Multi-modal molecular attribute prediction method based on cue word graph convolution

Abstract

The invention belongs to the technical field of artificial intelligence assisted drug design, and particularly relates to a cue graph convolution-based multi-modal molecular attribute prediction method, which comprises the following steps of: selecting and establishing a data set: selecting a plurality of molecular attributes so as to construct an initial data set of a multi-modal method, dividing the initial data set into a training set, a verification set and a test set, wherein the initial data set comprises a plurality of molecular attributes and a plurality of molecules corresponding to the molecular attributes; and data preprocessing: constructing molecular attribute prediction by the method. The invention provides a brand-new molecular attribute prediction method, multi-modal molecular characteristics are fused through potential association, and the importance of a molecular key structure part is highlighted. The characteristics of molecules are captured from multiple angles, and the prediction accuracy and reliability are improved. Meanwhile, more abundant and comprehensive feature representation can be learned by combining information of multiple modes, so that the method has better generalization ability on different data sets and tasks.

Description

Technical Field

[0001] The present invention belongs to the technical field of artificial intelligence-assisted drug design, and specifically relates to a multimodal molecular property prediction method based on cue word graph convolution. Background Art

[0002] Drug development is a long, costly and complex process, and molecular property prediction is a basic and important task in the drug development process. Determining molecular properties as early as possible, such as biological activity and toxicity, can accelerate the drug development process, avoid failures due to property problems in later experiments as much as possible, and reduce the number and cost of experiments. The traditional method of obtaining molecular properties through experiments and theoretical calculations is accurate, but time-consuming and costly. With the development of artificial intelligence, applying it to the field of molecular property prediction to improve prediction accuracy has become an important research direction, greatly reducing the time and cost of prediction.

[0003] At present, there are two main methods for artificial intelligence-assisted drug design: one is machine learning, and the other is deep learning. For both methods, there is an important problem: how to select the most important descriptors related to the target properties from a large number of predefined and computable molecular descriptors. This is not only about the accuracy of the method, but also about the interpretability of the method. Machine learning methods rely on the past experience of experts and have a certain degree of subjectivity. The deep learning method entrusts this task to the neural network, so that the most valuable features of the original input data can be extracted, and the relationship between the structure and its properties can be more accurately discovered. Most of the existing methods are based on a single molecular mode. However, it is difficult to fully reveal the complex properties of molecules based on the information of a single mode alone. In addition, there is currently no graph readout method designed for molecular properties, and these methods tend to ignore some of the chemical information that comes with the molecules themselves. Therefore, it is necessary to provide a new property prediction method to solve the above problems. Summary of the invention

[0004] In view of the technical problem that currently there is no graph readout method designed for molecular properties, the present invention provides a multimodal molecular property prediction method based on cue word graph convolution.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0006] A multimodal molecular property prediction method based on cue word graph convolution comprises the following steps:

[0007] S1. Selection and establishment of data set: multiple molecular attributes are selected to construct an initial data set of the multimodal method, and the initial data set is divided into a training set, a validation set and a test set, wherein the initial data set contains multiple molecular attributes and their corresponding molecules;

[0008] S2. Data preprocessing: Perform graph embedding and preliminary chemical feature calculation operations on each molecule in the training set of the initial data set to obtain the features of two dimensions corresponding to each molecule; combine the molecules and their corresponding graph and sequence features to obtain the training set of the target data set for training the multimodal method;

[0009] S3, method construction: constructing a multimodal method, wherein the multimodal method comprises a multimodal fusion module, a feature extraction module and a molecular graph readout module connected in sequence; training the multimodal method using a training set in a target data set, and obtaining a trained multimodal method after verifying that the trained method is qualified through a validation set and a test set;

[0010] S4. Molecular property prediction: The molecules with properties to be determined are input into the trained multimodal method for property prediction, and the multimodal method outputs the molecular properties to be predicted.

[0011] The method for performing graph embedding and preliminary chemical feature calculation operations on each molecule in S2 is: using the RDKit extension package to calculate the chemical features of each molecule in the training set of the initial data set, and embedding them in the form of graphs and sequences.

[0012] The method for calculating the chemical characteristics of each molecule in the training set of the initial data set is:

[0013] S21. Use the RDKit extension package to calculate the chiral and aromatic chemical features of the molecule and embed them as node and bond features to obtain the graph features of the molecule.

[0014] S22. Use the RDKit extension package to calculate the MACCS fingerprint, PubChem fingerprint, pharmacophore fingerprint and Morgan fingerprint of the molecule to obtain the sequence characteristics of the molecule.

[0015] The multimodal molecular property prediction method based on the cue word graph convolution in S3 is:

[0016] S31, multimodal fusion module: extract the substructure features of the molecule by decoding the one-dimensional molecular fingerprint information; then, map these screened substructure features onto the molecular graph to achieve effective fusion of the one-dimensional sequence features and the two-dimensional molecular graph features;

[0017] S32, feature extraction module: After fusing multi-dimensional features, a graph convolutional neural network is used to extract features, and a gating mechanism and residual connection are added to control the flow of information, so that the network can better focus on important information in the node;

[0018] S33, molecular graph readout module: The present invention chooses to construct and output a complete molecular graph based on the elements contained in the molecule, and designs a convolution-based readout module; in the readout module, the molecular graph information processed by graph convolution is extracted and fused into a one-dimensional representation through a three-layer convolutional network.

[0019] The method for achieving effective fusion of one-dimensional sequence features and two-dimensional molecular graph features in S31 is:

[0020] S311. Initialize the molecular graph as an undirected graph, in which atoms are represented as nodes and chemical bonds are represented as edges. Each node in the graph is represented by 6 atomic features, and each edge is represented by 4 bond features. Most of the features used to represent nodes and edges are directly calculated using RDKit. For ease of processing, these features are encoded using one-hot encoding.

[0021] S312, using the Morgan fingerprint in dictionary form to extract substructure information in the molecule, after obtaining the dictionary containing the substructure information, using the hash operation function in RDKit to map the substructure fragments back to the original molecule; in order to ensure the accuracy of the extracted information, only the substructure information that can be uniquely mapped is mapped back to the original molecule;

[0022] S313. Add a predefined value to the last-dimensional embedding feature of the corresponding atom and chemical bond to indicate that the atom and chemical bond are located in a substructure fragment.

[0023] The method for extracting and fusing the molecular graph information processed by graph convolution into a one-dimensional representation through a three-layer convolutional network in S33 is:

[0024] S331, reading out the molecular graph obtained by graph convolution processing according to element categories to obtain a molecular feature graph classified according to elements;

[0025] S332. Use dilated convolution to perform convolution operation on each molecular feature map. The operation process is divided into three layers. The number of convolutions in each layer is inversely proportional to the number of layers. The final representation that integrates all molecular features is obtained:

[0026]

[0027] Among them, n is the number of convolutions in each layer, C i is the output of each convolutional layer, C j is the output of the jth convolution block in each layer. By using multiple identical convolution units to extract features and then averaging the features, the model's dependence on a single convolution block is reduced, thereby increasing the model's stability. At the same time, each convolution block can extract different features from the data, making the extracted features more comprehensive and rich.

[0028] The classification method is evaluated using the area under the ROC curve AUC, and the prediction method is evaluated using the root mean square error RMSE.

[0029] Compared with the prior art, the present invention has the following beneficial effects:

[0030] The present invention provides a new molecular property prediction method, which integrates multimodal molecular features through potential associations and highlights the importance of key structural parts of molecules. It captures the characteristics of molecules from multiple angles and improves the accuracy and reliability of predictions. At the same time, by combining information from multiple modalities, the method can learn richer and more comprehensive feature representations, thereby having better generalization capabilities on different data sets and tasks. In addition, the present invention also designs a three-layer dilated convolution of the graph readout module based on the molecular properties. This readout method can more effectively integrate the molecular node information processed by the graph neural network. These strategies enable deep learning methods to focus on important atomic information in molecules, which not only helps to understand complex chemical structures, but also can effectively improve the accuracy of molecular property predictions, thereby promoting the drug development process. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the implementation methods of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the implementation methods or the description of the prior art. Obviously, the drawings in the following description are only exemplary, and for ordinary technicians in this field, other implementation drawings can be derived from the provided drawings without creative work.

[0032] The structures, proportions, sizes, etc. illustrated in this specification are only used to match the contents disclosed in the specification so as to facilitate understanding and reading by persons familiar with the technology. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantial technical significance. Any structural modification, change in proportion or adjustment of size shall still fall within the scope of the technical contents disclosed in the present invention without affecting the effects and purposes that can be achieved by the present invention.

[0033] Figure 1 A schematic diagram of a prediction method flow chart of an embodiment of the present invention;

[0034] Figure 2 Schematic diagram of the prediction method structure of an embodiment of the present invention.

[0035] Figure 3 It is a schematic diagram of the fusion module structure of an embodiment of the present invention.

[0036] Figure 4 Schematic diagram of the feature extraction module structure of an embodiment of the present invention.

[0037] Figure 5 Schematic diagram of the attention mechanism module structure of an embodiment of the present invention. DETAILED DESCRIPTION

[0038] 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. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. These descriptions are only to further illustrate the features and advantages of the present invention, rather than to limit the claims of the present invention. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0039] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0040] The purpose of this embodiment is to propose a new multimodal molecular property prediction method based on cue word graph convolution, which effectively solves the problem that the existing methods do not fully consider the potential associations between molecular features, the characteristics of molecular graphs, and other factors, and realize the prediction or classification of molecular properties. Figure 1-5 As shown, the program includes:

[0041] S1, data set selection and establishment: select multiple molecular attributes to construct an initial data set for the multimodal method, and divide the initial data set into a training set, a validation set, and a test set. The initial data set contains multiple molecular attributes and their corresponding multiple molecules.

[0042] S2, data preprocessing: perform graph embedding and preliminary chemical feature calculation operations on each molecule in the training set of the initial dataset to obtain features of two dimensions corresponding to each molecule; combine the molecules and their corresponding graph and sequence features to obtain the training set of the target dataset for training the multimodal method.

[0043] In a specific embodiment, for ease of description, feature calculation and multimodal embedding are performed for any molecule in the training set of the initial data set. On this basis, when the chemical feature calculation of each molecule in the training set of the initial data set is performed using the RDKit extension package and embedded in the form of graphs and sequences, the S21 may include the following steps S211 and S212:

[0044] S211, using the RDKit extension package to calculate the chemical features of the molecule, such as chirality and aromaticity, as shown in Table 1. Atoms are represented as nodes, and chemical bonds are represented as edges connecting nodes, so that the molecule is embedded in the form of a graph. The initial embedding feature of atoms is 28 dimensions, and the initial embedding feature of chemical bonds is 7 dimensions, which is the graph feature of the molecule;

[0045] S212, using the RDKit extension package to calculate the MACCS fingerprint, PubChem fingerprint, pharmacophore fingerprint (Pharmacophore ErG fingerprint) and Morgan fingerprint of the molecule, which are the sequence characteristics of the molecule.

[0046] S3, method construction: construct a multimodal method, the multimodal method comprising a multimodal fusion module, a feature extraction module and a molecular graph readout module connected in sequence. The multimodal method is trained using a training set in the target data set, and the trained method is verified to be qualified by a validation set and a test set, thereby obtaining a trained multimodal method.

[0047] Specifically, when the multimodal method is trained using the training set in the target data set, the multimodal method is trained using a plurality of molecules in the training set of the target data set and their corresponding graphs and sequence features.

[0048] Further, when S3 uses the training set in the target data set to train the multimodal method, it may include the following steps S31 to S33:

[0049] S31, multimodal fusion module: For each molecule's fingerprint and graph features, this embodiment designs an algorithm to fuse them. By decoding the one-dimensional molecular fingerprint information, the substructure features of the molecule can be extracted. Subsequently, these screened substructure features are mapped to the molecular graph to achieve effective fusion of the one-dimensional sequence features and the two-dimensional molecular graph features.

[0050] Furthermore, when S31 is specifically implemented, it includes the following steps S311 to S313:

[0051] S311, initialize the molecular graph as an undirected graph, where atoms are represented as nodes and chemical bonds are represented as edges. As shown in Table 1, each node in the graph is represented by 6 atomic features, and each edge is represented by 4 bond features. Most of the features used to represent nodes and edges are directly calculated using RDKit. For ease of processing, these features are encoded using one-hot encoding.

[0052] S312, using the Morgan fingerprint in dictionary form to extract substructure information in the molecule, after obtaining the dictionary containing the substructure information, using the hash function in RDKit to map the substructure fragments back to the original molecule. In order to ensure the accuracy of the extracted information, only the substructure information that can be uniquely mapped is mapped back to the original molecule.

[0053] S313, adding a predefined value to the last dimension embedding feature of the corresponding atom and chemical bond to indicate that the atom and chemical bond are located in a substructure fragment.

[0054] Table 1 Description of atomic and chemical bond embedding characteristics

[0055]

[0056]

[0057] S32, feature extraction module: After fusing the multimodal features, this embodiment uses a graph convolutional neural network to extract features. In order to avoid the problem of gradient vanishing and gradient exploding during network training, a residual connection module is added to each layer of the graph convolutional neural network, so that the node representation learned in the previous layer can be fused with the new representation of the current layer to retain the important features in the previous layer. At the same time, this method also uses a gating mechanism to control the flow of information so that the network can better focus on the important information in the node.

[0058] S33, molecular graph readout module: This embodiment chooses to construct and output a complete molecular graph based on the elements contained in the molecule, and designs a convolution-based readout module. In the readout module, the molecular graph information processed by graph convolution is extracted and fused into a one-dimensional representation through a three-layer convolutional network.

[0059] Further, when using the training set in the target data set to train the multimodal method, S33 includes the following steps S331 and S332:

[0060] S331, reading out the molecular graph obtained by graph convolution processing according to element categories, thereby obtaining a molecular feature graph classified according to elements;

[0061] S332, use dilated convolution to perform convolution operation on each molecular feature map. The operation process is divided into three layers. The number of convolutions in each layer is inversely proportional to the number of layers. The final representation that integrates all molecular features can be obtained. The formula table can be used as follows:

[0062]

[0063] Among them, n is the number of convolutions in each layer, C i is the output of each convolutional layer, Cj is the output of the jth convolution block in each layer. By using multiple identical convolution units for feature extraction and then averaging, the model's dependence on a single convolution block can be reduced, thereby increasing the model's stability. At the same time, each convolution block can extract different features from the data, which can make the extracted features more comprehensive and rich.

[0064] S4, molecular attribute prediction: the molecule with the to-be-determined attributes is input into the trained multimodal method for attribute prediction, and the multimodal method outputs the molecular attributes to be predicted.

[0065] In the training phase, the method learns the relationship between molecular features and properties through a large amount of data. In the verification or actual application, adding prompt words composed of fingerprints can guide the method to more accurately understand the requirements of specific tasks, thereby generating outputs that are more in line with expectations. The prompt words are composed of structural MACCS fingerprints, pharmacophore fingerprints (Pharmacophore ErGfingerprint) and PubChem fingerprints.

[0066] Furthermore, the classification method of this embodiment uses the area under the ROC curve AUC for evaluation, and the higher the index, the better the effect; the prediction method uses the root mean square error RMSE for evaluation, and the lower the index, the better the effect. The results are shown in Table 2.

[0067] Table 2 Results of this embodiment on the data set

[0068]

[0069] Among them, GCN is a graph convolution method, GAT is a graph attention convolution method, and MPNN is an information transfer method. All three are deep learning methods, and GMP is the abbreviation of the method of this embodiment. It can be seen that compared with the existing methods, the accuracy of the attribute prediction and classification method proposed in this embodiment is the best in most cases, proving that the molecular attribute prediction method proposed in this embodiment has relatively good classification and prediction capabilities.

[0070] Only the preferred embodiments of the present invention are described in detail above, but the present invention is not limited to the above embodiments. Various changes can be made within the knowledge scope of ordinary technicians in this field without departing from the purpose of the present invention, and various changes should be included in the protection scope of the present invention.

Claims

1. A multimodal molecular property prediction method based on cue word graph convolution, characterized in that: The following steps are involved: S1. Selection and establishment of data set: multiple molecular attributes are selected to construct an initial data set of the multimodal method, and the initial data set is divided into a training set, a validation set and a test set, wherein the initial data set contains multiple molecular attributes and their corresponding molecules; S2. Data preprocessing: Perform graph embedding and preliminary chemical feature calculation operations on each molecule in the training set of the initial data set to obtain the features of each molecule in two dimensions; Combining molecules and their corresponding graph and sequence features to obtain a training set of target datasets for training multimodal methods; S3, method construction: constructing a multimodal method, wherein the multimodal method comprises a multimodal fusion module, a feature extraction module and a molecular graph readout module connected in sequence; training the multimodal method using a training set in a target data set, and obtaining a trained multimodal method after verifying that the trained method is qualified through a validation set and a test set; S4. Molecular property prediction: The molecules with properties to be determined are input into the trained multimodal method for property prediction, and the multimodal method outputs the molecular properties to be predicted.

2. A multimodal molecular property prediction method based on prompt word graph convolution according to claim 1, characterized in that: The method for performing graph embedding and preliminary chemical feature calculation operations on each molecule in S2 is: using the RDKit extension package to calculate the chemical features of each molecule in the training set of the initial data set, and embedding them in the form of graphs and sequences.

3. A multimodal molecular property prediction method based on prompt word graph convolution according to claim 2, characterized in that: The method for calculating the chemical characteristics of each molecule in the training set of the initial data set is: S21. Use the RDKit extension package to calculate the chiral and aromatic chemical features of the molecule and embed them as node and bond features to obtain the graph features of the molecule. S22. Use the RDKit extension package to calculate the MACCS fingerprint, PubChem fingerprint, pharmacophore fingerprint and Morgan fingerprint of the molecule to obtain the sequence characteristics of the molecule.

4. The multimodal molecular property prediction method based on prompt word graph convolution according to claim 1, characterized in that: The multimodal molecular property prediction method based on the cue word graph convolution in S3 is: S31, multimodal fusion module: extract the substructure features of the molecule by decoding the one-dimensional molecular fingerprint information; then, map these screened substructure features onto the molecular graph to achieve effective fusion of the one-dimensional sequence features and the two-dimensional molecular graph features; S32, feature extraction module: After fusing multi-dimensional features, a graph convolutional neural network is used to extract features, and a gating mechanism and residual connection are added to control the flow of information, so that the network can better focus on important information in the node; S33, molecular graph readout module: The present invention chooses to construct and output a complete molecular graph based on the elements contained in the molecule, and designs a convolution-based readout module; in the readout module, the molecular graph information processed by graph convolution is extracted and fused into a one-dimensional representation through a three-layer convolutional network.

5. The multimodal molecular property prediction method based on prompt word graph convolution according to claim 4, characterized in that: The method for achieving effective fusion of one-dimensional sequence features and two-dimensional molecular graph features in S31 is: S311, initializing the molecular graph into an undirected graph, in which atoms are represented as nodes and chemical bonds are represented as edges; in the graph, each node is represented by 6 atomic features and each edge is represented by 4 bond features; Most of the features used to characterize nodes and edges are calculated directly using RDKit. For ease of processing, these features are encoded using one-hot encoding. S312, using the Morgan fingerprint in dictionary form to extract substructure information in the molecule, after obtaining the dictionary containing the substructure information, using the hash operation function in RDKit to map the substructure fragments back to the original molecule; in order to ensure the accuracy of the extracted information, only the substructure information that can be uniquely mapped is mapped back to the original molecule; S313. Add a predefined value to the last-dimensional embedding feature of the corresponding atom and chemical bond to indicate that the atom and chemical bond are located in a substructure fragment.

6. The multimodal molecular property prediction method based on prompt word graph convolution according to claim 4, characterized in that: The method for extracting and fusing the molecular graph information processed by graph convolution into a one-dimensional representation through a three-layer convolutional network in S33 is: S331, reading out the molecular graph obtained by graph convolution processing according to element categories to obtain a molecular feature graph classified according to elements; S332. Use dilated convolution to perform convolution operation on each molecular feature map. The operation process is divided into three layers. The number of convolutions in each layer is inversely proportional to the number of layers. The final representation that integrates all molecular features is obtained: Among them, n is the number of convolutions in each layer, C i is the output of each convolutional layer, C j is the output of the jth convolution block in each layer. By using multiple identical convolution units to extract features and then averaging the features, the model's dependence on a single convolution block is reduced, thereby increasing the model's stability. At the same time, each convolution block can extract different features from the data, making the extracted features more comprehensive and rich.

7. The multimodal molecular property prediction method based on prompt word graph convolution according to claim 6, characterized in that: The classification method is evaluated using the area under the ROC curve AUC, and the prediction method is evaluated using the root mean square error RMSE.

Images