Processing method and device for atomic scale reaction site prediction

Abstract

The embodiment of the invention relates to a processing method and device for atomic scale reaction site prediction. The method comprises the following steps: constructing a reaction site prediction model; constructing a first original data set through big data collection; performing model training data set construction based on the first original data set to obtain a first data set; training a reaction site prediction model based on the first data set; after the training is finished, inputting the 3D molecular structure of any compound molecule input by the client into the reaction site prediction model to carry out atomic-scale electrophilic / nucleophilic reaction site prediction processing so as to obtain a corresponding current reaction site prediction tensor; and according to the current 3D molecular structure and the current reaction site prediction tensor, carrying out electrophilic / nucleophilic reaction site atomic information summarization to obtain corresponding summarized information, and feeding back the summarized information to the current client. Based on the reaction site prediction model provided by the invention, the prediction efficiency can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a processing method and device for atomic-level reaction site prediction. Background Art

[0002] Fukui function is an important concept in density functional theory (DFT). By calculating and analyzing the Fukui function, the electrophilic / nucleophilic reaction sites of compound molecules can be predicted at the atomic level. At present, the traditional prediction method based on the Fukui function consists of three steps: 1) The first step is to use the Fukui function calculation interface provided by quantum chemical calculation tools (such as Gaussian software, ORCA software, CP2K software, etc.) to perform atomic-level Fukui function calculations based on the 3D molecular structure of the compound molecule to obtain the nucleophilic and electrophilic reaction function values ​​of each atom. 2) In the second step, the three types of reaction sites (electrophilic reaction sites, nucleophilic reaction sites, and non-nucleophilic reaction sites) of each atom on the compound molecule are estimated using quantum chemical calculation tools based on the three-dimensional spatial configuration of the 3D molecular structure and the calculation results of the Fukui function; 3) In the third step, the correctness of the reaction site estimation results output by the quantum chemical calculation tool is verified by chemical experimental means, and the estimation results of the tool are output as the final prediction results when the verification passes. Through practice, it can be found that the second and third steps of the above three processing steps are relatively time-consuming and the corresponding calculation and experimental costs are relatively high. In other words, if the traditional prediction method is used, there will inevitably be problems such as long prediction cycle, high prediction cost, and low prediction efficiency. Summary of the invention

[0003] The purpose of the present invention is to provide a processing method, device, electronic device and computer-readable storage medium for atomic-level reaction site prediction in view of the defects of the prior art. The present invention first constructs a processing model for atomic-level electrophilic / nucleophilic reaction site prediction of 3D molecular structures, which is recorded as the corresponding reaction site prediction model; then a first original data set containing a large number of 3D molecular structures is constructed by big data acquisition, and each 3D molecular structure of the first original data set is predicted at the atomic level of electrophilic / nucleophilic reaction sites according to the traditional prediction method, and label data is constructed based on the prediction results to obtain a model training data set, i.e., the first data set, and the reaction site prediction model is trained based on the first data set; after the model training is completed, the end-to-end atomic-level electrophilic / nucleophilic reaction site prediction is performed based on the 3D molecular structure of any compound based on the reaction site prediction model. The reaction site prediction model provided by the present invention integrates the first step processing process in the traditional prediction method, but omits the second and third steps of the processing process; based on the reaction site prediction model provided by the present invention, the prediction cycle can be shortened, the prediction cost can be reduced, and the prediction efficiency can be improved.

[0004] To achieve the above-mentioned purpose, a first aspect of an embodiment of the present invention provides a method for atomic-level reaction site prediction, the method comprising:

[0005] Construct a processing model for predicting electrophilic / nucleophilic reaction sites at the atomic level for 3D molecular structures, denoted as the corresponding reaction site prediction model;

[0006] A first original data set is constructed by collecting big data of the 3D molecular structure; and a model training data set is constructed based on the first original data set and preset quantum chemical calculation tools and experimental means to obtain a corresponding first data set; and the reaction site prediction model is trained based on the first data set;

[0007] After the model training is completed, the 3D molecular structure of any compound molecule input by the client is received as the corresponding current 3D molecular structure; the current 3D molecular structure is input into the reaction site prediction model for atomic-level electrophilic / nucleophilic reaction site prediction processing to obtain the corresponding current reaction site prediction tensor; and the electrophilic / nucleophilic reaction site atomic information is summarized according to the current 3D molecular structure and the current reaction site prediction tensor to obtain the corresponding summary information and feedback it to the current client.

[0008] Preferably, the quantum chemical calculation tools include at least Gaussian software, ORCA software and CP2K software.

[0009] Preferably, the reaction site prediction model is used to perform atomic-level electrophilic / nucleophilic reaction site prediction processing according to the 3D molecular structure input by the model and output a corresponding reaction site prediction tensor;

[0010] The model input end of the reaction site prediction model is used to receive the pair of 3D molecular structures input by the model, and the model output end is used to output the corresponding reaction site prediction tensor;

[0011] The 3D molecular structure consists of a first set of atoms and a first set of chemical bonds;

[0012] The first atom set includes multiple first atoms; the atomic properties of the first atoms include at least a first atom identifier, a first atom type, a first atom coordinates, a first atomic charge, and a first atom chiral property set; the first atom identifier is a unique identifier of the current first atom; the first atom type is a chemical element type of the current first atom; the first atomic charge is an atomic charge number of the current first atom; the first atom chiral property set consists of multiple first chiral properties; each of the first chiral properties consists of a corresponding property type and a property value; if the first atom is a chiral center atom, all property values ​​in the corresponding first atom chiral property set are not empty; if the first atom is not a chiral center atom, all property values ​​in the corresponding first atom chiral property set are empty;

[0013] The first chemical bond set includes multiple first chemical bonds; the chemical bond attributes of the first chemical bond include at least a first chemical bond identifier, a first chemical bond type and a first atom identifier group; the first chemical bond identifier is a unique identifier of the current first chemical bond; the first chemical bond type includes multiple chemical bond types; the first atom identifier group consists of the first atom identifiers of the two first atoms currently connected by the first chemical bond;

[0014] The reaction site prediction tensor includes a plurality of reaction site prediction vectors; the reaction site prediction vectors correspond to the first atoms one by one; the reaction site prediction vectors are composed of three prediction probabilities whose sum is 1, namely: electrophilic reaction site prediction probability, nucleophilic reaction site prediction probability, and non-reaction site prediction probability;

[0015] The reaction site prediction model includes a Fukui function calculation module, a molecular graph construction module, a molecular graph feature extraction network, a fully connected network and a Softmax activation layer;

[0016] The input end of the Fukui function calculation module is connected to the input end of the model, and the output end is connected to the second input end of the molecular graph construction module; the first input end of the molecular graph construction module is connected to the input end of the model, and the output end is connected to the input end of the molecular graph feature extraction network; the output end of the molecular graph feature extraction network is connected to the input end of the fully connected network; the output end of the fully connected network is connected to the Softmax activation layer; the Softmax activation layer is connected to the output end of the model;

[0017] The Fukui function calculation module is implemented based on the Fukui function calculation interface provided by the quantum chemical calculation tool; the Fukui function calculation module is used to perform atomic-level Fukui function calculation according to the 3D molecular structure input at the model input end to obtain the corresponding first Fukui function value tensor and send it to the molecular graph construction module; the first Fukui function value tensor is composed of multiple first Fukui function value vectors; the first Fukui function value vector corresponds to the first atom of the 3D molecular structure one by one; the first Fukui function value vector includes a nucleophilic reaction function value and electrophilic reaction function values

[0018] The molecular graph construction module is used to construct a corresponding molecular graph node set according to the first atom set of the 3D molecular structure input by the model input end and the first Fukui function value tensor to obtain a corresponding first node set; and to construct a corresponding molecular graph node edge set according to the first chemical bond set of the 3D molecular structure input by the model input end to obtain a corresponding first node edge set; and to form a corresponding first molecular graph composed of the first node set and the first node edge set to send to the molecular graph feature extraction network; the first molecular graph includes the first node set and the first node edge set; the first node set includes a plurality of first nodes; the first node corresponds to the first atom one by one; the node attribute of the first node is composed of the corresponding first atom identifier, the first atom type, the first atom coordinate, the first atom charge, the first atom chirality attribute set and the first Fukui function value vector; the first node edge set includes a plurality of first node edges; the edge attribute of the first node edge is composed of the corresponding first chemical bond identifier, the first chemical bond type and the first atom identifier group;

[0019] The molecular graph feature extraction network is realized by sequentially connecting a plurality of GINEConv layers; the molecular graph feature extraction network is used to perform node feature extraction processing with an edge feature fusion mechanism according to the first molecular graph to obtain a corresponding first node feature tensor and send it to the fully connected network; the first node feature tensor includes a plurality of first node feature vectors; the first node feature vector corresponds to the first node one by one;

[0020] The fully connected network is realized by sequentially connecting one or more fully connected layers; the fully connected network is used to perform a linear transformation process from a feature tensor to a prediction tensor on the first node feature tensor to obtain a corresponding first node prediction tensor and send it to the Softmax activation layer; the first node prediction tensor includes a plurality of first node prediction vectors; the first node prediction vector corresponds to the first node feature vector one by one; the first node prediction vector consists of three types of prediction values, namely, an electrophilic reaction site prediction value, a nucleophilic reaction site prediction value, and a non-reaction site prediction value;

[0021] The Softmax activation layer is composed of a plurality of parallel Softmax activation functions; the Softmax activation functions correspond to the first node prediction vectors one by one; the Softmax activation layer is used to obtain the corresponding reaction site prediction vector by performing prediction probability normalization processing according to the corresponding first node prediction vector by each of the Softmax activation functions; and the corresponding reaction site prediction tensor is formed by all the obtained reaction site prediction vectors; and the reaction site prediction tensor is sent to the model output end;

[0022] The Softmax activation layer is specifically used for, when each of the Softmax activation functions performs prediction probability normalization processing according to the corresponding first node prediction vector to obtain the corresponding reaction site prediction vector, setting the electrophilic reaction site prediction value, the nucleophilic reaction site prediction value, and the non-reaction site prediction value of each of the first node prediction vectors to be the corresponding z1, z2, and z3, respectively, and setting the electrophilic reaction site prediction probability, the nucleophilic reaction site prediction probability, and the non-reaction site prediction probability of the reaction site prediction vector corresponding to each of the first node prediction vectors to be the corresponding p1, p2, and p3, respectively, and setting the calculation formulas of p1, p2, p3 and z1, z2, and z3 to:

[0023] Preferably, the first original data set includes a plurality of first original records; each of the first original records includes a first original molecular formula and a first original molecular structure; each of the first original records corresponds to a compound molecule; the first original molecular formula is the chemical molecular formula of the current compound molecule; the first original molecular structure is the 3D molecular structure of the current compound molecule;

[0024] The first data set includes multiple first data records; the first data record includes a first training structure and a first label tensor; the first training structure is a 3D molecular structure; the first label tensor includes multiple first label vectors; the first label vector is composed of three label probabilities whose sum is 1, namely: electrophilic reaction site label probability, nucleophilic reaction site label probability, non-reaction site label probability; one and only one of the three label probabilities in the first label vector is 1, and the other two are 0; the first label vector corresponds one-to-one to the first atom of the first atom set of the first training structure.

[0025] Preferably, the step of constructing the first original data set by collecting big data of the 3D molecular structure specifically includes:

[0026] Data collection is performed on the 3D molecular structures of various compound molecules through multiple types of public data channels, and the chemical formula and the 3D molecular structure of each collected compound molecule are used as the corresponding first original molecular formula and the first original molecular structure to form the corresponding first original record; and all the obtained first original records form the corresponding first original data set; the multiple types of public data channels include at least various types of public data sets, journals / magazines / scientific and technological documents / papers that have published molecular three-dimensional structures.

[0027] Preferably, the construction of the model training data set based on the first original data set and the preset quantum chemical calculation tools and experimental means to obtain the corresponding first data set specifically includes:

[0028] Each of the first original records of the first original data set is used as the corresponding current original record; and the first original molecular formula and the first original molecular structure of the current original record are used as the corresponding current molecular formula and current molecular structure; and based on the quantum chemical calculation tool, a corresponding three-dimensional molecular model is constructed according to the current molecular formula and the current molecular structure to obtain the corresponding current molecular model; and based on the Fukui function calculation interface provided by the current quantum chemical calculation tool, an atomic-level Fukui function calculation is performed according to the current molecular model and the calculation result is used as the corresponding current Fukui function value tensor; and according to the current Fukui function value tensor and the three-dimensional spatial configuration of the current molecular model, three types of reaction sites of each atom on the current molecular structure are estimated to obtain the corresponding atomic reaction site type set; and through chemical The correctness of the estimated atomic reaction site type set is verified experimentally; and when the correctness verification passes, the current molecular structure is taken as a corresponding first training structure, and the label tensor conversion is performed according to the atomic reaction site type set corresponding to the current molecular structure to obtain the corresponding first label tensor, and the first training structure corresponding to the current molecular structure and the first label vector form a corresponding first data record; and all the obtained first data records form the corresponding first data set; the atomic reaction site type set includes multiple atomic reaction site types; the atomic reaction site type corresponds one-to-one to the first atom of the first atom set of the corresponding 3D molecular structure; the atomic reaction site types include electrophilic reaction sites, nucleophilic reaction sites and non-reaction sites;

[0029] Among them, the label tensor conversion according to the atomic reaction site type set corresponding to the current molecular structure is performed to obtain the corresponding first label tensor, specifically: each atomic reaction site type in the current atomic reaction site type set is used as the corresponding current reaction site type; and a corresponding first label vector is generated for the current reaction site type as the corresponding current label vector; and the electrophilic reaction site label probability, the nucleophilic reaction site label probability or the non-reaction site label probability corresponding to the current reaction site type in the current label vector is used as the corresponding current label probability, and the current label probability is set to 1, and the other two label probabilities except the current label probability are set to 0; and all the first label vectors corresponding to the current atomic reaction site type set form a corresponding first label tensor.

[0030] Preferably, the training of the reaction site prediction model based on the first data set specifically includes:

[0031] Step 71, based on a preset first segmentation ratio, the first data set is divided into two sub-data sets which are recorded as a corresponding first training set and a first evaluation set;

[0032] Wherein, both the first training set and the first evaluation set include a plurality of the first data records; the ratio of the total number of records in the first training set to the first evaluation set satisfies the first segmentation ratio;

[0033] Step 72, extracting the first first data record of the first training set as the corresponding current training record;

[0034] Step 73, inputting the first training structure of the current training record as the 3D molecular structure of the model input into the reaction site prediction model to perform atomic-level electrophilic / nucleophilic reaction site prediction processing, and using the reaction site prediction tensor output by the model as the corresponding first prediction tensor; and bringing the first prediction tensor and the first label tensor of the current training record into a preset first model loss function to calculate and obtain a corresponding first loss value;

[0035] Wherein, the first model loss function at least includes a cross entropy loss function;

[0036] Step 74, identifying whether the first loss value satisfies a preset first loss value range; if the first loss value satisfies the first loss value range, identifying whether the current training record is the last first data record of the first training set, if so, going to step 75, if not, extracting the next first data record of the first training set as the new current training record and returning to step 73; if the first loss value does not satisfy the first loss value range, performing a round of modulation on the model parameters of the molecular graph feature extraction network and the fully connected network of the reaction site prediction model based on a preset first model optimizer, and returning to step 73 at the end of this round of modulation;

[0037] Wherein, the first model optimizer includes at least an SGD optimizer and an Adam optimizer;

[0038] Step 75, perform a round of traversal on all the first data records of the first evaluation set; and in this round of traversal, use the first data record currently traversed as the corresponding current evaluation record; and use the first training structure of the current evaluation record as the 3D molecular structure input into the reaction site prediction model to perform atomic-level electrophilic / nucleophilic reaction site prediction processing, and use the reaction site prediction tensor output by the model as the corresponding second prediction tensor; and form a corresponding prediction-label data pair with the second prediction tensor and the first label tensor of the current evaluation record; and at the end of this round of traversal, calculate the accuracy, precision, recall, and F1 score based on all the obtained prediction-label data pairs to obtain the corresponding first accuracy, first precision, first recall, and first F1 score;

[0039] Step 76, identifying the first accuracy, the first precision, the first recall and the first F1 score; if the first accuracy does not meet the preset first accuracy range, or the first accuracy does not meet the preset first accuracy range, or the first recall does not meet the preset first recall range, or the first F1 score does not meet the preset first F1 score range, return to step 72 to continue training; if the first accuracy meets the first accuracy range, and the first accuracy meets the first accuracy range, and the first recall meets the first recall range, and the first F1 score meets the first F1 score range, stop training and confirm that the model training is completed.

[0040] Preferably, the step of aggregating the electrophilic / nucleophilic reaction site atomic information according to the current 3D molecular structure and the current reaction site prediction tensor to obtain corresponding summary information and feeding it back to the current client specifically includes:

[0041] The first atom set and the first chemical bond set corresponding to the current 3D molecular structure are taken as the corresponding current atom set and current chemical bond set; and a round of traversal is performed on all the first atoms of the current atom set; and in this round of traversal, the first atom currently traversed is taken as the corresponding current atom; and a reaction site type attribute is added to the atomic properties of the current atom as the corresponding first atom reaction site type; and the reaction site prediction vector corresponding to the current atom in the current reaction site prediction tensor is taken as the corresponding current prediction vector; and the electrophilic reaction site, nucleophilic reaction site or non-reaction site corresponding to the maximum probability in the current prediction vector is taken as the corresponding current reaction site type; and the first atom reaction site type of the current atom is set based on the current reaction site type; and at the end of this round of traversal, the latest new molecular structure composed of the current atom set and the current chemical bond set is fed back to the current client as the corresponding summary information.

[0042] A second aspect of an embodiment of the present invention provides a device for implementing the processing method for atomic-level reaction site prediction described in the first aspect, the device comprising: a model building module, a model training module and a model application module;

[0043] The model building module is used to build a processing model for predicting electrophilic / nucleophilic reaction sites at the atomic level for 3D molecular structures, which is recorded as a corresponding reaction site prediction model;

[0044] The model training module is used to construct a first original data set by collecting big data of the 3D molecular structure; and to construct a model training data set based on the first original data set and preset quantum chemical calculation tools and experimental means to obtain a corresponding first data set; and to train the reaction site prediction model based on the first data set;

[0045] The model application module is used to receive the 3D molecular structure of any compound molecule input by the client as the corresponding current 3D molecular structure after the model training is completed; and input the current 3D molecular structure into the reaction site prediction model to perform atomic-level electrophilic / nucleophilic reaction site prediction processing to obtain the corresponding current reaction site prediction tensor; and summarize the electrophilic / nucleophilic reaction site atomic information according to the current 3D molecular structure and the current reaction site prediction tensor to obtain the corresponding summary information and feedback it to the current client.

[0046] A third aspect of an embodiment of the present invention provides an electronic device, including: a memory, a processor, and a transceiver;

[0047] The processor is used to be coupled to the memory, read and execute instructions in the memory, so as to implement the method steps described in the first aspect above;

[0048] The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

[0049] A fourth aspect of an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions. When the computer instructions are executed by a computer, the computer executes the instructions of the method described in the first aspect above.

[0050] The embodiment of the present invention provides a processing method, device, electronic device and computer-readable storage medium for atomic-level reaction site prediction. As can be seen from the above content, the embodiment of the present invention first constructs a processing model for atomic-level electrophilic / nucleophilic reaction site prediction of 3D molecular structures, which is recorded as the corresponding reaction site prediction model; then a first original data set containing a large number of 3D molecular structures is constructed by big data acquisition, and each 3D molecular structure of the first original data set is predicted at the atomic level of electrophilic / nucleophilic reaction sites according to the traditional prediction method, and label data is constructed based on the prediction results to obtain a model training data set, i.e., the first data set, and the reaction site prediction model is trained based on the first data set; after the model training is completed, the end-to-end atomic-level electrophilic / nucleophilic reaction site prediction is performed based on the 3D molecular structure of any compound based on the reaction site prediction model. The reaction site prediction model provided by the embodiment of the present invention integrates the first step of the traditional prediction method, but omits the second and third steps; the prediction task processing is performed based on the reaction site prediction model provided by the embodiment of the present invention, which not only shortens the prediction cycle, reduces the prediction cost, but also improves the prediction efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A schematic diagram of a processing method for atomic-level reaction site prediction provided in Example 1 of the present invention;

[0052] Figure 2 This is a module structure diagram of the reaction site prediction model provided in Example 1 of the present invention;

[0053] Figure 3 A module structure diagram of a processing device for atomic-level reaction site prediction provided in Embodiment 2 of the present invention;

[0054] Figure 4 A schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention. DETAILED DESCRIPTION

[0055] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] Embodiment 1 of the present invention provides a method for predicting atomic-level reaction sites. Figure 1 As shown in the schematic diagram of a processing method for atomic-level reaction site prediction provided in Example 1 of the present invention, the method mainly includes the following steps:

[0057] Step 1: construct a processing model for predicting electrophilic / nucleophilic reaction sites at the atomic level for 3D molecular structures, which is recorded as the corresponding reaction site prediction model.

[0058] Here, the 3D molecular structures mentioned in the embodiments of the present invention are all composed of a first atom set and a first chemical bond set; wherein: 1) the first atom set includes multiple first atoms; the atomic properties of the first atom include at least a first atom identifier, a first atom type, a first atom coordinate, a first atom charge, and a first atom chiral property set; the first atom identifier is the unique identifier of the current first atom; the first atom type is the chemical element type of the current first atom; the first atomic charge is the atomic charge number of the current first atom; the first atomic chiral property set is composed of multiple first chiral properties; each first chiral property is composed of a corresponding property type and property value composition; if the first atom is a chiral center atom, all attribute values ​​in the corresponding first atom chiral attribute set are not empty; if the first atom is not a chiral center atom, all attribute values ​​in the corresponding first atom chiral attribute set are empty; 2) the first chemical bond set includes multiple first chemical bonds; the chemical bond attributes of the first chemical bond include at least a first chemical bond identifier, a first chemical bond type and a first atom identifier group; the first chemical bond identifier is the unique identifier of the current first chemical bond; the first chemical bond type includes multiple chemical bond types; the first atom identifier group is composed of the first atom identifiers of the two first atoms connected by the current first chemical bond.

[0059] The reaction site prediction model of an embodiment of the present invention is used to perform atomic-level electrophilic / nucleophilic reaction site prediction processing based on the 3D molecular structure input by the model and output the corresponding reaction site prediction tensor; wherein the reaction site prediction tensor includes multiple reaction site prediction vectors; the reaction site prediction vector corresponds one-to-one to the first atom; the reaction site prediction vector is composed of three prediction probabilities whose sum is 1, namely: electrophilic reaction site prediction probability, nucleophilic reaction site prediction probability, and non-reaction site prediction probability.

[0060] The model input of the reaction site prediction model is used to receive the 3D molecular structure of the model input, and the model output is used to output the corresponding reaction site prediction tensor. The model components of the reaction site prediction model include Fukui function calculation module, molecular graph construction module, molecular graph feature extraction network, fully connected network and Softmax activation layer, such as Figure 2 This is a module structure diagram of the reaction site prediction model provided in Example 1 of the present invention.

[0061] The connection relationship between each model component is as follows: the input end of the Fukui function calculation module is connected to the model input end, and the output end is connected to the second input end of the molecular graph construction module; the first input end of the molecular graph construction module is connected to the model input end, and the output end is connected to the input end of the molecular graph feature extraction network; the output end of the molecular graph feature extraction network is connected to the input end of the fully connected network; the output end of the fully connected network is connected to the Softmax activation layer; the Softmax activation layer is connected to the model output end.

[0062] The component functions of each model component are described below.

[0063] 1) Fukui function calculation module:

[0064] The Fukui function calculation module is implemented based on the Fukui function calculation interface provided by the quantum chemical calculation tool; here, the quantum chemical calculation tool mentioned in the embodiment of the present invention includes at least Gaussian software, ORCA software and CP2K software.

[0065] The Fukui function calculation module is used to perform atomic-level Fukui function calculation according to the 3D molecular structure input by the model input terminal to obtain the corresponding first Fukui function value tensor and send it to the molecular graph construction module; wherein the first Fukui function value tensor is composed of multiple first Fukui function value vectors; the first Fukui function value vector corresponds to the first atom of the 3D molecular structure one by one; the first Fukui function value vector includes a nucleophilic reaction function value and electrophilic reaction function values

[0066] The Fukui function calculation module of the reaction site prediction model in the embodiment of the present invention is actually similar to the first step of the traditional prediction method, but the subsequent molecular graph construction module, molecular graph feature extraction network, fully connected network and Softmax activation layer are used to replace the second and third steps of the traditional prediction method.

[0067] 2) Molecular graph building module:

[0068] The molecular graph construction module is used to construct a corresponding molecular graph node set according to the first atom set of the 3D molecular structure input by the model input end and the first Fukui function value tensor to obtain a corresponding first node set; and to construct a corresponding molecular graph node edge set according to the first chemical bond set of the 3D molecular structure input by the model input end to obtain a corresponding first node edge set; and to send the corresponding first molecular graph composed of the first node set and the first node edge set to the molecular graph feature extraction network.

[0069] Here, the first molecular graph of an embodiment of the present invention includes a first node set and a first node edge set; wherein, the first node set includes multiple first nodes; the first node corresponds one-to-one to the first atom; the node attributes of the first node are composed of the corresponding first atom identifier, the first atom type, the first atom coordinates, the first atom charge, the first atom chirality attribute set and the first Fukui function value vector; the first node edge set includes multiple first node edges; the edge attributes of the first node edge are composed of the corresponding first chemical bond identifier, the first chemical bond type and the first atom identifier group.

[0070] 3) Molecular graph feature extraction network:

[0071] The molecular graph feature extraction network is implemented by sequentially connecting multiple Graph Isomorphism Network with Edge Features Convolution (GINEConv) layers.

[0072] The molecular graph feature extraction network is used to perform node feature extraction processing with an edge feature fusion mechanism according to the first molecular graph to obtain the corresponding first node feature tensor and send it to the fully connected network; wherein the first node feature tensor includes multiple first node feature vectors; the first node feature vector corresponds one-to-one to the first node.

[0073] 4) Fully connected network:

[0074] A fully connected network is realized by sequentially connecting one or more fully connected layers.

[0075] The fully connected network is used to perform linear transformation processing on the first node feature tensor from the feature tensor to the prediction tensor to obtain the corresponding first node prediction tensor and send it to the Softmax activation layer; wherein the first node prediction tensor includes multiple first node prediction vectors; the first node prediction vector corresponds one-to-one to the first node feature vector; the first node prediction vector consists of three types of prediction values, namely, electrophilic reaction site prediction value, nucleophilic reaction site prediction value, and non-reaction site prediction value.

[0076] 5) Softmax activation layer:

[0077] The Softmax activation layer consists of multiple parallel Softmax activation functions; the Softmax activation function corresponds one-to-one to the first node prediction vector.

[0078] The Softmax activation layer is used to obtain the corresponding reaction site prediction vector by performing prediction probability normalization processing according to the corresponding first node prediction vector by each Softmax activation function; and the corresponding reaction site prediction tensor is composed of all the obtained reaction site prediction vectors; and the reaction site prediction tensor is sent to the output end of the model.

[0079] Here, the Softmax activation layer is specifically used for obtaining the corresponding reaction site prediction vector by normalizing the prediction probability of each Softmax activation function according to the corresponding first node prediction vector: the electrophilic reaction site prediction value, the nucleophilic reaction site prediction value, and the non-reaction site prediction value of each first node prediction vector are set to the corresponding z1, z2, and z3 respectively; and the electrophilic reaction site prediction probability, the nucleophilic reaction site prediction probability, and the non-reaction site prediction probability of the reaction site prediction vector corresponding to each first node prediction vector are set to the corresponding p1, p2, and p3 respectively; and the calculation formulas of p1, p2, p3 and z1, z2, and z3 are set to be:

[0080] Step 2, constructing a first original data set by collecting big data of 3D molecular structures; and constructing a model training data set based on the first original data set and preset quantum chemical calculation tools and experimental means to obtain a corresponding first data set; and training a reaction site prediction model based on the first data set;

[0081] Specifically comprising: step 21, constructing a first original data set by collecting big data of 3D molecular structures;

[0082] Specifically, it includes: collecting data on the 3D molecular structures of various compound molecules through multiple public data channels, and using the chemical molecular formula and 3D molecular structure of each compound molecule collected as the corresponding first original molecular formula and the first original molecular structure to form a corresponding first original record; and forming a corresponding first original data set from all the obtained first original records;

[0083] Among them, the multiple public data channels include at least various public data sets, journals / magazines / scientific and technological documents / papers that publish molecular three-dimensional structures; the first original data set includes multiple first original records; each first original record includes a first original molecular formula and a first original molecular structure; each first original record corresponds to a compound molecule; the first original molecular formula is the chemical molecular formula of the current compound molecule; the first original molecular structure is the 3D molecular structure of the current compound molecule;

[0084] Step 22, constructing a model training data set based on the first original data set and the preset quantum chemical calculation tools and experimental means to obtain a corresponding first data set;

[0085] The first data set includes a plurality of first data records; the first data record includes a first training structure and a first label tensor; the first training structure is a 3D molecular structure; the first label tensor includes a plurality of first label vectors; the first label vector is composed of three label probabilities whose sum is 1, namely: electrophilic reaction site label probability, nucleophilic reaction site label probability, and non-reaction site label probability; one and only one of the three label probabilities in the first label vector is 1, and the other two are 0; the first label vector corresponds one-to-one to the first atom of the first atom set of the first training structure;

[0086] Specifically, it includes: taking each first original record of the first original data set as the corresponding current original record; taking the first original molecular formula and the first original molecular structure of the current original record as the corresponding current molecular formula and the current molecular structure; and constructing a corresponding three-dimensional molecular model according to the current molecular formula and the current molecular structure based on the quantum chemical calculation tool to obtain the corresponding current molecular model; and performing atomic-level Fukui function calculation according to the current molecular model based on the Fukui function calculation interface provided by the current quantum chemical calculation tool and taking the calculation result as the corresponding current Fukui function value tensor; and estimating the three types of reaction site types of each atom on the current molecular structure according to the current Fukui function value tensor and the three-dimensional spatial configuration of the current molecular model to obtain a corresponding atomic reaction site type set; and verifying the correctness of the estimated atomic reaction site type set by chemical experiments; and when the correctness verification passes, taking the current molecular structure as a corresponding first training structure, and performing label tensor conversion according to the atomic reaction site type set corresponding to the current molecular structure to obtain a corresponding first label tensor, and forming a corresponding first data record by the first training structure and the first label vector corresponding to the current molecular structure; and forming a corresponding first data set by all the obtained first data records;

[0087] The atomic reaction site type set includes multiple atomic reaction site types; the atomic reaction site type corresponds one-to-one to the first atom of the first atomic set of the corresponding 3D molecular structure; the atomic reaction site type includes an electrophilic reaction site, a nucleophilic reaction site and a non-reaction site;

[0088] In the current step 22, the label tensor is converted according to the atomic reaction site type set corresponding to the current molecular structure to obtain the corresponding first label tensor, specifically: each atomic reaction site type of the current atomic reaction site type set is used as the corresponding current reaction site type; and a corresponding first label vector is generated for the current reaction site type as the corresponding current label vector; and the electrophilic reaction site label probability, nucleophilic reaction site label probability or non-reaction site label probability corresponding to the current reaction site type in the current label vector is used as the corresponding current label probability, and the current label probability is set to 1, and the remaining two label probabilities except the current label probability are set to 0; and all the first label vectors corresponding to the current atomic reaction site type set form a corresponding first label tensor;

[0089] Step 23, training a reaction site prediction model based on the first data set;

[0090] Specifically, it includes: step 231, dividing the first data set into two sub-data sets based on a preset first division ratio and recording them as a corresponding first training set and a first evaluation set;

[0091] The first segmentation ratio is a preset ratio parameter, such as 8:2; the first training set and the first evaluation set both include a plurality of first data records; the ratio of the total number of records in the first training set to the total number of records in the first evaluation set satisfies the first segmentation ratio;

[0092] Step 232, extracting the first first data record of the first training set as the corresponding current training record;

[0093] Step 233, input the first training structure of the current training record as the 3D molecular structure of the model input into the reaction site prediction model to perform atomic-level electrophilic / nucleophilic reaction site prediction processing, and use the reaction site prediction tensor output by the model as the corresponding first prediction tensor; and bring the first prediction tensor and the first label tensor of the current training record into the preset first model loss function to calculate and obtain the corresponding first loss value;

[0094] Wherein, the first model loss function at least includes a cross entropy loss function;

[0095] Step 234, identifying whether the first loss value satisfies a preset first loss value range; if the first loss value satisfies the first loss value range, identifying whether the current training record is the last first data record of the first training set, if so, turning to step 235, if not, extracting the next first data record of the first training set as a new current training record and returning to step 233; if the first loss value does not satisfy the first loss value range, performing a round of modulation on the model parameters of the molecular graph feature extraction network and the fully connected network of the reaction site prediction model based on the preset first model optimizer, and returning to step 233 at the end of this round of modulation;

[0096] Wherein, the first loss value range is a preset numerical range; the first model optimizer includes at least an SGD optimizer and an Adam optimizer;

[0097] Step 235, perform a round of traversal on all first data records of the first evaluation set; and in this round of traversal, use the first data record currently traversed as the corresponding current evaluation record; and use the first training structure of the current evaluation record as the 3D molecular structure input of the model to input the reaction site prediction model to perform atomic-level electrophilic / nucleophilic reaction site prediction processing, and use the reaction site prediction tensor output by the model as the corresponding second prediction tensor; and form a corresponding prediction-label data pair by the second prediction tensor and the first label tensor of the current evaluation record; and at the end of this round of traversal, calculate the accuracy, precision, recall, and F1 score based on all the obtained prediction-label data pairs to obtain the corresponding first accuracy, first precision, first recall, and first F1 score;

[0098] Step 236, identify the first accuracy, the first precision, the first recall rate and the first F1 score; if the first accuracy does not meet the preset first accuracy range, or the first accuracy does not meet the preset first accuracy range, or the first recall rate does not meet the preset first recall rate range, or the first F1 score does not meet the preset first F1 score range, return to step 232 to continue training; if the first accuracy meets the first accuracy range, and the first accuracy meets the first accuracy range, and the first recall rate meets the first recall rate range, and the first F1 score meets the first F1 score range, stop training and confirm that the model training is completed.

[0099] Here, the first accuracy range, the first precision range, the first recall range and the first F1 score range are four preset value ranges.

[0100] Step 3: After the model training is completed, the 3D molecular structure of any compound molecule input by the client is received as the corresponding current 3D molecular structure; the current 3D molecular structure is input into the reaction site prediction model to perform atomic-level electrophilic / nucleophilic reaction site prediction processing to obtain the corresponding current reaction site prediction tensor; and the electrophilic / nucleophilic reaction site atomic information is summarized according to the current 3D molecular structure and the current reaction site prediction tensor to obtain the corresponding summary information and feedback it to the current client;

[0101] Specifically, it includes: step 31, after the model training is completed, receiving the 3D molecular structure of any compound molecule input by the client as the corresponding current 3D molecular structure; and inputting the current 3D molecular structure into the reaction site prediction model to perform atomic-level electrophilic / nucleophilic reaction site prediction processing to obtain the corresponding current reaction site prediction tensor;

[0102] Step 32, summarizing the electrophilic / nucleophilic reaction site atomic information according to the current 3D molecular structure and the current reaction site prediction tensor to obtain corresponding summary information and feed it back to the current client;

[0103] Specifically, it includes: taking the first atom set and the first chemical bond set corresponding to the current 3D molecular structure as the corresponding current atom set and the current chemical bond set; and performing a round of traversal on all the first atoms of the current atom set; and in this round of traversal, taking the first atom currently traversed as the corresponding current atom; and adding a reaction site type attribute to the atomic properties of the current atom as the corresponding first atom reaction site type; and taking the reaction site prediction vector corresponding to the current atom in the current reaction site prediction tensor as the corresponding current prediction vector; and taking the electrophilic reaction site, nucleophilic reaction site or non-reaction site corresponding to the maximum probability in the current prediction vector as the corresponding current reaction site type; and setting the first atom reaction site type of the current atom based on the current reaction site type; and at the end of this round of traversal, feeding back the new molecular structure composed of the latest current atom set and current chemical bond set as the corresponding summary information to the current client.

[0104] Figure 3 This is a module structure diagram of a processing device for atomic-level reaction site prediction provided in Embodiment 2 of the present invention. The device is a terminal device or server that implements the aforementioned method embodiment, or can be a device that enables the aforementioned terminal device or server to implement the aforementioned method embodiment. For example, the device can be a device or chip system of the aforementioned terminal device or server. Figure 3 As shown, the device includes: a model building module 201, a model training module 202 and a model application module 203.

[0105] The model building module 201 is used to build a processing model for predicting electrophilic / nucleophilic reaction sites at the atomic level for a 3D molecular structure, which is recorded as a corresponding reaction site prediction model.

[0106] The model training module 202 is used to construct a first original data set by collecting big data of 3D molecular structures; and to construct a model training data set based on the first original data set and preset quantum chemical calculation tools and experimental means to obtain a corresponding first data set; and to train a reaction site prediction model based on the first data set.

[0107] The model application module 203 is used to receive the 3D molecular structure of any compound molecule input by the client as the corresponding current 3D molecular structure after the model training is completed; and input the current 3D molecular structure into the reaction site prediction model to perform atomic-level electrophilic / nucleophilic reaction site prediction processing to obtain the corresponding current reaction site prediction tensor; and summarize the electrophilic / nucleophilic reaction site atomic information based on the current 3D molecular structure and the current reaction site prediction tensor to obtain the corresponding summary information and feedback it to the current client.

[0108] An atomic-level reaction site prediction processing device provided in an embodiment of the present invention can execute the method steps in the above method embodiment. Its implementation principle and technical effects are similar and will not be repeated here.

[0109] It should be noted that it should be understood that the division of the various modules of the above device is only a division of logical functions. In actual implementation, they can be fully or partially integrated into one physical entity, or they can be physically separated. And these modules can all be implemented in the form of software called by processing elements; they can also be all implemented in the form of hardware; some modules can also be implemented in the form of software called by processing elements, and some modules can be implemented in the form of hardware. For example, the model building module can be a separately established processing element, or it can be integrated in a chip of the above device. In addition, it can also be stored in the memory of the above device in the form of program code, and called and executed by a processing element of the above device. The function of the above-mentioned module is determined. The implementation of other modules is similar. In addition, these modules can be fully or partially integrated together, or they can be implemented independently. The processing element described here can be an integrated circuit with signal processing capabilities. In the implementation process, each step of the above method or each module above can be completed by an integrated logic circuit of hardware in the processor element or instructions in the form of software.

[0110] For example, 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 (ASIC), or one or more digital signal processors (DSP), or one or more field programmable gate arrays (FPGA). For another example, when a module above 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).

[0111] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the above method embodiments are generated. The above computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The above-mentioned computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the above-mentioned computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, Bluetooth, microwave, etc.) methods. The above-mentioned computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that includes one or more available media integrated. The above-mentioned available medium can be a magnetic medium (such as a floppy disk, a hard disk, a tape), an optical medium (such as a DVD), or a semiconductor medium (such as a solid state disk (SSD)), etc.

[0112] Figure 4 This is a schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention. The electronic device may be a terminal device or a server for implementing the method of the aforementioned embodiment, or may be a terminal device or a server for implementing the method of the aforementioned embodiment connected to the aforementioned terminal device or server. Figure 4As shown, the electronic device may include: a processor 301 (such as a CPU), a memory 302, and a transceiver 303; the transceiver 303 is coupled to the processor 301, and the processor 301 controls the transceiver 303. Various instructions may be stored in the memory 302 to complete various processing functions and implement the processing steps described in the aforementioned embodiment method. Preferably, the electronic device involved in the embodiment of the present invention also includes: a power supply 304, a system bus 305 and a communication port 306. The system bus 305 is used to realize the communication connection between components. The above-mentioned communication port 306 is used for connecting and communicating between the electronic device and other peripherals.

[0113] exist Figure 4 The system bus 305 mentioned in the figure can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The system bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus. The communication interface is used to realize the communication between the database access device and other devices (such as clients, read-write libraries, and read-only libraries). The memory may include random access memory (RAM), and may also include non-volatile memory (Non-Volatile Memory), such as at least one disk storage.

[0114] The above-mentioned processor can be a general-purpose processor, including a central processing unit CPU, a network processor (Network Processor, NP), a graphics processing unit (Graphics Processing Unit, GPU), etc.; it can also be a digital signal processor DSP, an application-specific integrated circuit ASIC, a field programmable gate array FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0115] It should be noted that an embodiment of the present invention further provides a computer-readable storage medium, in which instructions are stored. When the computer-readable storage medium is run on a computer, the computer executes the method and processing process provided in the above embodiments.

[0116] The embodiment of the present invention provides a processing method, device, electronic device and computer-readable storage medium for atomic-level reaction site prediction. As can be seen from the above content, the embodiment of the present invention first constructs a processing model for atomic-level electrophilic / nucleophilic reaction site prediction of 3D molecular structures, which is recorded as the corresponding reaction site prediction model; then a first original data set containing a large number of 3D molecular structures is constructed by big data acquisition, and each 3D molecular structure of the first original data set is predicted at the atomic level of electrophilic / nucleophilic reaction sites according to the traditional prediction method, and label data is constructed based on the prediction results to obtain a model training data set, i.e., the first data set, and the reaction site prediction model is trained based on the first data set; after the model training is completed, the end-to-end atomic-level electrophilic / nucleophilic reaction site prediction is performed based on the 3D molecular structure of any compound based on the reaction site prediction model. The reaction site prediction model provided by the embodiment of the present invention integrates the first step of the traditional prediction method, but omits the second and third steps; the prediction task processing is performed based on the reaction site prediction model provided by the embodiment of the present invention, which not only shortens the prediction cycle, reduces the prediction cost, but also improves the prediction efficiency.

[0117] The steps of the method or algorithm described in conjunction with the embodiments disclosed herein may be implemented using hardware, a software module executed by a processor, or a combination of the two. The software module may be placed in a random access memory (RAM), a memory, a read-only memory (ROM), an electrically programmable ROM, an electrically erasable programmable ROM, a register, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.

[0118] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for atomic-level reaction site prediction, characterized in that: The method comprises: Construct a processing model for predicting electrophilic / nucleophilic reaction sites at the atomic level for 3D molecular structures, denoted as the corresponding reaction site prediction model; A first original data set is constructed by collecting big data of the 3D molecular structure; and a model training data set is constructed based on the first original data set and preset quantum chemical calculation tools and experimental means to obtain a corresponding first data set; and the reaction site prediction model is trained based on the first data set; After the model training is completed, the 3D molecular structure of any compound molecule input by the client is received as the corresponding current 3D molecular structure; the current 3D molecular structure is input into the reaction site prediction model for atomic-level electrophilic / nucleophilic reaction site prediction processing to obtain the corresponding current reaction site prediction tensor; and the electrophilic / nucleophilic reaction site atomic information is summarized according to the current 3D molecular structure and the current reaction site prediction tensor to obtain the corresponding summary information and feedback it to the current client.

2. The method for atomic-level reaction site prediction according to claim 1, characterized in that: The quantum chemical calculation tools at least include Gaussian software, ORCA software and CP2K software.

3. The method for atomic-level reaction site prediction according to claim 1, characterized in that: The reaction site prediction model is used to perform atomic-level electrophilic / nucleophilic reaction site prediction processing according to the 3D molecular structure input by the model and output a corresponding reaction site prediction tensor; The model input end of the reaction site prediction model is used to receive the pair of 3D molecular structures input by the model, and the model output end is used to output the corresponding reaction site prediction tensor; The 3D molecular structure consists of a first set of atoms and a first set of chemical bonds; The first atom set includes a plurality of first atoms; the atomic properties of the first atoms include at least a first atom identifier, a first atom type, a first atom coordinates, a first atom charge, and a first atom chiral property set; the first atom identifier is a unique identifier of the current first atom; the first atom type is a chemical element type of the current first atom; The first atomic charge is the current atomic charge number of the first atom; The first atomic chiral attribute set consists of a plurality of first chiral attributes; each of the first chiral attributes consists of a corresponding attribute type and an attribute value; if the first atom is a chiral center atom, all attribute values ​​in the corresponding first atomic chiral attribute set are not empty; If the first atom is not a chiral center atom, all attribute values ​​in the corresponding first atom chiral attribute set are empty; The first chemical bond set includes multiple first chemical bonds; the chemical bond attributes of the first chemical bond include at least a first chemical bond identifier, a first chemical bond type and a first atom identifier group; the first chemical bond identifier is a unique identifier of the current first chemical bond; the first chemical bond type includes multiple chemical bond types; the first atom identifier group consists of the first atom identifiers of the two first atoms currently connected by the first chemical bond; The reaction site prediction tensor includes a plurality of reaction site prediction vectors; the reaction site prediction vectors correspond to the first atoms one by one; the reaction site prediction vectors are composed of three prediction probabilities whose sum is 1, namely: electrophilic reaction site prediction probability, nucleophilic reaction site prediction probability, and non-reaction site prediction probability; The reaction site prediction model includes a Fukui function calculation module, a molecular graph construction module, a molecular graph feature extraction network, a fully connected network and a Softmax activation layer; The input end of the Fukui function calculation module is connected to the input end of the model, and the output end is connected to the second input end of the molecular graph construction module; the first input end of the molecular graph construction module is connected to the input end of the model, and the output end is connected to the input end of the molecular graph feature extraction network; the output end of the molecular graph feature extraction network is connected to the input end of the fully connected network; the output end of the fully connected network is connected to the Softmax activation layer; the Softmax activation layer is connected to the output end of the model; The Fukui function calculation module is implemented based on the Fukui function calculation interface provided by the quantum chemical calculation tool; The Fukui function calculation module is used to perform atomic-level Fukui function calculation according to the 3D molecular structure input at the model input end to obtain a corresponding first Fukui function value tensor and send it to the molecular graph construction module; the first Fukui function value tensor is composed of multiple first Fukui function value vectors; the first Fukui function value vector corresponds to the first atom of the 3D molecular structure one by one; the first Fukui function value vector includes a nucleophilic reaction function value f + A and electrophilic reaction function values The molecular graph construction module is used to construct a corresponding molecular graph node set according to the first atom set of the 3D molecular structure input by the model input end and the first Fukui function value tensor to obtain a corresponding first node set; And according to the first chemical bond set of the 3D molecular structure input at the model input end, a corresponding molecular graph node edge set is constructed to obtain a corresponding first node edge set; and the first node set and the first node edge set form a corresponding first molecular graph and send it to the molecular graph feature extraction network; the first molecular graph includes the first node set and the first node edge set; the first node set includes a plurality of first nodes; the first node corresponds to the first atom one by one; the node attribute of the first node is composed of the corresponding first atom identifier, the first atom type, the first atom coordinates, the first atom charge, the first atom chirality attribute set and the first Fukui function value vector; the first node edge set includes a plurality of first node edges; the edge attribute of the first node edge is composed of the corresponding first chemical bond identifier, the first chemical bond type and the first atom identifier group; The molecular graph feature extraction network is realized by sequentially connecting a plurality of GINEConv layers; the molecular graph feature extraction network is used to perform node feature extraction processing with an edge feature fusion mechanism according to the first molecular graph to obtain a corresponding first node feature tensor and send it to the fully connected network; the first node feature tensor includes a plurality of first node feature vectors; the first node feature vector corresponds to the first node one by one; The fully connected network is realized by sequentially connecting one or more fully connected layers; the fully connected network is used to perform a linear transformation process from a feature tensor to a prediction tensor on the first node feature tensor to obtain a corresponding first node prediction tensor and send it to the Softmax activation layer; the first node prediction tensor includes a plurality of first node prediction vectors; the first node prediction vector corresponds to the first node feature vector one by one; the first node prediction vector consists of three types of prediction values, namely, an electrophilic reaction site prediction value, a nucleophilic reaction site prediction value, and a non-reaction site prediction value; The Softmax activation layer is composed of a plurality of parallel Softmax activation functions; the Softmax activation functions correspond to the first node prediction vectors one by one; the Softmax activation layer is used to obtain the corresponding reaction site prediction vector by performing prediction probability normalization processing according to the corresponding first node prediction vector by each of the Softmax activation functions; and the corresponding reaction site prediction tensor is formed by all the obtained reaction site prediction vectors; and the reaction site prediction tensor is sent to the model output end; The Softmax activation layer is specifically used for, when each of the Softmax activation functions performs prediction probability normalization processing according to the corresponding first node prediction vector to obtain the corresponding reaction site prediction vector, setting the electrophilic reaction site prediction value, the nucleophilic reaction site prediction value, and the non-reaction site prediction value of each of the first node prediction vectors to be the corresponding z1, z2, and z3, respectively, and setting the electrophilic reaction site prediction probability, the nucleophilic reaction site prediction probability, and the non-reaction site prediction probability of the reaction site prediction vector corresponding to each of the first node prediction vectors to be the corresponding p1, p2, and p3, respectively, and setting the calculation formulas of p1, p2, p3 and z1, z2, and z3 to:

4. The method for atomic-level reaction site prediction according to claim 3, characterized in that: The first original data set includes a plurality of first original records; each of the first original records includes a first original molecular formula and a first original molecular structure; each of the first original records corresponds to a compound molecule; the first original molecular formula is the chemical molecular formula of the current compound molecule; the first original molecular structure is the 3D molecular structure of the current compound molecule; The first data set includes multiple first data records; the first data record includes a first training structure and a first label tensor; the first training structure is a 3D molecular structure; the first label tensor includes multiple first label vectors; the first label vector is composed of three label probabilities whose sum is 1, namely: electrophilic reaction site label probability, nucleophilic reaction site label probability, non-reaction site label probability; one and only one of the three label probabilities in the first label vector is 1, and the other two are 0; the first label vector corresponds one-to-one to the first atom of the first atom set of the first training structure.

5. The method for atomic-level reaction site prediction according to claim 4, characterized in that: The step of constructing a first original data set by collecting big data of the 3D molecular structure specifically includes: Data collection is performed on the 3D molecular structures of various compound molecules through multiple types of public data channels, and the chemical formula and the 3D molecular structure of each collected compound molecule are used as the corresponding first original molecular formula and the first original molecular structure to form the corresponding first original record; and all the obtained first original records form the corresponding first original data set; the multiple types of public data channels include at least various types of public data sets, journals / magazines / scientific and technological documents / papers that have published molecular three-dimensional structures.

6. The method for atomic-level reaction site prediction according to claim 4, characterized in that: The constructing of a model training data set based on the first original data set and a preset quantum chemical calculation tool and experimental means to obtain a corresponding first data set specifically includes: Each of the first original records of the first original data set is used as the corresponding current original record; and the first original molecular formula and the first original molecular structure of the current original record are used as the corresponding current molecular formula and current molecular structure; and based on the quantum chemical calculation tool, a corresponding three-dimensional molecular model is constructed according to the current molecular formula and the current molecular structure to obtain the corresponding current molecular model; and based on the Fukui function calculation interface provided by the current quantum chemical calculation tool, an atomic-level Fukui function calculation is performed according to the current molecular model and the calculation result is used as the corresponding current Fukui function value tensor; and according to the current Fukui function value tensor and the three-dimensional spatial configuration of the current molecular model, three types of reaction sites of each atom on the current molecular structure are estimated to obtain the corresponding atomic reaction site type set; and through chemical The correctness of the estimated atomic reaction site type set is verified experimentally; and when the correctness verification passes, the current molecular structure is taken as a corresponding first training structure, and the label tensor conversion is performed according to the atomic reaction site type set corresponding to the current molecular structure to obtain the corresponding first label tensor, and the first training structure corresponding to the current molecular structure and the first label vector form a corresponding first data record; and all the obtained first data records form the corresponding first data set; the atomic reaction site type set includes multiple atomic reaction site types; the atomic reaction site type corresponds one-to-one to the first atom of the first atom set of the corresponding 3D molecular structure; the atomic reaction site types include electrophilic reaction sites, nucleophilic reaction sites and non-reaction sites; Among them, the label tensor conversion according to the atomic reaction site type set corresponding to the current molecular structure is performed to obtain the corresponding first label tensor, specifically: each atomic reaction site type in the current atomic reaction site type set is used as the corresponding current reaction site type; and a corresponding first label vector is generated for the current reaction site type as the corresponding current label vector; and the electrophilic reaction site label probability, the nucleophilic reaction site label probability or the non-reaction site label probability corresponding to the current reaction site type in the current label vector is used as the corresponding current label probability, and the current label probability is set to 1, and the other two label probabilities except the current label probability are set to 0; and all the first label vectors corresponding to the current atomic reaction site type set form a corresponding first label tensor.

7. The method for atomic-level reaction site prediction according to claim 4, characterized in that: The training of the reaction site prediction model based on the first data set specifically includes: Step 71, based on a preset first segmentation ratio, the first data set is divided into two sub-data sets which are recorded as a corresponding first training set and a first evaluation set; Wherein, both the first training set and the first evaluation set include a plurality of the first data records; the ratio of the total number of records in the first training set to the first evaluation set satisfies the first segmentation ratio; Step 72, extracting the first first data record of the first training set as the corresponding current training record; Step 73, inputting the first training structure of the current training record as the 3D molecular structure of the model input into the reaction site prediction model to perform atomic-level electrophilic / nucleophilic reaction site prediction processing, and using the reaction site prediction tensor output by the model as the corresponding first prediction tensor; and bringing the first prediction tensor and the first label tensor of the current training record into a preset first model loss function to calculate and obtain a corresponding first loss value; Wherein, the first model loss function at least includes a cross entropy loss function; Step 74, identifying whether the first loss value satisfies a preset first loss value range; if the first loss value satisfies the first loss value range, identifying whether the current training record is the last first data record of the first training set, if so, going to step 75, if not, extracting the next first data record of the first training set as the new current training record and returning to step 73; if the first loss value does not satisfy the first loss value range, performing a round of modulation on the model parameters of the molecular graph feature extraction network and the fully connected network of the reaction site prediction model based on a preset first model optimizer, and returning to step 73 at the end of this round of modulation; Wherein, the first model optimizer includes at least an SGD optimizer and an Adam optimizer; Step 75, perform a round of traversal on all the first data records of the first evaluation set; and in this round of traversal, use the first data record currently traversed as the corresponding current evaluation record; and use the first training structure of the current evaluation record as the 3D molecular structure input into the reaction site prediction model to perform atomic-level electrophilic / nucleophilic reaction site prediction processing, and use the reaction site prediction tensor output by the model as the corresponding second prediction tensor; and form a corresponding prediction-label data pair with the second prediction tensor and the first label tensor of the current evaluation record; and at the end of this round of traversal, calculate the accuracy, precision, recall, and F1 score based on all the obtained prediction-label data pairs to obtain the corresponding first accuracy, first precision, first recall, and first F1 score; Step 76, identifying the first accuracy, the first precision, the first recall and the first F1 score; if the first accuracy does not meet the preset first accuracy range, or the first accuracy does not meet the preset first accuracy range, or the first recall does not meet the preset first recall range, or the first F1 score does not meet the preset first F1 score range, return to step 72 to continue training; if the first accuracy meets the first accuracy range, and the first accuracy meets the first accuracy range, and the first recall meets the first recall range, and the first F1 score meets the first F1 score range, stop training and confirm that the model training is completed.

8. The method for atomic-level reaction site prediction according to claim 3, characterized in that: The step of aggregating the electrophilic / nucleophilic reaction site atomic information according to the current 3D molecular structure and the current reaction site prediction tensor to obtain corresponding summary information and feeding back the corresponding summary information to the current client specifically includes: The first atom set and the first chemical bond set corresponding to the current 3D molecular structure are taken as the corresponding current atom set and current chemical bond set; and a round of traversal is performed on all the first atoms of the current atom set; and in this round of traversal, the first atom currently traversed is taken as the corresponding current atom; and a reaction site type attribute is added to the atomic properties of the current atom as the corresponding first atom reaction site type; and the reaction site prediction vector corresponding to the current atom in the current reaction site prediction tensor is taken as the corresponding current prediction vector; and the electrophilic reaction site, nucleophilic reaction site or non-reaction site corresponding to the maximum probability in the current prediction vector is taken as the corresponding current reaction site type; and the first atom reaction site type of the current atom is set based on the current reaction site type; and at the end of this round of traversal, the latest new molecular structure composed of the current atom set and the current chemical bond set is fed back to the current client as the corresponding summary information.

9. A device for executing the processing method for atomic-level reaction site prediction according to any one of claims 1 to 8, characterized in that: The device comprises: a model building module, a model training module and a model application module; The model building module is used to build a processing model for predicting electrophilic / nucleophilic reaction sites at the atomic level for 3D molecular structures, which is recorded as the corresponding reaction site prediction model; The model training module is used to construct a first original data set by collecting big data of the 3D molecular structure; and to construct a model training data set based on the first original data set and preset quantum chemical calculation tools and experimental means to obtain a corresponding first data set; and to train the reaction site prediction model based on the first data set; The model application module is used to receive the 3D molecular structure of any compound molecule input by the client as the corresponding current 3D molecular structure after the model training is completed; and input the current 3D molecular structure into the reaction site prediction model to perform atomic-level electrophilic / nucleophilic reaction site prediction processing to obtain the corresponding current reaction site prediction tensor; and summarize the electrophilic / nucleophilic reaction site atomic information according to the current 3D molecular structure and the current reaction site prediction tensor to obtain the corresponding summary information and feedback it to the current client.

10. An electronic device, characterized in that: include: memory, processors, and transceivers; The processor is used to couple with the memory, read and execute instructions in the memory, so as to implement the method according to any one of claims 1 to 8; The transceiver is coupled to the processor, and the processor controls the transceiver to send and receive messages.

11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and when the computer instructions are executed by a computer, the computer is enabled to execute the method according to any one of claims 1 to 8.

Images