 and 220:

[0109] Step 210: Use the RWR algorithm to process the initial drug similarity matrix and the initial protein similarity matrix to obtain an intermediate drug similarity matrix and an intermediate protein similarity matrix.

[0110] The RWR algorithm is shown in the following formula (17):

[0111] r=cW ij r+(1-c)e (17)

[0112] Among them, the size of c is between [0, 1]; W ij is the transition probability from node i to node j; e is the starting point vector; r is the end point vector.

[0113] Since the initial drug similarity matrix includes four initial similarity matrices related to drugs, and the initial protein similarity matrix includes three initial similarity matrices related to proteins; therefore, the initial similarity matrices after being processed by the RWR algorithm in this step are also seven, including four initial similarity matrices related to drugs and three initial similarity matrices related to proteins. Furthermore, in actual operation, the initial similarity matrices processed by the RWR algorithm can be integrated, that is, the four initial similarity matrices related to drugs processed by the RWR algorithm are spliced ​​to obtain the above-mentioned intermediate drug similarity matrix, and the three initial similarity matrices related to proteins processed by the RWR algorithm are spliced ​​to obtain the above-mentioned intermediate protein similarity matrix. In this way, the intermediate drug similarity matrix and the intermediate protein similarity matrix obtained are both single matrices.

[0114] Step 220: Use the DAE algorithm to denoise and reduce the dimension of the intermediate drug similarity matrix and the intermediate protein similarity matrix to obtain a target drug similarity matrix and a target protein similarity matrix.

[0115] Among them, the denoising autoencoder (DAE) is a deep learning model that achieves denoising by adding noise to the input data and then training the model to recover the original data from the noisy data. At the same time, DAE can also achieve data dimensionality reduction through its encoding and decoding process.

[0116] The DAE algorithm is shown in the following formula (18):

[0117]

[0118] in, is the noise distribution; x is the original input; is the input after adding noise; W1, b1 are the encoding weight and bias respectively; W2, b2 are the decoding weight and bias respectively; h1 represents the hidden layer after encoding; y represents the reconstructed output; σ e is the activation function used in the encoding process; σ d is the activation function used in the decoding process.

[0119] The drug-target interaction prediction method based on deep neural network provided in the embodiment of the present application, first, the initial drug similarity matrix and the initial protein similarity matrix are processed by RWR algorithm to obtain the intermediate drug similarity matrix and the intermediate protein similarity matrix; secondly, the intermediate drug similarity matrix and the intermediate protein similarity matrix are denoised and reduced in dimension by DAE algorithm to obtain the target drug similarity matrix and the target protein similarity matrix. In this way, based on the characteristics of drugs and proteins, drug-related networks (including drug-drug relationship network, drug-related disease network, drug and side effect network, drug-chemical structure similarity network) and protein-related networks (including protein-related disease network, protein-protein relationship network, protein sequence similarity network) are established, and the eigenvectors of the above seven networks are directly learned by convolutional neural network, and the original high-dimensional data is mapped to low-dimensional space to achieve data dimensionality reduction, and the topological adjacency matrix and the characteristic adjacency matrix are obtained, which has positive significance for finding the main dimensions of data distribution.

[0120] The following is a continued explanation of the process of extracting features from the initial drug similarity matrix and the initial protein similarity matrix using the graph embedding module in the above embodiment to obtain multi-view information in combination with examples. Figure 3 The following is a schematic diagram of a method for predicting drug-target interactions based on a deep neural network provided in an embodiment of the present application. Figure 3 As shown, the above step 140 may include the following steps 310 and 320:

[0121] Step 310: Use a graph embedding module to perform feature extraction on the initial drug similarity matrix and the initial protein similarity matrix to obtain an embedding matrix.

[0122] Among them, the graph embedding module includes a convolutional network, and the convolutional network is composed of multiple convolutional layers.

[0123] The preferred mode of a convolutional network is that the convolutional network is composed of multiple convolutional layers. Of course, the convolutional network can also be composed of one convolutional layer, which is not limited here.

[0124] Specifically, Figure 4 As shown, the above step 310 may include the following steps 410 to 430:

[0125] Step 410: Initialize the initial drug similarity matrix and the initial protein similarity matrix using random initialization weights and random initialization biases to obtain an initialization matrix.

[0126] The initialization process is shown in the following formula (19):

[0127] r=W0x+b0(19)

[0128] Among them, W0 is the randomly initialized weight; b0 is the randomly initialized bias; x is the input; r is the initialization matrix.

[0129] Step 420: Use the SUM function to update the initialization matrix to obtain an updated matrix.

[0130] The SUM function is shown in the following formula (20):

[0131] y i =W i x+b i (20)

[0132] Among them, W i is the weight; b i For bias.

[0133] The update matrix r′ is shown in the following formula (21):

[0134] r′=r+y i (twenty one)

[0135] Step 430: perform normalization, activation, and regularization on the update matrix in sequence to obtain an embedding matrix.

[0136] The normalization, activation, and regularization processes are shown in the following formula (22):

[0137] r=Dropout(ReLU(BathNorm(r)))(22)

[0138] Among them, Dropout represents regularization processing; ReLU represents activation processing; BathNorm represents normalization processing.

[0139] Taking the lth layer as an example, the forward propagation of the graph embedding module is shown in the following formula (23):

[0140] x l =ConvLayer l (x l-1 )(twenty three)

[0141] Among them, ConvLayer l represents the lth convolutional layer.

[0142] Update input x = x l .

[0143] Step 320: learn corresponding weighting parameters for the embedding matrix to obtain multi-view information.

[0144] Among them, since the initial drug similarity matrix includes four initial similarity matrices related to drugs, and the initial protein similarity matrix includes three initial similarity matrices related to proteins; therefore, the embedding matrix also includes four embedding matrices related to drugs and three embedding matrices related to proteins.

[0145] In actual operation, each embedding matrix in this step corresponds to a weighting parameter, that is, each embedding matrix learns its corresponding weighting parameter to obtain its corresponding weighted embedding matrix. Finally, these seven weighted embedding matrices constitute the above-mentioned multi-view information.

[0146] The learning process of this step is shown in the following formula (24):

[0147]

[0148] Among them, x1, x2, x3, x4, x5, x6, and x7 are four drug-related embedding matrices and three protein-related embedding matrices respectively; beta1, beta2, beta3, beta4, beta5, beta6, and beta7 are the weighted parameters of the four drug-related embedding matrices and the three protein-related embedding matrices respectively, and the weighted parameters are automatically learned by the network according to the loss function; x1′, x2′, x3′, x4′, x5′, x6′, and x7′ are four drug-related weighted embedding matrices and three protein-related weighted embedding matrices respectively.

[0149] like Figure 5 , which is a schematic diagram of a drug-target interaction prediction method based on a deep neural network provided in an embodiment of the present application. In the figure, the spatial information is the above-mentioned topological adjacency matrix; the spectral information is the above-mentioned feature adjacency matrix.

[0150] The drug-target interaction prediction method based on deep neural network provided in the embodiment of the present application firstly uses the graph embedding module to extract features from the initial drug similarity matrix and the initial protein similarity matrix to obtain an embedding matrix; secondly, the corresponding weighting parameters are learned for the embedding matrix to obtain multi-view information. In this way, the multi-view information is generated by using the multi-view dilated convolution, ResNet and K Nearest Neighbors algorithm in the graph embedding module. The complementary information and relationships in the multi-view information can greatly improve the learning effect of the deep neural network, making it easier to obtain key features that affect the prediction of drug-target interactions.

[0151] After introducing the drug-target interaction prediction method based on deep neural network according to the exemplary embodiment of the present disclosure, next, reference is made to Figure 6 A drug-target interaction prediction device 600 based on a deep neural network according to an exemplary embodiment of the present disclosure is described.

[0152] refer to Figure 6 The drug-target interaction prediction device 600 based on deep neural network includes: a preprocessing module 610, which is configured to calculate the similarity between different elements in the drug-related network and the protein-related network to obtain an initial drug similarity matrix and an initial protein similarity matrix; preprocess the initial drug similarity matrix and the initial protein similarity matrix to obtain a target drug similarity matrix and a target protein similarity matrix; splice the target drug similarity matrix and the target protein similarity matrix to obtain a DPP; is configured to use a KNN algorithm to extract features from the DPP to obtain a feature adjacency matrix of the DPP; processes the DPP according to a preset rule to obtain a topological adjacency matrix of the DPP; a graph embedding module 620, which is configured to use a graph embedding module 620 performs feature extraction on the initial drug similarity matrix and the initial protein similarity matrix to obtain multi-view information; the feature extraction module 630 is configured to use the feature extraction module 630 to perform feature extraction on the multi-view information to obtain a feature matrix; the prediction module 640 is configured to obtain a feature map and a topological map based on the feature adjacency matrix, the topological adjacency matrix, and the feature matrix; the feature map and the topological map are respectively input into the graph attention network for feature extraction to obtain a first target feature and a second target feature; the feature map and the topological map are input into the graph convolution network for feature extraction to obtain a third target feature; the first target feature, the second target feature, and the third target feature are input into the attention network for feature fusion to obtain a fused feature; the fused feature is input into the fully connected layer to obtain a prediction result.

[0153] In one embodiment, the preprocessing module 610 is configured to: use the RWR algorithm to process the initial drug similarity matrix and the initial protein similarity matrix to obtain an intermediate drug similarity matrix and an intermediate protein similarity matrix; use the DAE algorithm to denoise and reduce the dimension of the intermediate drug similarity matrix and the intermediate protein similarity matrix to obtain a target drug similarity matrix and a target protein similarity matrix.

[0154] In one embodiment, the preset rule is: if two DPPs have the same drug or protein, then the two DPPs are considered to have the same topological edge.

[0155] In one embodiment, the graph embedding module 620 is configured as follows: the graph embedding module 620 is used to perform feature extraction on the initial drug similarity matrix and the initial protein similarity matrix to obtain an embedding matrix; the graph embedding module 620 includes a convolutional network, and the convolutional network is composed of multiple convolutional layers; the corresponding weighted parameters are learned for the embedding matrix to obtain multi-view information.

[0156] In one embodiment, the graph embedding module 620 is configured to: initialize the initial drug similarity matrix and the initial protein similarity matrix using random initialization weights and random initialization biases to obtain an initialization matrix; update the initialization matrix using a SUM function to obtain an update matrix; and perform normalization, activation, and regularization on the update matrix in sequence to obtain an embedding matrix.

[0157] In one embodiment, the prediction module 640 is configured to obtain a feature map and a topological map based on the feature adjacency matrix, the topological adjacency matrix, and the feature matrix, including: combining the feature adjacency matrix and the feature matrix to obtain the feature map; combining the topological adjacency matrix and the feature matrix to obtain the topological map.

[0158] In one embodiment, the preprocessing module 610 is configured to calculate the Jaccard similarity coefficients between different elements in the drug-related network and the protein-related network to obtain an initial drug similarity matrix and an initial protein similarity matrix.

[0159] In one embodiment, the drug-related network includes a drug-drug relationship network, a drug-related disease network, a drug-side effect network, and a drug-chemical structure similarity network; the drug-drug relationship network is used to represent the interaction between different drugs; the drug-related disease network is used to represent the relationship between drugs and diseases; the drug-side effect network is used to represent the relationship between drugs and side effects; the drug-chemical structure similarity network is used to represent the relationship between the chemical structures of different drugs; the protein-related network includes a protein-related disease network, a protein-protein relationship network, and a protein sequence similarity network; the protein-related disease network is used to represent the relationship between proteins and diseases; the protein-protein relationship network is used to represent the interaction between different proteins; the protein sequence similarity network is used to represent the relationship between different protein sequences.

[0160] In one embodiment, the feature extraction module 630 includes an ECR network; the ECR network includes an ECR input layer, an ECR convolution layer, an ECR pooling layer, a CECJC layer, and an ECR output layer connected in sequence.

[0161] The above-mentioned device is used to execute the method provided by the aforementioned embodiment, and its implementation principle and technical effect are similar, which will not be repeated here.

[0162] The above modules may be one or more integrated circuits configured to implement the above methods, such as one or more application specific integrated circuits (ASICs), or one or more microprocessors, or one or more field programmable gate arrays (FPGAs). For another example, when a module is implemented in the form of a processing element scheduling program code, the processing element may be a general-purpose processor, such as a central processing unit (CPU) or other processor that can call program code. For another example, these modules may be integrated together and implemented in the form of a system-on-a-chip (SOC).

[0163] Figure 7 A schematic diagram of a computer device provided in an embodiment of the present application. The device can be integrated into a terminal device or a chip of a terminal device. The terminal can be a computing device with data processing capabilities.

[0164] The device includes: a processor 701 , a storage medium 702 and a bus 703 .

[0165] The storage medium 702 stores program instructions executable by the processor 701. When the computer device 700 is running, the processor 701 communicates with the storage medium 702 via the bus 703, and the processor 701 executes the program instructions to perform the above method embodiment. The specific implementation method and technical effect are similar and will not be repeated here.

[0166] Optionally, the present invention also provides a program product, such as a computer-readable storage medium, comprising a program, which is used to execute the above method embodiment when executed by a processor.

[0167] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0168] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0169] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0170] The above-mentioned integrated unit implemented in the form of a software functional unit can be stored in a computer-readable storage medium. The above-mentioned software functional unit is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) or a processor (English: processor) to perform some steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (English: Read-Only Memory, abbreviated: ROM), random access memory (English: Random Access Memory, abbreviated: RAM), disk or optical disk and other media that can store program codes.

[0171] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art who is familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A drug-target interaction prediction method based on deep neural network, characterized in that: include: Calculate the similarities between different elements in the drug-related network and the protein-related network to obtain the initial drug similarity matrix and the initial protein similarity matrix; Preprocessing the initial drug similarity matrix and the initial protein similarity matrix to obtain a target drug similarity matrix and a target protein similarity matrix; splicing the target drug similarity matrix and the target protein similarity matrix to obtain a DPP; Using a KNN algorithm to extract features from the DPP to obtain a feature adjacency matrix of the DPP; Processing the DPP according to a preset rule to obtain a topological adjacency matrix of the DPP; Using a graph embedding module to perform feature extraction on the initial drug similarity matrix and the initial protein similarity matrix to obtain multi-view information; Using a feature extraction module to extract features from the multi-view information to obtain a feature matrix; A feature map and a topological map are obtained according to the feature adjacency matrix, the topological adjacency matrix, and the feature matrix; the feature map and the topological map are respectively input into a graph attention network for feature extraction to obtain a first target feature and a second target feature; the feature map and the topological map are input into a graph convolutional network for feature extraction to obtain a third target feature; the first target feature, the second target feature, and the third target feature are input into an attention network for feature fusion to obtain a fused feature; the fused feature is input into a fully connected layer to obtain a prediction result.

2. The method according to claim 1, characterized in that The preprocessing of the initial drug similarity matrix and the initial protein similarity matrix to obtain a target drug similarity matrix and a target protein similarity matrix includes: The initial drug similarity matrix and the initial protein similarity matrix are processed by using a RWR algorithm to obtain an intermediate drug similarity matrix and an intermediate protein similarity matrix; The DAE algorithm is used to denoise and reduce the dimension of the intermediate drug similarity matrix and the intermediate protein similarity matrix to obtain the target drug similarity matrix and the target protein similarity matrix.

3. The method according to claim 1, characterized in that The preset rule is: if two DPPs have the same drug or protein, then the two DPPs are considered to have the same topological edge.

4. The method according to claim 1, characterized in that: The graph embedding module is used to extract features from the initial drug similarity matrix and the initial protein similarity matrix to obtain multi-view information, including: The graph embedding module is used to perform feature extraction on the initial drug similarity matrix and the initial protein similarity matrix to obtain an embedding matrix; the graph embedding module includes a convolutional network, and the convolutional network is composed of a plurality of convolutional layers; Corresponding weighting parameters are learned for the embedding matrix to obtain the multi-view information.

5. The method according to claim 4, characterized in that The graph embedding module is used to extract features from the initial drug similarity matrix and the initial protein similarity matrix to obtain an embedding matrix, including: Initializing the initial drug similarity matrix and the initial protein similarity matrix using random initialization weights and random initialization biases to obtain an initialization matrix; The initialization matrix is ​​updated using a SUM function to obtain an updated matrix; The update matrix is ​​subjected to normalization processing, activation processing, and regularization processing in sequence to obtain the embedding matrix.

6. The method according to claim 1, characterized in that The step of obtaining a feature graph and a topological graph according to the feature adjacency matrix, the topological adjacency matrix, and the feature matrix includes: Combining the feature adjacency matrix and the feature matrix to obtain the feature graph; The topological adjacency matrix and the characteristic matrix are combined to obtain the topological graph.

7. The method according to claim 1, characterized in that The calculating of the similarities between different elements in the drug-related network and the protein-related network to obtain an initial drug similarity matrix and an initial protein similarity matrix includes: The Jaccard similarity coefficients between different elements in the drug-related network and the protein-related network are calculated to obtain the initial drug similarity matrix and the initial protein similarity matrix.

8. The method according to claim 1, characterized in that The drug-related network includes a drug-drug relationship network, a drug-related disease network, a drug-side effect network, and a drug-chemical structure similarity network; the drug-drug relationship network is used to represent the interaction between different drugs; the drug-related disease network is used to represent the relationship between drugs and diseases; the drug-side effect network is used to represent the relationship between drugs and side effects; The drug-chemical structure similarity network is used to represent the relationship between the chemical structures of different drugs; The protein-related network includes a protein-related disease network, a protein-protein relationship network, and a protein sequence similarity network; the protein-related disease network is used to represent the relationship between protein and disease; the protein-protein relationship network is used to represent the interaction between different proteins; and the protein sequence similarity network is used to represent the relationship between different protein sequences.

9. The method according to claim 1, characterized in that: The feature extraction module includes an ECR network; the ECR network includes an ECR input layer, an ECR convolution layer, an ECR pooling layer, a CECJC layer, and an ECR output layer connected in sequence.

10. A drug-target interaction prediction device based on deep neural network, characterized in that: include: A preprocessing module is configured to calculate the similarities between different elements in the drug-related network and the protein-related network to obtain an initial drug similarity matrix and an initial protein similarity matrix; Preprocessing the initial drug similarity matrix and the initial protein similarity matrix to obtain a target drug similarity matrix and a target protein similarity matrix; splicing the target drug similarity matrix and the target protein similarity matrix to obtain a DPP; The method is configured to extract features of the DPP using a KNN algorithm to obtain a feature adjacency matrix of the DPP; Processing the DPP according to a preset rule to obtain a topological adjacency matrix of the DPP; A graph embedding module is configured to extract features from the initial drug similarity matrix and the initial protein similarity matrix using the graph embedding module to obtain multi-view information; A feature extraction module is configured to extract features from the multi-view information using a feature extraction module to obtain a feature matrix; The prediction module is configured to obtain a feature map and a topological map according to the feature adjacency matrix, the topological adjacency matrix, and the feature matrix; input the feature map and the topological map into a graph attention network for feature extraction to obtain a first target feature and a second target feature; input the feature map and the topological map into a graph convolution network for feature extraction to obtain a third target feature; input the first target feature, the second target feature, and the third target feature into an attention network for feature fusion to obtain a fused feature; The fused features are input into the fully connected layer to obtain the prediction result.

Images