Method for predicting properties of drug molecules and related devices

Abstract

The application discloses a drug molecule property prediction method and a related device, and belongs to the technical field of quantum computing. The method comprises the following steps: calculating the expected value of various edge types in a molecular structure graph through a quantum circuit, and taking the expected value as a first edge feature of the molecular structure graph, wherein the edges in the molecular structure graph are used to represent the chemical bonds of a drug molecule; performing noise reduction on the first edge feature through a clustering algorithm to obtain a second edge feature; constructing a new molecular structure graph according to the second edge feature and original node features in the molecular structure graph; and performing chemical property prediction on the drug molecule according to the graph features of the new molecular structure graph. The application applies the QRGCN model to drug molecule property prediction for the first time, opens up a new possibility for quantum acceleration processing of complex graph data, effectively suppresses the interference of quantum noise through the application of a clustering method, and greatly improves the robustness and computing performance of the QRGCN model in a real quantum environment.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of quantum computing, and particularly relates to a drug molecule property prediction method and a related device. BACKGROUND

[0002] At present, in the drug research and development process, accurately predicting various properties of candidate drug molecules, such as drug solubility, toxicity, etc., is crucial for accelerating drug screening. As an advanced quantum neural network model, the Quantum Relational Graph Convolutional Network (QRGCN) can efficiently process molecular structure graphs through the advantages of quantum computing, which is expected to greatly improve the prediction performance. However, when applying the QRGCN model to the prediction of drug molecule properties, due to the inherent noise in the quantum chip, the QRGCN model cannot be calculated according to the processing logic of the traditional relational graph convolutional network, which greatly reduces the performance of the model.

[0003] Therefore, the technical problem to be solved is to propose a method to reduce the influence of quantum chip noise to improve the performance advantage of QRGCN model in complex data processing. SUMMARY

[0004] The purpose of the application is to provide a drug molecule property prediction method and related device, which aims to suppress the interference of inherent noise in the quantum chip and greatly improve the robustness and computing performance of the QRGCN model in the real quantum environment.

[0005] One embodiment of the application provides a drug molecule property prediction method, which comprises:

[0006] calculating the expected value of each edge type in the molecular structure graph through a quantum circuit, and taking the expected value as the first edge feature of the molecular structure graph, wherein the edges in the molecular structure graph are used to represent the chemical bonds of the drug molecule;

[0007] de-noising the first edge feature through a clustering algorithm to obtain a second edge feature;

[0008] constructing a new molecular structure graph according to the second edge feature and the original node feature in the molecular structure graph;

[0009] performing chemical property prediction on the drug molecule according to the graph feature of the new molecular structure graph.

[0010] Optionally, the calculation of the expected value of each edge type in the molecular structure graph comprises:

[0011] obtaining original edge features of a molecular structure graph, and encoding the original edge features into parameters of quantum logic gates;

[0012] configuring a quantum circuit according to the parameters, and running the quantum circuit to obtain an output quantum state;

[0013] calculating expected values of various edge types in the molecular structure graph according to the output quantum state.

[0014] Optionally, the encoding the original edge features into parameters of quantum logic gates comprises:

[0015] encoding the original edge features into parameters of quantum logic gates according to a mapping function, the mapping function being:

[0016]

[0017] wherein, the is an encoded weight, the is an encoded bias, and the is an i-th edge feature of a j-th molecular structure graph.

[0018] Optionally, the number of the first edge features is a plurality, and the denoising the first edge features by a clustering algorithm to obtain second edge features comprises:

[0019] clustering the plurality of first edge features by a clustering algorithm to obtain a clustering result;

[0020] grouping the plurality of first edge features into predefined edge categories according to the clustering result, and generating an index for each edge category to obtain second edge features.

[0021] Optionally, before the predicting the chemical properties of the drug molecule according to the graph features of the new molecular structure graph, the method further comprises:

[0022] determining the graph features according to a graph convolution calculation formula, the graph convolution calculation formula being:

[0023]

[0024] wherein, A is an adjacency matrix determined based on the second edge features, D is a degree matrix, is a node feature of an l-th layer graph convolution network, is a weight of the l-th layer graph convolution network, and σ is an activation function.

[0025] Another embodiment of the present application provides a drug molecule property prediction device, the device comprising:

[0026] a calculation unit configured to calculate expectation values of various edge types in a molecular structure graph via a quantum circuit, and use the expectation values as first edge features of the molecular structure graph, wherein the molecular structure graph comprises a plurality of nodes and a plurality of edges, and each edge is used to represent a chemical bond of a drug molecule

[0027] the edge is used to represent a chemical bond of a drug molecule

[0028] a denoising unit configured to denoise the first edge features via a clustering algorithm to obtain second edge features

[0029] a reconstruction unit configured to construct a new molecular structure graph according to the second edge features and original node features in the molecular structure graph

[0030] a prediction unit configured to perform chemical property prediction on the drug molecule according to graph features of the new molecular structure graph

[0031] Optionally, in the aspect of calculating the expectation values of various edge types in the molecular structure graph via the quantum circuit, the calculation unit specifically comprises:

[0032] obtaining original edge features of the molecular structure graph, and encoding the original edge features into parameters of quantum logic gates

[0033] configuring the quantum circuit according to the parameters, and running the quantum circuit to obtain an output quantum state

[0034] calculating the expectation values of various edge types in the molecular structure graph according to the output quantum state

[0035] Optionally, in the aspect of encoding the original edge features into the parameters of the quantum logic gates, the calculation unit specifically comprises:

[0036] encoding the original edge features into the parameters of the quantum logic gates according to a mapping function, wherein the mapping function is:

[0037]

[0038] wherein, the is an encoded weight, the is an encoded bias, and the is an i-th edge feature of a j-th molecular structure graph

[0039] Another embodiment of the present application provides an electronic device, comprising: a processor and a memory;

[0040] The processor and the memory are connected, wherein the memory is used to store a computer program, and the processor is used to call the computer program to execute the method described in any of the above embodiments.

[0041] Still another embodiment of the present application provides a computer readable storage medium storing a computer program, the computer program comprising program instructions which, when executed by a processor, perform the method described in any of the above embodiments.

[0042] Compared with the prior art, the present application provides a drug molecule property prediction method, which first calculates the expected value of various edge types in a molecular structure graph through a quantum circuit, and takes the expected value as the first edge feature of the molecular structure graph, wherein the edges in the molecular structure graph are used to represent the chemical bonds of a drug molecule

[0043] ; then performs noise reduction on the first edge feature through a clustering algorithm to obtain a second edge feature; then constructs a new molecular structure graph according to the second edge feature and the original node feature in the molecular structure graph; and finally performs chemical property prediction on the drug molecule according to the graph feature of the new molecular structure graph.

[0044] The present application first calculates the expected value of various edge types in a molecular structure graph through a quantum circuit, and takes the expected value as the first edge feature in the molecular structure graph, wherein the quantum circuit is used to simulate the chemical bond types in the drug molecule and calculate the expected value of the chemical bond types as the first edge feature, thereby providing more accurate basic data for subsequent prediction; the first edge feature is subjected to noise reduction processing through a clustering algorithm to obtain a second edge feature, wherein the clustering effectively identifies and removes noise and redundant information in the first edge feature, improves the reliability of the data, and the second edge feature after noise reduction can more accurately reflect the key information of the drug molecule, which is conducive to the accuracy of subsequent analysis; a new molecular structure graph is constructed by combining the second edge feature after noise reduction and the original node feature, wherein the new molecular structure graph retains the original structure information and integrates the accuracy of quantum calculation and the noise reduction effect of the clustering algorithm, thereby providing a more reliable basis for subsequent chemical property prediction; and finally, the drug molecule is subjected to chemical property prediction based on the graph feature of the new molecular structure graph, which can more accurately reflect the real structure and properties of the drug molecule, improve the prediction accuracy, and accelerate the screening and optimization process of the drug. BRIEF DESCRIPTION OF DRAWINGS

[0045] Figure 1 A network block diagram of a drug molecule property prediction system provided by an embodiment of the present application.

[0046] Figure 2 A flowchart of a drug molecule property prediction method provided by an embodiment of the present application.

[0047] Figure 3 A flowchart of an edge type expected value calculation method provided by an embodiment of the present application.

[0048] Figure 4 A flow chart of an edge feature clustering method provided by an embodiment of the present application.

[0049] Figure 5 A structural schematic diagram of a drug molecule property prediction device provided by an embodiment of the present application.

[0050] Figure 6 A structural schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0051] The embodiments described below with reference to the drawings are exemplary and are only used to explain the present application, and cannot be explained as a limitation of the present application.

[0052] Figure 1 A network block diagram of a drug molecule property prediction system provided by an embodiment of the present application. The drug molecule property prediction system can include a network 110, a server 120, a wireless device 130, a client 140, a storage 150, a classical computing unit 160, a quantum computing unit 170, and can also include additional storage, classical processors, quantum processors and other devices not shown.

[0053] The network 110 is a medium for providing a communication link between various devices and computers connected together in the drug molecule property prediction system, including but not limited to the Internet, an intranet, a local area network, a mobile communication network and a combination thereof, and the connection mode can adopt a wired, wireless communication link or an optical fiber cable, etc.

[0054] The server 120, the wireless device 130 and the client 140 are conventional data processing systems, which can contain data and have application programs or software tools for performing conventional computing processes. The client 140 can be a personal computer or a network computer, so the data can also be provided by the server 120. The wireless device 130 can be a smart phone, a tablet, a notebook computer, a smart wearable device, etc. The storage unit 150 can include a database 151, which can be configured to store quantum qubit parameters, quantum logic gate parameters, quantum circuits, quantum programs and other data.

[0055] The classical computing unit 160 (quantum computing unit 170) can include a classical processor 161 (quantum processor 171) for processing classical data (quantum data) and a memory 162 (memory 172) for storing classical data (quantum data), and the classical data (quantum data) can be a boot file, an operating system image, and an application 163 (application 173), which can be used to implement a quantum algorithm compiled according to the drug molecule property prediction method provided by an embodiment of the present application.

[0056] Any data or information stored or generated in the classical computing unit 160 (quantum computing unit 170) can also be configured to be stored or generated in another classical (quantum) processing system in a similar manner, and any application program executed by it can also be configured to be executed in another classical (quantum) processing system in a similar manner.

[0057] It should be noted that a real quantum computer is a hybrid structure, which includes at least two parts: a classical computing unit 160 responsible for performing classical computing and control; and a quantum computing unit 170 responsible for running quantum programs to implement quantum computing. Figure 1

[0058] The classical computing unit 160 and the quantum computing unit 170 described above can be integrated in one device or distributed in two different devices. For example, a first device including the classical computing unit 160 runs a classical computer operating system, on which quantum application development tools and services are provided, and storage and network services required by quantum applications are also provided. A user develops a quantum program through the quantum application development tools and services thereon, and sends the quantum program to a second device including the quantum computing unit 170 through the network services thereon. The second device runs a quantum computer operating system, which parses and compiles the code of the quantum program into instructions that can be recognized and executed by the quantum processor 170, and the quantum processor 170 implements the quantum algorithm corresponding to the quantum program according to the instructions.

[0059] The computing unit of the classical processor 161 in the classical computing unit 160 is based on CMOS tubes of silicon chips, and such a computing unit is not limited by time and coherence, i.e., such a computing unit is not limited by the length of use and is available at any time. In addition, in silicon chips, the number of such computing units is also sufficient, and the number of computing units in a classical processor 161 is currently thousands or even tens of thousands. The number of computing units is sufficient and the computing logic of CMOS tubes is fixed, for example: AND logic. When operating with CMOS tubes, a large number of CMOS tubes are combined with limited logic functions to achieve the effect of operation.

[0060] ​The basic computing unit of the quantum processor 171 in the quantum computing unit 170 is a quantum bit, the input of the quantum bit is limited by coherence, and is also limited by coherence time, that is, the quantum bit is limited by the use time and is not available at any time. Sufficient use of the quantum bit within the available use time of the quantum bit is a key problem of quantum computing. In addition, the number of quantum bits in the quantum computer is one of the representative indicators of the performance of the quantum computer, and each quantum bit realizes a computing function through a logically configured function. In view of the limited number of quantum bits and the diversified logic functions in the field of quantum computing, such as Hadamard gate (H gate), Pauli-X gate (X gate), Pauli-Y gate (Y gate), Pauli-Z gate (Z gate), X gate, RY gate, RZ gate, CNOT gate, CR gate, iSWAP gate, Toffoli gate, and the like. When quantum computing, the limited quantum bits need to be combined with various logic functions to achieve the operation effect.

[0061] Based on these differences, the design of the classical logic function acting on the CMOS tube and the design of the quantum logic function acting on the quantum bit are significantly and essentially different; the design of the classical logic function acting on the CMOS tube does not need to consider the individuality of the CMOS tube, such as the individual identification, position, and available time of each CMOS tube in the silicon chip. Therefore, the classical algorithm composed of the classical logic function only expresses the operation relationship of the algorithm, and does not express the dependence of the algorithm on the individuality of the CMOS tube.

[0062] And the quantum logic function acting on the quantum bit needs to consider the individuality of the quantum bit, such as the individual identification, position, and relationship with the surrounding quantum bits of the quantum bit in the quantum chip, as well as the available time of each quantum bit. Therefore, the quantum algorithm composed of the quantum logic function not only expresses the operation relationship of the algorithm, but also expresses the dependence of the algorithm on the individuality of the quantum bit.

[0063] The quantum chip can include quantum bits and channels for regulating the quantum bits, and the quantum logic gate is realized by analog signals. Different combinations of analog signals are applied to the quantum bits through the channels for regulating the quantum bits, so as to realize quantum circuits with different functions and complete the processing of data. Therefore, the design of the quantum logic function acting on the quantum bit (including the design of whether to use the quantum bit and the design of the use efficiency of each quantum bit) is the key to improving the operation performance of the quantum computer, and special design is required. This is also the uniqueness of the quantum algorithm based on the quantum logic function, which is essentially and significantly different from the classical algorithm based on the classical logic function. The above design for the quantum bit is a technical problem that ordinary computing devices do not need to consider and face.

[0064] In the field of quantum computing, when running a QRGCN model on a real quantum chip, the inherent noise of the actual quantum chip will disturb the quantum state representation of the chemical bond type, causing the QRGCN model to be unable to run as expected; while the QRGCN model theoretically requires a clear and fixed edge type definition, so the type discretization caused by the disturbance of noise becomes a key obstacle to its application in a real quantum environment.

[0065] Therefore, in order to overcome the inherent noise disturbance of the quantum chip and ensure the effective application of the QRGCN model in a real quantum environment, the present application proposes a drug molecule property prediction method, device, electronic equipment and computer readable storage medium.

[0066] Therefore, in order to overcome the inherent noise disturbance of the quantum chip and ensure the effective application of the QRGCN model in a real quantum environment, the present application proposes a drug molecule property prediction method, device, electronic equipment and computer readable storage medium.

[0067] Referring to Figure 2 , Figure 2 A drug molecule property prediction method provided by an embodiment of the present application comprises the following steps:

[0068] Step S201, calculating the expected value of each edge type in the molecular structure graph through a quantum circuit, and taking the expected value as the first edge feature of the molecular structure graph, wherein the edges in the molecular structure graph are used to represent the chemical bonds of the drug molecule;

[0069] Wherein, the edge refers to the connection relationship between atoms in the molecular structure graph, and the atoms are usually connected by chemical bonds, the edge type is the type of the chemical bond, for example, the edge type can include single chemical bond, double chemical bond, multiple chemical bond, etc. The first edge feature refers to the expected value of the chemical bond obtained by simulating calculation through a quantum circuit using quantum computing technology.

[0070] Specifically, the chemical bond types in the drug molecule are simulated through a quantum circuit, and the expected value of the chemical bond types is calculated. These expected values reflect the average performance of the chemical bond in the quantum system, such as strength, stability and other key properties. The calculated expected value is then taken as the first edge feature of the molecular structure graph, providing important input information for subsequent drug molecule property prediction.

[0071] For example, assume that a new drug molecule of an "X-Drug" is being studied, and in order to understand the chemical properties and potential biological activity of the drug molecule, quantum computing technology is used to predict its properties. First, the expected value of different edges in the molecular structure graph is calculated by constructing a quantum circuit, where the edges in the molecular structure graph represent chemical bonds, including single bonds, double bonds, and aromatic bonds, etc. The calculated expected value represents the average performance of the chemical bond in the quantum system, which can reflect the strength and stability of the chemical bond. For example, it may be found that the expected value of a double bond is high, indicating that it exhibits strong interaction and stability in the quantum system, while the expected value of a single bond may be low, meaning that it is weak. The calculated expected value is used as the first edge feature of the "X-Drug" molecular structure graph, which provides important input information for the subsequent machine learning model to predict various properties of the "X-Drug" molecule, such as anticancer activity, pharmacokinetic parameters, etc.

[0072] Step S202, denoising the first edge feature by a clustering algorithm to obtain a second edge feature;

[0073] The clustering algorithm refers to dividing the first edge feature dataset into K clusters by the K-means algorithm, randomly selecting K data points as initial centroids, and iterating the data based on the Euclidean distance until the centroids are unchanged, the algorithm converges, and the clustering is completed.

[0074] Specifically, the K-means algorithm effectively identifies and removes noise and redundant information in the first edge feature, resulting in more accurate second edge features as important inputs for subsequent data prediction analysis.

[0075] Step S203, constructing a new molecular structure graph according to the second edge feature and the original node feature in the molecular structure graph;

[0076] The original node feature refers to the inherent attribute value of the atomic node in each molecular structure graph without any transformation processing.

[0077] Specifically, the second edge feature set after denoising processing and the original node feature are integrated to form a complete molecular structure representation.

[0078] Exemplary, taking ethane molecule as an example, assuming that its original node features only include atom types, carbon or hydrogen, representing these features by one-hot encoding, the feature vector of carbon atom (C) = [1, 0], the feature vector of hydrogen atom (H) = [0, 1], through quantum circuit calculation, the expected values of various edge types in the ethane molecule are obtained, assuming that the expected value of C-C bond = 0.9, the expected value of C-H bond = 0.8, in this example, it is assumed that the denoising process has no significant effect on the expected value, so the first edge feature (expected value) is directly used as the second edge feature, but in actual application, the denoising process may change these values; finally, a new molecular structure graph is constructed, and the second edge feature set and the original node feature set are integrated, in order to facilitate the explanation, this process is conceptualized as a "weighted graph", the new ethane molecular structure graph will have two carbon atom nodes (each node has a [1, 0] feature vector) and six hydrogen atom nodes (each node has a [0, 1] feature vector), the edges will include two C-C bonds (each bond has a weight of 0.9) and six C-H bonds (each bond has a weight of 0.8).

[0079] In step S204, the chemical property of the drug molecule is predicted according to the graph features of the new molecular structure graph.

[0080] Among them, the graph features refer to more accurate feature representations extracted by the classical convolutional neural network from the comprehensive graph representation of the new molecular structure graph.

[0081] Specifically, the comprehensive graph representation of the constructed new molecular structure graph is input into the classical convolutional neural network to obtain the graph features, the graph features are converted into label probabilities through the activation function, and the chemical property prediction result of the drug molecule is obtained through the label probabilities.

[0082] Exemplary, the comprehensive graph vector representation of the constructed new molecular structure graph is input into the classical convolutional neural network, which can extract key features in the molecular structure graph, such as the type of node (different atom types), the type of edge (different types of chemical bonds) and the neighbor relationship of node, etc., and encode these features into more accurate feature vector representation, i.e. graph features; then the graph features are sent to the output layer of the model, and are converted into label probabilities through the activation function (such as softmax function), the label probabilities represent the possibility of the drug molecule having various chemical properties, for example, for the property of anti-tumor activity, the model may output a probability value between 0 and 1, the closer to 1 indicates that the drug molecule is more likely to have anti-tumor activity.

[0083] In summary, the present application firstly calculates the expected value of various edge types in the molecular structure graph through quantum circuit, and takes the expected value as the first edge feature in the molecular structure graph, simulates the chemical bond type in the drug molecule through quantum circuit, and calculates the expected value of the chemical bond type as the first edge feature, which provides more accurate basic data for subsequent prediction; the first edge feature is denoised by the clustering algorithm to obtain the second edge feature, the clustering effectively identifies and removes the noise and redundant information in the first edge feature, improves the reliability of the data, and the second edge feature after denoising can more accurately reflect the key information of the drug molecule, which is beneficial to the accuracy of subsequent analysis; a new molecular structure graph is constructed by the second edge feature after denoising and the original node feature, the new molecular structure graph retains the original structure information and integrates the accuracy of quantum calculation and the denoising effect of the clustering algorithm, which provides a more reliable basis for subsequent chemical property prediction; finally, the chemical property of the drug molecule is predicted based on the graph feature of the new molecular structure graph, which can more accurately reflect the real structure and property of the drug molecule, improve the prediction accuracy, and speed up the screening and optimization process of the drug.

[0084] In step S301, the original edge feature of the molecular structure graph is obtained, and the original edge feature is encoded as the parameter of the quantum logic gate.

[0085] The original edge feature refers to obtaining the molecular structure graph of the drug molecule through chemical software or database, and the connection relationship between atoms in the drug molecule, i.e. chemical bond (edge), is described in the structure graph. These edges have various attributes, such as bond type (single bond, double bond, triple bond, etc.), and then the attribute features are converted into numerical vector representation, for example, a unique integer value can be assigned to each different bond type or a multi-dimensional vector can be used to represent bond length, bond angle and other continuous features, i.e. original edge feature.

[0086] Specifically, the feature vector representation of the original edge in the molecular structure graph is first obtained, and then converted into the parameter of the quantum logic gate through a specific mapping function.

[0087] For example, the original edge feature is encoded into the parameter of the quantum logic gate through the mapping function, and the specific mapping function is wherein is the weight of the encoding, is the bias of the encoding,

[0088] is the jth edge feature of the ith molecular structure graph, assuming that the molecular structure graph of compound A has two edge features, bond length and bond angle , the two features are respectively encoded into the parameters of two quantum logic gates and , assuming encoding weight = 0.5, = 1.0 and encoding bias =0.1, =0.2, and bond length is 1.5 and bond angle is 120 degrees; then the corresponding quantum logic gate parameters can be calculated as: = 0.5 *1.5 + 0.1 = 0.85, = 1.0 *120 + 0.2 =120.2.

[0089] Step S302, configuring a quantum circuit according to the parameters, and running the quantum circuit to obtain an output quantum state;

[0090] The output quantum state is composed of a series of quantum states and their corresponding probabilities.

[0091] Specifically, the quantum circuit is configured according to the quantum encoding parameters, and the output quantum state is obtained by running the quantum circuit. The output quantum state is usually represented as the state of one or more quantum bit sets and their corresponding probability distribution.

[0092] For example, assume that a quantum circuit has two quantum bits, and the output of the quantum circuit is a mixed state containing two possible pure states;

[0093] Pure state A: the first quantum bit is 0 and the second quantum bit is 0, i.e., state |00 ;

[0094] Pure state B: the first quantum bit is 1 and the second quantum bit is 1, i.e., state |11 ;

[0095] The density matrix ρ of the mixed state can be represented by a linear combination of two projection operators, each corresponding to a pure state;

[0096] The matrix corresponding to |00><00| is: ; The matrix corresponding to |11><11| is: ;

[0097] The mixed state ρ is composed of a linear combination of the two projection operators, where and are probabilities satisfying + =1; that is, the mixed state can be represented as:

[0098]

[0099] The measurement is performed by running the quantum circuit. Due to the randomness of quantum states, different results can be obtained each time the measurement is performed. Therefore, a sufficient number of shots (for example, 1000 times) are performed, the number of occurrences of the pure state A and the pure state B is counted, and the probabilities of the pure state A and the pure state B are calculated and , assuming that the pure state A (|00 ) occurs 550 times and the pure state B (|11 ) occurs 450 times, the probabilities of the pure states and the corresponding probabilities in the mixed state are estimated as follows: = 0.55, = 0.45.

[0100] In step S303, the expected values of various edge types in the molecular structure graph are calculated according to the output quantum state.

[0101] Specifically, the expected values of each edge type are calculated according to the output of the quantum circuit, which are the first edge features of the molecular structure graph.

[0102] In summary, by encoding the original edge features of the molecular structure graph into the parameters of the quantum logic gate and configuring and running the quantum circuit accordingly, the expected values of various edge types in the molecular structure graph can be efficiently calculated, thereby providing a new molecular structure analysis method based on quantum computing, which helps to more accurately understand the interaction and properties between molecules, and further improves the accuracy and reliability of the prediction of the properties of drug molecules.

[0103] Referring to Figure 4 , Figure 4 a flowchart of an edge feature clustering method provided by an embodiment of the present application, the method comprising the following steps:

[0104] In step S401, a plurality of first edge features are clustered by a clustering algorithm to obtain a clustering result.

[0105] The clustering result refers to a second edge feature obtained by performing noise reduction processing on the first edge feature.

[0106] Specifically, the first edge feature is processed by the K-means clustering algorithm to obtain the second edge feature.

[0107] For example, in this process, similar edge features are effectively clustered together through the K-means algorithm. Assume that there are three types of bonds in the drug molecule structure diagram: single bond, double bond and triple bond. In an ideal noise-free environment, after these bond types are processed by variational quantum circuit transformation, their numerical features are clearly mapped to different intervals. Assume that the value of a single bond varies from 1 to 10, the value of a double bond varies from 2 to 20, and the value of a triple bond varies from 3 to 30; however, in actual applications, the existence of quantum inherent noise requires running quantum circuits in a noisy environment. Assume that the quantum circuit is run 10 times for each bond type. For a single bond, 5 different results are obtained. Some results are centered around the standard value of 10, showing slight fluctuations, namely 10, 10.1, 9.8, 10.2... etc. (the standard value 10 may appear multiple times). The double bond results are centered around the standard value of 20, 20, 20.2, 19.8, 20... etc. (the standard value 20 may appear multiple times). The triple bond results are centered around the standard value of 30, 30, 30.2, 29.8, 30... etc. (the standard value 30 may appear multiple times). At this time, the interval disturbance range with the standard value of 10 is treated as the first cluster result, the interval disturbance range with the standard value of 20 is treated as the second cluster result, and the interval disturbance range with the standard value of 30 is treated as the third cluster result.

[0108] Step S402 : classifying the plurality of first edge features into predefined edge categories according to the clustering result, and generating an index for each edge category to obtain a second edge feature.

[0109] The predefined edge category refers to how many types of chemical bonds (edges) for a drug molecule are known in advance before clustering; the index can be a unique identifier (such as an integer, a string, etc.).

[0110] Specifically, based on the clustering results, similar edge features are classified into predefined edge categories, and an index is generated for each edge category for fast retrieval and identification of the edge category. The edge category with the index is called the second edge feature.

[0111] For example, from the above example, it can be seen that the first clustering result of the interval perturbation range with a standard value of 10 is divided into a single-key category, the second clustering result of the interval perturbation range with a standard value of 20 is divided into a double-key category, and the third clustering result of the interval perturbation range with a standard value of 30 is divided into a triple-key category; and an index 1 identifier is generated for the single-key category, an index 2 identifier is generated for the double-key category, and an index 3 identifier is generated for the triple-key category for quick retrieval and identification of the edge category.

[0112] In summary, by clustering the first edge features, the noise and redundant information can be separated from the main edge features to obtain the second edge features, the second edge features better retain the essential features of the original data, enhance the generalization ability of the model, and each edge category corresponds to an index, so that subsequent data processing can be quickly located to the relevant edge features based on the indexes, greatly improving the efficiency of data processing.

[0113] In an embodiment of the present application, before the chemical property of the drug molecule is predicted according to the graph features of the new molecular structure graph, the method further comprises:

[0114] The graph features are determined according to a graph convolution calculation formula, and the graph convolution calculation formula is:

[0115]

[0116] Wherein, A is an adjacency matrix determined based on the second edge features, D is a degree matrix, is the node feature of the lth layer graph convolution network, is the weight of the lth layer graph convolution network, and σ is an activation function.

[0117] Specifically, the second edge features obtained by noise reduction processing, a new molecular structure graph is constructed according to the second edge features and the original node features in the molecular structure graph, the comprehensive graph representation of the constructed new molecular structure is input into a classical convolutional neural network to obtain graph features, the graph features are converted into label probability through an activation function, and the prediction result is obtained through the label probability.

[0118] Referring to Figure 5 , Figure 5 A drug molecule property prediction device is provided in an embodiment of the present application, and the device can include:

[0119] The computing unit 501 is configured to calculate the expected value of each edge type in the molecular structure graph by using a quantum circuit, and the expected value is used as the first edge feature of the molecular structure graph, and the edges in the molecular structure graph are used to represent the chemical bonds of the drug molecule.

[0120] Wherein, the edge refers to the connection relationship between atoms in the molecular structure graph, and can also include information about the properties of the chemical bond, the edge type refers to the type of connection relationship, and the first edge feature refers to using quantum computing technology to simulate the type of chemical bond in the drug molecule by using a quantum circuit and calculating the expected value of the type of chemical bond.

[0121] Specifically, in terms of calculating the expected value of each edge type in the molecular structure graph by using a quantum circuit, the computing unit 501 specifically includes:

[0122] obtain original edge features of a molecular structure graph, and encode the original edge features into parameters of quantum logic gates;

[0123] configure a quantum circuit according to the parameters, and run the quantum circuit to obtain an output quantum state;

[0124] calculate expected values of various edge types in the molecular structure graph according to the output quantum state.

[0125] Specifically, in the aspect of encoding the original edge features into parameters of quantum logic gates, the computing unit specifically comprises:

[0126] encoding the original edge features into parameters of quantum logic gates according to a mapping function, the mapping function being:

[0127]

[0128] wherein, the is an encoded weight, the is an encoded bias, and the is an i-th edge feature of a j-th molecular structure graph.

[0129] a denoising unit 502, configured to denoise the first edge features by a clustering algorithm to obtain second edge features;

[0130] a reconstruction unit 503, configured to construct a new molecular structure graph according to the second edge features and original node features in the molecular structure graph;

[0131] a prediction unit 504, configured to perform chemical property prediction on the drug molecule according to graph features of the new molecular structure graph.

[0132] Please refer to Figure 6 The embodiments of the present specification also provide an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the drug molecule property prediction method in any of the above embodiments when executing the computer program. Please refer to Figure 6 The electronic device can be a computer device, which can be a classical computer or a quantum computer.

[0133] The embodiments of the present specification also provide a computer-readable storage medium, which stores a computer program, and the computer program makes the computer execute the drug molecule property prediction method in any of the above embodiments when executed by the computer.

[0134] It can be understood that the specific examples in the present specification are only to help those skilled in the art better understand the embodiments of the present specification, and not to limit the scope of the present application.

[0135] It can be understood that, in various embodiments of the present specification, the size of the sequence number of each process does not mean the order of execution, and the execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present specification.

[0136] It can be understood that the various embodiments described in the present specification can be implemented alone or in combination, and the embodiments of the present specification do not limit this.

[0137] Unless otherwise specified, all technical and scientific terms used in the embodiments of the present specification have the same meanings as those commonly understood by those skilled in the art of the present specification. The terms used in the present specification are only for the purpose of describing the specific embodiments and are not intended to limit the scope of the present specification. The term "and / or" used in the present specification includes any and all combinations of one or more related listed terms. The singular forms "a", "an" and "the" used in the embodiments of the present specification and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise.

[0138] It can be understood that the processor of the embodiments of the present specification can be an integrated circuit chip with signal processing capability. In the implementation process, each step of the above method embodiments can be completed by the integrated logic circuit of hardware in the processor or the instruction in the form of software. The above processor can be a general processor, a digital signal processor (Digital Signal Processor, DSP), an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field programmable gate array (Field Programmable Gate Array, FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component. The disclosed methods, steps and logic block diagrams in the embodiments of the present specification can be implemented or executed. The general processor can be a microprocessor or the processor can also be any conventional processor or the like. The steps of the method disclosed in combination with the embodiments of the present specification can be directly embodied as a hardware code processor for execution, or a combination of hardware and software modules in the code processor for execution. The software module can be located in a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register or other mature storage medium in the art. The storage medium is located in the memory, and the processor reads the information in the memory, and combines the hardware to complete the steps of the above method.

[0139] It is appreciated that the memory in the embodiments of the present specification can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM). It should be noted that the memory of the system and method described herein is intended to include, but not limited to, these and any other suitable types of memory.

[0140] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in connection with the embodiments disclosed herein can be realized in electronic hardware, or a combination of computer software and electronic hardware. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present specification.

[0141] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the above-described system, device and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be repeated here.

[0142] In several embodiments provided in the present specification, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the above-described device embodiments are merely schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0143] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, i.e. can be located in one place, or can be distributed on multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the present embodiment.

[0144] In addition, each functional unit in each embodiment of the present specification can be integrated in one processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit.

[0145] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present specification or the parts of the technical solutions that essentially contribute to the prior art can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present specification. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various media that can store program codes.

[0146] The above is only a specific embodiment of the present specification, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical scope disclosed in the present specification, which should be covered within the protection scope of the present specification. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for predicting drug molecular properties, characterized in that: The method comprises: Calculating expected values ​​of various edge types in a molecular structure graph through quantum circuits, and using the expected values ​​as first edge features of the molecular structure graph, where the edges in the molecular structure graph are used to represent chemical bonds of drug molecules; the expected values ​​represent average performance of the chemical bonds in the quantum circuits; Clustering the plurality of first edge features using a clustering algorithm to obtain a clustering result, classifying the plurality of first edge features into predefined edge categories based on the clustering result, and generating an index for each edge category to obtain a second edge feature; wherein the predefined edge category refers to the number of chemical bond types known in advance for a drug molecule before clustering; and constructing a new molecular structure graph based on the second edge features and the original node features in the molecular structure graph; The chemical properties of the drug molecule are predicted based on the graph features of the new molecular structure graph.

2. The method according to claim 1, characterized in that The expected values ​​of various edge types in the molecular structure graph are calculated by quantum circuits, including: Obtaining original edge features of a molecular structure graph, and encoding the original edge features into parameters of a quantum logic gate; Configuring a quantum circuit according to the parameters, and running the quantum circuit to obtain an output quantum state; Expected values ​​of various edge types in the molecular structure graph are calculated based on the output quantum state.

3. The method according to claim 2, characterized in that The encoding of the original edge features into parameters of a quantum logic gate includes: The original edge features are encoded into parameters of a quantum logic gate according to a mapping function, wherein the mapping function is: i ij =W*e ij +b Wherein, W is the encoding weight, b is the encoding bias, and e ij is the jth edge feature of the i-th molecular structure graph.

4. The method according to claim 1, wherein Before predicting the chemical properties of the drug molecule based on the graph features of the new molecular structure graph, the method further includes: The graph features are determined according to a graph convolution calculation formula, which is: Where A is the adjacency matrix determined based on the second edge feature, D is the degree matrix, and H l is the node feature of the l-th layer graph convolutional network, W l is the weight of the l-th layer graph convolutional network, and σ is the activation function.

5. A drug molecular property prediction device, characterized in that: The device comprises: a calculation unit, configured to calculate expected values ​​of various edge types in a molecular structure graph through a quantum circuit, and use the expected values ​​as first edge features of the molecular structure graph, wherein the edges in the molecular structure graph are used to represent chemical bonds of drug molecules; the expected values ​​represent average performance of the chemical bonds in the quantum circuit; a denoising unit, configured to cluster the plurality of first edge features using a clustering algorithm to obtain a clustering result, classify the plurality of first edge features into predefined edge categories based on the clustering result, and generate an index for each edge category to obtain a second edge feature; wherein the predefined edge category refers to the number of chemical bond types that are known in advance for a drug molecule before clustering; and a reconstruction unit, configured to construct a new molecular structure graph based on the second edge features and the original node features in the molecular structure graph; The prediction unit is used to predict the chemical properties of the drug molecule according to the graph features of the new molecular structure graph.

6. The device according to claim 5, characterized in that In the aspect of calculating the expected values ​​of various edge types in the molecular structure graph by using quantum circuits, the calculation unit specifically includes: Obtaining original edge features of a molecular structure graph, and encoding the original edge features into parameters of a quantum logic gate; Configuring a quantum circuit according to the parameters, and running the quantum circuit to obtain an output quantum state; Expected values ​​of various edge types in the molecular structure graph are calculated based on the output quantum state.

7. The device according to claim 6, characterized in that In terms of encoding the original edge features into parameters of a quantum logic gate, the calculation unit specifically includes: The original edge features are encoded into parameters of a quantum logic gate according to a mapping function, wherein the mapping function is: i ij =W*e ij +b Wherein, W is the encoding weight, b is the encoding bias, and e ij is the jth edge feature of the i-th molecular structure graph.

8. An electronic device, characterized in that: include: processor and memory; The processor is connected to a memory, wherein the memory is used to store a computer program, and the processor is used to call the computer program to execute the method according to any one of claims 1 to 4.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, wherein the computer program includes program instructions. When the program instructions are executed by a processor, the method according to any one of claims 1 to 4 is executed.

Images