Method for predicting protein affinity changes and related devices

Abstract

The application relates to the fields of artificial intelligence and digital medical technologies, and provides a protein affinity change prediction method and related equipment, a first initial graph network of a protein before mutation is constructed, a second initial graph network of a protein after mutation is constructed; the first initial graph network is updated in the graph at least once, and the first initial graph network and the second initial graph network are updated between graphs at least once until a first target graph network is obtained; the second initial graph network is updated in the graph at least once, and the second initial graph network and the first initial graph network are updated between graphs at least once until a second target graph network is obtained; according to the first target graph network and the second target graph network, the affinity change between the protein before mutation and the protein after mutation is predicted, so that the prediction accuracy of the protein affinity change is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence and digital medical technology, in particular to a method for predicting protein affinity change and related equipment. BACKGROUND

[0002] The research on protein-protein interaction (PPI) plays a crucial role in many basic biological processes, and affinity is usually used as a thermodynamic indicator to measure the interaction between proteins, thereby measuring the binding degree between two proteins.

[0003] For example, antibodies (Ab) are the core components of the human immune system, and the interaction between antibodies and corresponding antigens can trigger an immune response. The higher the affinity between antibodies and antigens, the better the binding degree, and the better the effect of antibody therapy.

[0004] It is very challenging to directly design antibodies with the desired binding affinity to antigens, and usually requires multiple mutations based on antibody templates to determine whether the affinity between the mutated antibodies and antigens is enhanced, thereby selecting target antibodies from the mutated antibodies with enhanced affinity. However, the existing method for predicting protein affinity change has the problem of low prediction accuracy. SUMMARY

[0005] In view of the above, it is necessary to propose a method for predicting protein affinity change and related equipment, which can assist in predicting protein affinity change and improve the accuracy of predicting protein affinity change.

[0006] The first aspect of the present application provides a method for predicting protein affinity change, the method comprising: constructing a first initial graph network of a protein before mutation, and constructing a second initial graph network of a protein after mutation;

[0007] performing at least one intra-graph update on the first initial graph network, and performing at least one inter-graph update between the first initial graph network and the second initial graph network, until a first target graph network is obtained;

[0008] performing at least one intra-graph update on the second initial graph network, and performing at least one inter-graph update between the second initial graph network and the first initial graph network, until a second target graph network is obtained; and predicting the affinity change between the protein before mutation and the protein after mutation according to the first target graph network and the second target graph network.

[0009] According to an optional embodiment of the present application, any one of the at least one intra-graph update on the first initial graph network comprises: updating node information of a first node in the first initial graph network according to a first graph structure of the first initial graph network;

[0010] Any one of the at least one intra-graph update on the second initial graph network comprises: updating node information of a second node in the second initial graph network according to a second graph structure of the second initial graph network;

[0011] Any one of the at least one inter-graph update between the first initial graph network and the second initial graph network comprises: connecting a first node of the first initial graph network with a second node of the second initial graph network to form a heterogeneous graph, and fusing node information of the second node to node information of the first node in the form of a unidirectional edge in the heterogeneous graph;

[0012] Any one of the at least one inter-graph update between the second initial graph network and the first initial graph network comprises: connecting a first node of the first initial graph network with a second node of the second initial graph network to form a heterogeneous graph, and fusing node information of the first node to node information of the second node in the form of a unidirectional edge in the heterogeneous graph.

[0013] According to an optional embodiment of the present application, the predicting the affinity change between the protein before the mutation and the protein after the mutation according to the first target graph network and the second target graph network comprises:

[0014] taking node information of a first node of the first target graph network as a first target node representation of the first node, and taking node information of a second node of the second target graph network as a second target node representation of the second node;

[0015] obtaining a difference between a first vector space of the first target graph network and a second vector space of the second target graph network based on the first target node representation and the second target node representation by using a preset multi-layer perception neural network, and taking the difference as the affinity change.

[0016] According to an optional embodiment of the present application, the constructing the first initial graph network of the protein before the mutation comprises:

[0017] determining first coordinates of a plurality of first nodes of the first initial graph network according to a three-dimensional space structure of the protein before the mutation;

[0018] determine a first distance between the first nodes according to the first coordinates, and determine a first edge between the first nodes based on the first distance;

[0019] determine a first initial node representation of the first nodes according to the protein before the mutation; and

[0020] construct a first initial graph network based on the first nodes, the first edges and the first initial node representation.

[0021] According to an optional embodiment of the present application, the first coordinates of a plurality of first nodes of the first initial graph network are determined according to the three-dimensional spatial structure of the protein before the mutation, comprising:

[0022] obtain the coordinates of all first alpha carbon atoms in the protein before the mutation according to the three-dimensional spatial structure of the protein before the mutation;

[0023] take a first amino acid in the protein before the mutation as the first node, and take the coordinates of the first alpha carbon atom of the first amino acid as the first coordinates of the corresponding first node.

[0024] According to an optional embodiment of the present application, the first distance between the first nodes is determined according to the first coordinates, and the first edge between the first nodes is determined based on the first distance, comprising:

[0025] when the first distance is less than a preset distance threshold, the first edge is established between the first nodes corresponding to the first distance; or

[0026] when the first distance is greater than or equal to the distance threshold, the first edge is not established between the first nodes corresponding to the first distance.

[0027] According to an optional embodiment of the present application, the first initial node representation of the first nodes is determined according to the protein before the mutation, comprising:

[0028] construct a low-dimensional dense vector of each first amino acid according to the type of each first amino acid, and determine the first initial node representation based on the low-dimensional dense vector.

[0029] According to an optional embodiment of the present application, the second initial graph network of the protein after the mutation comprises:

[0030] determine the three-dimensional spatial structure of the protein after the mutation based on the three-dimensional spatial structure of the protein before the mutation and a predetermined mutation site;

[0031] determine second coordinates of a plurality of second nodes of the second initial graph network according to a three-dimensional spatial structure of the mutated protein;

[0032] determine second distances between the second nodes according to the second coordinates, and determine second edges between the second nodes based on the second distances;

[0033] determine second initial node representations of the second nodes according to the mutated protein; and

[0034] construct the second initial graph network based on the second nodes, the second edges, and the second initial node representations.

[0035] A second aspect of the present application provides a device for predicting a change in protein affinity, the device comprising: the device comprising a constructing module, an updating module, and a predicting module:

[0036] the constructing module is configured to construct a first initial graph network of a pre-mutated protein and a second initial graph network of a post-mutated protein;

[0037] the updating module is configured to perform at least one intra-graph update on the first initial graph network and at least one inter-graph update between the first initial graph network and the second initial graph network until a first target graph network is obtained, and perform at least one intra-graph update on the second initial graph network and at least one inter-graph update between the second initial graph network and the first initial graph network until a second target graph network is obtained;

[0038] the predicting module is configured to predict a change in affinity between the pre-mutated protein and the post-mutated protein according to the first target graph network and the second target graph network.

[0039] A third aspect of the present application provides an electronic device, the electronic device comprising a processor and a memory, the processor being configured to implement the method for predicting a change in protein affinity when executing a computer program stored in the memory.

[0040] A fourth aspect of the present application provides a computer-readable storage medium, the computer-readable storage medium storing a computer program, the computer program being configured to implement the method for predicting a change in protein affinity when executed by a processor.

[0041] In summary, the protein affinity change prediction method and related device provided by the present application can construct a first initial graph network of the protein before mutation and a second initial graph network of the protein after mutation; perform at least one intra-graph update and at least one inter-graph update on the first initial graph network and the second initial graph network until a first target graph network and a second target graph network are obtained; and predict the affinity change between the protein before mutation and the protein after mutation according to the first target graph network and the second target graph network. Compared with the method in the prior art in which only inter-graph update is performed on the first initial graph network and the second initial graph network respectively, the information interaction between the first initial graph network and the second initial graph network is realized by inter-graph update, thereby improving the accuracy of the prediction of the protein affinity change. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 FIG. 1 is a flowchart of a protein affinity change prediction method provided by an embodiment of the present application.

[0043] Figure 2 FIG. 2 is a flowchart of constructing a first initial graph network provided by an embodiment of the present application.

[0044] Figure 3 FIG. 3 is an example graph of protein structures before and after mutation provided by an embodiment of the present application.

[0045] Figure 4 FIG. 4 is an example graph of intra-graph update and inter-graph update provided by an embodiment of the present application.

[0046] Figure 5 FIG. 5 is an example graph of a heterogeneous graph provided by an embodiment of the present application.

[0047] Figure 6 FIG. 6 is a structural diagram of a protein affinity change prediction device provided by an embodiment two of the present application.

[0048] Figure 7 FIG. 7 is a structural schematic diagram of an electronic device provided by an embodiment four of the present application. DETAILED DESCRIPTION

[0049] In order to more clearly understand the above-mentioned purposes, features and advantages of the present application, the present application will be described in detail below with reference to the drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0050] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs. The terms used in the specification of the present application are only for the purpose of describing the embodiments of the present application in one optional embodiment, and are not intended to limit the present application.

[0051] The embodiment of the present application provides a protein affinity change prediction method, which can be based on a graph structure deep learning method, takes the spatial conformation of a protein as input, and interactively updates the node information in and between graphs in the form of a graph structure to obtain node representation after the interaction of the in-graph and inter-graph information. Then the node representation is obtained through a multilayer perceptron (MLP, Multilayer Perceptron) to obtain the representation of the graph and the difference between the graphs, and the representation result of the graph difference is taken as the network output to obtain the change of the protein affinity. The accuracy of the prediction of the change of the protein affinity can be improved.

[0052] The protein affinity change prediction method provided by the embodiment of the present application is executed by an electronic device, and accordingly, the protein affinity change prediction device runs in the electronic device.

[0053] Embodiment one

[0054] Figure 1 is a flowchart of the protein affinity change prediction method provided by the embodiment one of the present application. The protein affinity change prediction method specifically includes the following steps, and the order of the steps in the flowchart can be changed, and some steps can be omitted according to different requirements.

[0055] S11, constructing a first initial graph network of a protein before mutation and a second initial graph network of a protein after mutation.

[0056] In an optional implementation, the protein comprises a plurality of amino acids (for example, aliphatic amino acids, aromatic amino acids, heterocyclic amino acids, etc.), wherein the amino acids can be combined together through a peptide bond (or an amide bond), and the structure center of each amino acid comprises an alpha carbon atom. The protein can be regarded as a three-dimensional spatial structure composed of amino acids and peptide bonds. Among them, the three-dimensional spatial structure of the protein before mutation is known, for example, the position of each amino acid in the known antibody template is known, the position of the alpha carbon atom at the center of the amino acid is known, and the combination mode of the peptide bond between the amino acids is known.

[0057] In addition, by comparing the three-dimensional spatial structure of the protein with the three-dimensional graph neural network structure, it can be found that the amino acids of the protein can be regarded as points in the graph neural network, and the peptide bonds between the amino acids can be regarded as edges in the graph neural network. Therefore, a corresponding graph neural network (hereinafter referred to as a graph network) can be established according to the three-dimensional spatial structure of the protein, so as to analyze the correlation between the amino acids in the protein and the affinity between the proteins.

[0058] In an optional implementation, the constructing the first initial graph network of the protein before mutation includes the following steps S111 to S114 for constructing the first initial graph network as shown in the figure. Figure 2 ​

[0059] Step S111: Determine the first coordinates of multiple first nodes of the first initial graph network based on the three-dimensional spatial structure of the protein before mutation.

[0060] In an optional implementation, determining the first coordinates of multiple first nodes of the first initial graph network based on the three-dimensional spatial structure of the protein before mutation includes:

[0061] Based on the three-dimensional spatial structure of the protein before mutation, obtain the coordinates of all first α carbon atoms in the protein before mutation;

[0062] The first amino acid in the protein before the mutation is taken as the first node, and the coordinates of the first α carbon atom of the first amino acid are taken as the first coordinates of the corresponding first node.

[0063] In an alternative implementation, a first three-dimensional coordinate system of the protein before mutation can be constructed, thereby determining the coordinates of all first α carbon atoms in the first three-dimensional coordinate system based on the position of the first α carbon atom in the three-dimensional spatial structure of the protein before mutation.

[0064] Step S112: Determine the first distance between the first nodes based on the first coordinates, and determine the first edge between the first nodes based on the first distance.

[0065] In an optional implementation, determining the first distance between the first nodes based on the first coordinates, and determining the first edge between the first nodes based on the first distance, includes:

[0066] When the first distance is less than a preset distance threshold, the first edge is established between the first nodes corresponding to the first distance; or

[0067] When the first distance is greater than or equal to the distance threshold, the first edge is not established between the first nodes corresponding to the first distance.

[0068] In an alternative implementation, the distance threshold can be determined based on factors such as the length of the peptide bonds connecting amino acids. For example, if the CN bond length in the peptide bond is 0.132 nm, the CN single bond length is 0.147 nm, and the C=N bond length is 0.127 nm, the distance threshold can be set to 0.12 nm.

[0069] Step S113: Determine the first initial node characterization of the first node based on the protein before mutation.

[0070] In an optional implementation, the determining the first initial node representation of the first node according to the protein before the mutation comprises:

[0071] The low-dimensional dense vector of each first amino acid is constructed according to the category of each first amino acid, and the first initial node representation is determined based on the low-dimensional dense vector.

[0072] In an optional implementation, the same low-dimensional dense vector can be constructed for first amino acids of the same category, and the length of the peptide bond between first amino acids of different categories is mapped to the distance between corresponding low-dimensional dense vectors, and the shorter the length of the peptide bond, the smaller the distance between the low-dimensional dense vectors. The low-dimensional dense vector can be directly used as the first initial node representation, wherein the node representation is used to represent the characteristics of the node, and the node representation can be obtained by encoding the nodes in the graph based on embedding technology, for example, the first initial node representation is encoded into a vector form.

[0073] In an optional implementation, the first initial node representation indicates the relationship between adjacent first amino acids, and does not consider the relationship between all first amino acids.

[0074] In step S114, the first initial graph network is constructed based on the first node, the first edge and the first initial node representation.

[0075] In an optional implementation, after obtaining the first node, the first edge between the pair of first nodes, and the first initial node representation of the first node, the first initial graph network is preliminarily constructed.

[0076] In an optional implementation, the constructing the second initial graph network of the protein after the mutation comprises:

[0077] The three-dimensional spatial structure of the protein after the mutation is determined based on the three-dimensional spatial structure of the protein before the mutation and the predetermined mutation site.

[0078] The second coordinates of a plurality of second nodes of the second initial graph network are determined according to the three-dimensional spatial structure of the protein after the mutation.

[0079] The second distances between the second nodes are determined according to the second coordinates, and the second edges between the second nodes are determined based on the second distances.

[0080] The second initial node representation of the second node is determined according to the protein after the mutation; and

[0081] The second initial graph network is constructed based on the second node, the second edge and the second initial node representation.

[0082] In an optional embodiment, the protein before mutation can be a known antibody template of an antigen, the mutation sites of the protein before mutation can be determined in advance, the proteins after mutation are obtained, and whether the affinity of the proteins after mutation is increased compared with the protein before mutation is compared, so as to obtain an antibody with higher affinity to the antigen, and achieve a better antibody treatment effect.

[0083] In an optional embodiment, after the mutation sites are determined in advance, the three-dimensional spatial structure of the protein after mutation can be determined based on the three-dimensional spatial structure of the protein before mutation, for example Figure 3 As shown in the example diagram of the protein structure before and after mutation provided by the embodiment of the present application.

[0084] In an optional embodiment, after the three-dimensional spatial structure of the protein after mutation is determined, the step of constructing the second initial graph network is similar to the steps S111-S114 described above, and will not be repeated.

[0085] S12, at least one intra-graph update is performed on the first initial graph network, and at least one inter-graph update is performed between the first initial graph network and the second initial graph network, until a first target graph network is obtained;

[0086] At least one intra-graph update is performed on the second initial graph network, and at least one inter-graph update is performed between the second initial graph network and the first initial graph network, until a second target graph network is obtained.

[0087] In an optional embodiment, the first initial node representation indicates the relationship between adjacent first nodes, and does not consider the relationship between all first nodes, therefore, at least one intra-graph update is required to be performed on the first initial graph network to obtain the global relationship between all first nodes, and the same is true for the second initial graph network. In addition, after each intra-graph update, an inter-graph update is performed to complete the interaction of node information between the two graph networks. In other embodiments, an intra-graph update can also be performed after each inter-graph update, or an inter-graph update can be performed after all intra-graph updates are completed. The present application does not limit the order of each intra-graph update and inter-graph update, as long as the interaction of node information between the two graph networks can be completed.

[0088] In an optional embodiment, any one of the at least one intra-graph update of the first initial graph network includes updating the node information of the first node in the first initial graph network according to the first graph structure of the first initial graph network;

[0089] The any one of the at least one intra-graph update on the second initial graph network comprises: updating node information of a second node in the second initial graph network according to a second graph structure of the second initial graph network;

[0090] The any one of the at least one inter-graph update between the first initial graph network and the second initial graph network comprises: connecting a first node of the first initial graph network and a second node of the second initial graph network to form a heterogeneous graph, and fusing node information of the second node to node information of the first node in the form of a unidirectional edge in the heterogeneous graph;

[0091] The any one of the at least one inter-graph update between the first initial graph network and the second initial graph network comprises: connecting a first node of the first initial graph network and a second node of the second initial graph network to form a heterogeneous graph, and fusing node information of the second node to node information of the first node in the form of a unidirectional edge in the heterogeneous graph;

[0092] In an optional embodiment, the graph structure represents a structure composed of points and edges in a graph network, and the node information comprises node representation of the node. Each intra-graph update comprises one update of the node information of the first node and one update of the node information of the second node. In addition, in each intra-graph update, node information containing more relationships between nodes is obtained.

[0093] For example, when the node information of the first node in the first initial graph network is updated for the first time, the first initial node representation obtained according to each two adjacent first nodes is updated to a node representation obtained according to the distance between each three adjacent first nodes, and the same can be sequentially applied to subsequent intra-graph updates, and node information containing more relationships between first nodes is obtained. For example, Figure 4 As shown in the figure, the intra-graph update and the inter-graph update provided by the embodiments of the present application are shown, wherein inner represents intra-graph, outer represents inter-graph, update represents update, x i represents node information of the first initial graph network, x j represents node information of the second initial graph network.

[0094] In an optional embodiment, when the inter-graph information update is performed, the first node of the first initial graph network and the second node of the second initial graph network are connected to form a heterogeneous graph, for example Figure 5 As shown in the figure, an example of a heterogeneous graph provided by the embodiments of the present application is shown, wherein a light-colored thin line without an arrow represents a heterogeneous connection between graphs.

[0095] In an optional implementation, during any one of the at least one inter-graph updates between the first initial graph network and the second initial graph network, the node information of the second node is fused to the node information of the first node in the form of a one-way connection in the heterogeneous graph, including: introducing the node where the mutation site is located in the second initial graph network G2 as a supernode into the first initial graph network G1, and performing a one-way update from the second initial graph network G2 to the first initial graph network G1 on other nodes besides the node where the mutation site is located.

[0096] Specifically, in G1, for the N1i node where mutation site i is located, the node information x of the N2i node where mutation site i is located in G2 is used. j Node information x for node N1i i After updating, the updated node information obtained by node N1i can be represented as (x j ,x i ).

[0097] In an optional implementation, during any one of the at least one inter-graph updates between the second initial graph network and the first initial graph network, the node information of the first node is fused into the node information of the second node in the form of a one-way connection in the heterogeneous graph, including: introducing the node where the mutation site is located in the first initial graph network G1 as a supernode into the second initial graph network G2, and performing a one-way update from the first initial graph network G1 to the second initial graph network G2 on other nodes besides the node where the mutation site is located.

[0098] Specifically, in G2, for mutation site i, the node information x of the N1i node in G1 is used. i Node information x for N2i node j After updating, the updated node information obtained by the N2i node can be represented as (x i ,x j This completes the update of node information between graphs in the current update cycle.

[0099] Furthermore, during inter-graph updates, the node information updates are directed, (x j ,x i ) is different from (x) i ,x j ).

[0100] In an optional implementation, the first target graph network represents the graph network obtained after the first initial graph network completes all intra-graph updates and inter-graph updates, and the second target graph network represents the graph network obtained after the second initial graph network completes all intra-graph updates and inter-graph updates.

[0101] S13, predicting the affinity change between the protein before the mutation and the protein after the mutation according to the first target graph network and the second target graph network.

[0102] In an optional implementation, the step of predicting the affinity change between the protein before the mutation and the protein after the mutation according to the first target graph network and the second target graph network comprises:

[0103] characterizing the node information of the first node of the first target graph network as a first target node representation of the first node, and characterizing the node information of the second node of the second target graph network as a second target node representation of the second node;

[0104] obtaining the difference between the first vector space of the first target graph network and the second vector space of the second target graph network based on the first target node representation and the second target node representation, and taking the difference as the affinity change.

[0105] In an optional implementation, the multi-layer perception neural network comprises an input layer, a plurality of hidden layers, and an output layer. The first vector space of the first target graph network and the second vector space of the second target graph network can be input into the multi-layer perception neural network, and the difference between the two vector spaces is calculated and output by the multi-layer perception neural network, so as to obtain the affinity change between the protein before the mutation and the protein after the mutation.

[0106] The method for predicting the affinity change of the protein provided in the embodiments of the present application comprises the following steps: constructing a first initial graph network of the protein before the mutation, and constructing a second initial graph network of the protein after the mutation; performing at least one intra-graph update and at least one inter-graph update on the first initial graph network and the second initial graph network until a first target graph network and a second target graph network are obtained; and predicting the affinity change between the protein before the mutation and the protein after the mutation according to the first target graph network and the second target graph network, thereby improving the accuracy of the prediction of the affinity change of the protein.

[0107] Embodiment Two

[0108] Figure 6 is a structural diagram of the device for predicting the affinity change of the protein provided in Embodiment Two of the present application.

[0109] In some embodiments, the device 20 for predicting the affinity change of the protein can comprise a plurality of functional modules composed of computer program segments. The computer programs of each program segment in the device 20 for predicting the affinity change of the protein can be stored in the memory of an electronic device and executed by at least one processor to perform the following steps (see the following description in detail):Figure 1 a function of predicting the change of protein affinity.

[0110] In this embodiment, the protein affinity change prediction device 20 can be divided into a plurality of functional modules according to the functions performed thereby. The functional modules can include a construction module 201, an update module 202, and a prediction module 203. The term "module" as used herein refers to a series of computer program segments that can be executed by at least one processor and that can complete a fixed function, which are stored in a memory. In this embodiment, the definition of the protein affinity change prediction device 20 can refer to the definition of the protein affinity change prediction method described above, and will not be described in detail here.

[0111] The construction module 201 is configured to construct a first initial graph network of a protein before mutation and a second initial graph network of a protein after mutation.

[0112] The update module 202 is configured to perform at least one intra-graph update on the first initial graph network and at least one inter-graph update between the first initial graph network and the second initial graph network until a first target graph network is obtained, and to perform at least one intra-graph update on the second initial graph network and at least one inter-graph update between the second initial graph network and the first initial graph network until a second target graph network is obtained.

[0113] The prediction module 203 is configured to predict the change of affinity between the protein before mutation and the protein after mutation according to the first target graph network and the second target graph network.

[0114] In an optional implementation, the update module 203 is further configured to:

[0115] Any one of the at least one intra-graph update on the first initial graph network includes updating node information of a first node in the first initial graph network according to a first graph structure of the first initial graph network.

[0116] Any one of the at least one intra-graph update on the second initial graph network includes updating node information of a second node in the second initial graph network according to a second graph structure of the second initial graph network.

[0117] Any one of the at least one inter-graph update between the first initial graph network and the second initial graph network includes connecting edges between the first node of the first initial graph network and the second node of the second initial graph network to form a heterogeneous graph, and fusing the node information of the second node into the node information of the first node in the form of a unidirectional edge in the heterogeneous graph.

[0118] Any one of the inter-graph updates between the second initial graph network and the first initial graph network comprises: connecting edges between a first node of the first initial graph network and a second node of the second initial graph network and forming a heterogeneous graph, and fusing node information of the first node to node information of the second node in the form of a unidirectional edge in the heterogeneous graph.

[0119] In an optional implementation, the prediction module 203 is further configured to: take node information of a first node of the first target graph network as a first target node representation of the first node, and take node information of a second node of the second target graph network as a second target node representation of the second node; and based on the first target node representation and the second target node representation, obtain a difference between a first vector space of the first target graph network and a second vector space of the second target graph network by using a preset multi-layer perception neural network, and take the difference as the affinity change.

[0120] In an optional implementation, the construction module 201 is further configured to:

[0121] The first initial graph network of the protein before the mutation comprises:

[0122] According to the three-dimensional spatial structure of the protein before the mutation, first coordinates of a plurality of first nodes of the first initial graph network are determined;

[0123] According to the first coordinates, first distances between the first nodes are determined, and first edges between the first nodes are determined based on the first distances;

[0124] According to the protein before the mutation, first initial node representations of the first nodes are determined; and

[0125] Based on the first nodes, the first edges, and the first initial node representations, the first initial graph network is constructed.

[0126] In an optional implementation, the construction module 201 is further configured to:

[0127] According to the three-dimensional spatial structure of the protein before the mutation, first coordinates of a plurality of first nodes of the first initial graph network are determined, comprising:

[0128] According to the three-dimensional spatial structure of the protein before the mutation, coordinates of all first alpha carbon atoms in the protein before the mutation are obtained;

[0129] The first node is a first amino acid in the protein before the mutation, and a first coordinate of a first alpha carbon atom of the first amino acid is the first coordinate of the first node.

[0130] In an optional implementation, the construction module 201 is further configured to:

[0131] The first distance between the first nodes is determined according to the first coordinates, the first edges between the first nodes are determined based on the first distance, and the method comprises:

[0132] When the first distance is less than a preset distance threshold, the first edges are established between the first nodes corresponding to the first distance; or

[0133] When the first distance is greater than or equal to the distance threshold, the first edges are not established between the first nodes corresponding to the first distance.

[0134] In an optional implementation, the construction module 201 is further configured to:

[0135] The first initial node representation of the first node is determined according to the protein before the mutation, and the method comprises:

[0136] A low-dimensional dense vector of each first amino acid is constructed according to the type of each first amino acid, and the first initial node representation is determined based on the low-dimensional dense vector.

[0137] In an optional implementation, the construction module 201 is further configured to:

[0138] The second initial graph network of the protein after the mutation comprises:

[0139] The three-dimensional space structure of the protein after the mutation is determined based on the three-dimensional space structure of the protein before the mutation and the predetermined mutation site;

[0140] The second coordinates of a plurality of second nodes of the second initial graph network are determined according to the three-dimensional space structure of the protein after the mutation;

[0141] The second distance between the second nodes is determined according to the second coordinates, and the second edges between the second nodes are determined based on the second distance;

[0142] The second initial node representation of the second node is determined according to the protein after the mutation; and

[0143] The second initial graph network is constructed based on the second nodes, the second edges and the second initial node representation.

[0144] Embodiment Three

[0145] The embodiment provides a computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the steps in the above-mentioned protein affinity change prediction embodiment, for example Figure 1 S11-S13 shown in the figure:

[0146] S11, constructing a first initial graph network of a protein before mutation and a second initial graph network of a protein after mutation;

[0147] S12, performing at least one intra-graph update on the first initial graph network and at least one inter-graph update between the first initial graph network and the second initial graph network until a first target graph network is obtained; performing at least one intra-graph update on the second initial graph network and at least one inter-graph update between the second initial graph network and the first initial graph network until a second target graph network is obtained; S13, predicting an affinity change between the protein before mutation and the protein after mutation according to the first target graph network and the second target graph network.

[0148] Alternatively, the computer program is executed by a processor to implement the functions of various modules / units in the above-mentioned device embodiment, for example Figure 6 Modules 201-203 in the figure:

[0149] The construction module 201 is configured to construct a first initial graph network of a protein before mutation and a second initial graph network of a protein after mutation.

[0150] The update module 202 is configured to perform at least one intra-graph update on the first initial graph network and at least one inter-graph update between the first initial graph network and the second initial graph network until a first target graph network is obtained; and perform at least one intra-graph update on the second initial graph network and at least one inter-graph update between the second initial graph network and the first initial graph network until a second target graph network is obtained.

[0151] The prediction module 203 is configured to predict an affinity change between the protein before mutation and the protein after mutation according to the first target graph network and the second target graph network.

[0152] Embodiment four

[0153] Referring to Figure 7 The electronic device 3 provided in the fourth embodiment of the present application includes a memory 31, at least one processor 32, at least one communication bus 33, and a transceiver 34.

[0154] Those skilled in the art should understand that, Figure 7 The structure of the electronic device shown does not constitute a limitation of the embodiments of the present application, and can be a bus structure or a star structure. The electronic device 3 can also include more or fewer other hardware or software, or different component arrangements than those shown.

[0155] In some embodiments, the electronic device 3 is a device capable of automatically performing numerical calculation and / or information processing according to pre-set or stored instructions. The hardware of the electronic device 3 includes, but is not limited to, a microprocessor, an application-specific integrated circuit, a programmable gate array, a digital processor, and an embedded device. The electronic device 3 can also include a client device, which includes, but is not limited to, any electronic product that can interact with a client through a keyboard, a mouse, a remote control, a touchpad, or a voice control device, such as a personal computer, a tablet computer, a smartphone, a digital camera, and the like.

[0156] It should be noted that the electronic device 3 is only an example, and other existing or future electronic products, such as those that can be adapted to the present application, should also be included within the scope of the present application and are hereby incorporated by reference.

[0157] In some embodiments, the memory 31 stores a computer program that, when executed by the at least one processor 32, implements all or part of the steps of the method for predicting changes in protein affinity as described. The memory 31 includes a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electrically-erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disk memory, a magnetic disk memory, a magnetic tape memory, or any other computer-readable medium capable of carrying or storing data.

[0158] Further, the computer readable storage medium can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, application programs required by at least one function, and the like; and the data storage area can store data created according to the use of the blockchain node, and the like.

[0159] The blockchain referred to in the present application is a new application mode of distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm and other computer technologies. The blockchain is essentially a decentralized database, which is a series of data blocks associated using cryptographic methods, each data block contains information of a batch of network transactions, and is used to verify the validity (anti-fake) of the information and generate the next block. The blockchain can include a blockchain underlying platform, a platform product service layer, and an application service layer, etc.

[0160] In some embodiments, the at least one processor 32 is a control core of the electronic device 3, which connects various components of the entire electronic device 3 through various interfaces and lines, and performs various functions of the electronic device 3 and processes data by running or executing programs or modules stored in the memory 31 and calling data stored in the memory 31. For example, the at least one processor 32 implements all or part of the steps of the prediction method of the change in protein affinity in the embodiments of the present application when executing the computer program stored in the memory, or implements all or part of the functions of the prediction device of the change in protein affinity. The at least one processor 32 can be composed of integrated circuits, for example, can be composed of a single packaged integrated circuit, or can be composed of multiple packaged integrated circuits with the same function or different functions, including one or more central processing units (CPU), microprocessors, digital processing chips, graphics processors, and combinations of various control chips, etc.

[0161] In some embodiments, the at least one communication bus 33 is configured to realize the connection and communication between the memory 31, the at least one processor 32, and the like.

[0162] Although not shown, the electronic device 3 can also include a power supply (such as a battery) for powering various components. Preferably, the power supply can be logically connected to the at least one processor 32 through a power management device, so as to realize the functions of managing charging, discharging, and power consumption management, etc. through the power management device. The power supply can also include one or more direct current or alternating current power supplies, recharging devices, power supply fault detection circuits, power supply converters or inverters, power supply status indicators, and the like. The electronic device 3 can also include various sensors, Bluetooth modules, Wi-Fi modules, camera devices, and the like, which are not described here again.

[0163] The integrated unit implemented as a software functional module described above can be stored in a computer-readable storage medium. This software functional module, stored in a storage medium, includes several instructions to cause a computer device (which may be a personal computer, electronic device, or network device, etc.) or processor to execute portions of the methods described in the various embodiments of this application.

[0164] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0165] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0166] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0167] It will be apparent to those skilled in the art that this application is not limited to the details of the exemplary embodiments described above, and that it can be implemented in other specific forms without departing from the spirit or essential characteristics of this application. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of this application is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within this application. No reference numerals in the claims should be construed as limiting the scope of the claims. Furthermore, it is clear that the word "comprising" does not exclude other elements or, and the singular does not exclude the plural. Multiple elements or devices recited in the specification may also be implemented by a single element or device through software or hardware. The terms "first," "second," etc., are used to indicate names and do not indicate any particular order.

[0168] Finally, it should be noted that the above examples are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present application.

Claims

1. A method for predicting a change in protein affinity, characterized by, The method comprises: constructing a first initial graph network of the protein before mutation, comprising: determining first coordinates of a plurality of first nodes of the first initial graph network according to a three-dimensional spatial structure of the protein before mutation; determining first distances between the first nodes according to the first coordinates, and determining first edges between the first nodes based on the first distances; determining first initial node representations of the first nodes according to the protein before mutation; and constructing the first initial graph network based on the first nodes, the first edges, and the first initial node representations; constructing a second initial graph network of the protein after mutation, comprising: determining a three-dimensional spatial structure of the protein after mutation based on the three-dimensional spatial structure of the protein before mutation and a predetermined mutation site; determining second coordinates of a plurality of second nodes of the second initial graph network according to the three-dimensional spatial structure of the protein after mutation; determining second distances between the second nodes according to the second coordinates, and determining second edges between the second nodes based on the second distances; determining second initial node representations of the second nodes according to the protein after mutation; and constructing the second initial graph network based on the second nodes, the second edges, and the second initial node representations; performing at least one intra-graph update on the first initial graph network, and performing at least one inter-graph update between the first initial graph network and the second initial graph network until a first target graph network is obtained; any one of the at least one inter-graph update between the first initial graph network and the second initial graph network comprises: connecting edges between first nodes of the first initial graph network and second nodes of the second initial graph network to form a heterogeneous graph, and fusing node information of the second nodes to node information of the first nodes in the form of one-way edges in the heterogeneous graph; performing at least one intra-graph update on the second initial graph network, and performing at least one inter-graph update between the second initial graph network and the first initial graph network until a second target graph network is obtained; any one of the at least one inter-graph update between the second initial graph network and the first initial graph network comprises: connecting edges between first nodes of the first initial graph network and second nodes of the second initial graph network to form a heterogeneous graph, and fusing node information of the first nodes to node information of the second nodes in the form of one-way edges in the heterogeneous graph; predicting an affinity change between the protein before mutation and the protein after mutation according to the first target graph network and the second target graph network.

2. The method for predicting a protein affinity change according to claim 1, wherein: any one of the at least one intra-graph update on the first initial graph network comprises: updating node information of first nodes in the first initial graph network according to a first graph structure of the first initial graph network. The any one of the at least one intra-graph update on the second initial graph network comprises: updating node information of a second node in the second initial graph network according to a second graph structure of the second initial graph network.

3. The method of predicting changes in protein affinity according to claim 2, wherein, The predicting the affinity change between the protein before the mutation and the protein after the mutation according to the first target graph network and the second target graph network comprises: taking node information of a first node of the first target graph network as a first target node representation of the first node, and taking node information of a second node of the second target graph network as a second target node representation of the second node; obtaining a difference between a first vector space of the first target graph network and a second vector space of the second target graph network based on the first target node representation and the second target node representation by using a preset multi-layer perception neural network, and taking the difference as the affinity change.

4. The method for predicting a change in protein affinity according to claim 2 or 3, wherein The determining the first coordinates of the plurality of first nodes of the first initial graph network according to the three-dimensional spatial structure of the protein before the mutation comprises: obtaining coordinates of all first alpha carbon atoms in the protein before the mutation according to the three-dimensional spatial structure of the protein before the mutation; taking a first amino acid in the protein before the mutation as the first node, and taking the coordinates of the first alpha carbon atom of the first amino acid as the first coordinate of the corresponding first node.

5. The method of predicting changes in protein affinity according to claim 1, wherein, The determining the first distances between the first nodes according to the first coordinates, and determining the first edges between the first nodes based on the first distances comprises: establishing the first edge between the first nodes corresponding to the first distance when the first distance is less than a preset distance threshold; or not establishing the first edge between the first nodes corresponding to the first distance when the first distance is greater than or equal to the distance threshold.

6. The method of predicting changes in protein affinity according to claim 1, wherein, The determining the first initial node representation of the first node according to the protein before the mutation comprises: constructing a low-dimensional dense vector of each first amino acid according to a category of each first amino acid, and determining the first initial node representation based on the low-dimensional dense vector.

7. An electronic device, comprising: The electronic device comprises a processor and a memory, and the processor is configured to implement the prediction method of the protein affinity change as claimed in any one of claims 1 to 6 when executing a computer program stored in the memory.

8. A computer-readable storage medium having stored thereon a computer program, characterized in that The computer program is configured to implement the prediction method of the protein affinity change as claimed in any one of claims 1 to 6 when executed by the processor.

Images