Method and device for training relation prediction model and method and device for predicting interaction relation

Abstract

The invention relates to a relationship prediction model training method and device and an interaction relationship prediction method and device, and relates to the technical field of machine learning, and the method comprises the steps: constructing a first data set according to first historical proteins and an actual association relationship value between the first historical proteins; training the local feature extraction model according to the first data set to obtain a first interaction relationship prediction model, and inputting the second historical protein into the first interaction relationship prediction model to obtain a first prediction association relationship value; constructing a second data set according to the second historical protein and the first prediction association relationship value, and training the global feature extraction model according to the first data set and the second data set to obtain a second interaction relationship prediction model; and constructing a target relationship prediction model according to the first interaction relationship prediction model and the second interaction relationship prediction model. The accuracy of the target relation prediction model is improved.

Description

Background Art

[0002] In existing model methods, local information of proteins is utilized while global information is ignored, which results in low accuracy of the model.

[0003] It should be noted that the information disclosed in the above background technology section is only used to enhance the understanding of the background of the present disclosure, and therefore may include information that does not constitute the prior art known to ordinary technicians in the field. Summary of the invention

[0004] The purpose of the present disclosure is to provide a training method for a relationship prediction model, a training device for a relationship prediction model, a method for predicting an interaction relationship, a device for predicting an interaction relationship, a computer-readable storage medium, and an electronic device, thereby at least to a certain extent overcoming the problem of low model accuracy caused by the limitations and defects of related technologies.

[0005] According to one aspect of the present disclosure, a method for training a relationship prediction model is provided, comprising:

[0006] Extracting first historical proteins with correlation and second historical proteins without correlation in the historical biological data, and constructing a first data set according to the first historical proteins and the actual correlation values ​​between the first historical proteins;

[0007] Training a local feature extraction model according to the first data set to obtain a first interaction relationship prediction model, and inputting the second historical protein into the first interaction relationship prediction model to obtain a first predicted association relationship value;

[0008] constructing a second data set according to the second historical protein and the first predicted association relationship value, and training a global feature extraction model according to the first data set and the second data set to obtain a second interaction relationship prediction model;

[0009] A target relationship prediction model is constructed based on the first interaction relationship prediction model and the second interaction relationship prediction model.

[0010] In an exemplary embodiment of the present disclosure, a first data set is constructed according to the first historical protein and the actual association relationship value between the first historical proteins, including:

[0011] A first historical protein pair is formed according to the first historical proteins having an association relationship, and a first data set is constructed according to the first historical protein pair and the actual association relationship values ​​between the first historical proteins in the first historical protein pair.

[0012] In an exemplary embodiment of the present disclosure, a local feature extraction model is trained according to a first data set to obtain a first interaction relationship prediction model, including:

[0013] Inputting the first historical protein pair in the first data set into the local feature extraction model to obtain a first prediction result between the first historical proteins in the first historical protein pair;

[0014] According to the first prediction result and the actual association relationship value, a first target loss function is constructed, and based on the first target loss function, the first model parameters in the local feature extraction model are adjusted to obtain a first interaction relationship prediction model.

[0015] In an exemplary embodiment of the present disclosure, the local feature extraction model includes a first feature encoding layer, a first feature mapping layer, a first feature sequence extraction layer and a first classification layer; the first historical protein pair includes a first sub-protein and a second sub-protein;

[0016] The first historical protein pair in the first data set is input into the local feature extraction model to obtain a first prediction result between the first historical proteins in the first historical protein pair, including:

[0017] Obtaining a first amino acid sequence of a first subprotein and a second amino acid sequence of a second subprotein in a first historical protein pair in a first data set, and constructing a first amino acid vector and a second amino acid vector according to a first amino acid fragment in the first amino acid sequence and a second amino acid fragment in the second amino acid sequence;

[0018] Using a first feature encoding layer to perform one-hot encoding on the first amino acid vector and the second amino acid vector to obtain a first encoding result and a second encoding result, and using a first feature mapping layer to perform dense mapping on the first encoding result and the second encoding result to obtain a first dense mapping result and a second dense mapping result;

[0019] Using a first feature sequence extraction layer to perform temporal feature extraction on the first dense mapping result and the second dense mapping result, to obtain a first feature vector of the first sub-protein based on the first amino acid sequence and a second feature vector of the second sub-protein based on the second amino acid sequence;

[0020] The first feature vector and the second feature vector are input into a first classification layer to obtain a first prediction result between the first sub-protein and the second sub-protein.

[0021] In an exemplary embodiment of the present disclosure, the first feature sequence extraction layer includes a plurality of feature sequence prediction networks and a first concatenation layer, and the number of the feature sequence prediction networks is consistent with the number of first amino acid fragments in the first amino acid vector;

[0022] The first dense mapping result is subjected to temporal feature extraction using the first feature sequence extraction layer to obtain a first feature vector of the first sub-protein based on the first amino acid sequence, including:

[0023] According to the fragment position of the first amino acid fragment in the first amino acid vector, the first sub-mapping result corresponding to each first amino acid fragment in the first dense mapping result is input into the feature sequence prediction network corresponding to the fragment position to obtain a first sub-sequence vector;

[0024] A first weight value of the first subsequence vector is calculated, and the first subsequence vector and the first weight value are weighted and summed through a first concatenation layer to obtain a first feature vector of the first subprotein based on the first amino acid sequence.

[0025] In an exemplary embodiment of the present disclosure, the first model parameters include a first encoding parameter corresponding to a first feature encoding layer, a first mapping parameter corresponding to a first feature mapping layer, a first feature extraction parameter corresponding to a first feature sequence extraction layer, and a first classification parameter corresponding to a first classification layer;

[0026] The first model parameter in the local feature extraction model is adjusted based on the first objective loss function to obtain a first interaction relationship prediction model, including:

[0027] Based on the first objective loss function, the first encoding parameter, the first mapping parameter, the first feature extraction parameter and the first classification parameter in the local feature extraction model are adjusted to obtain a first interaction relationship prediction model.

[0028] In an exemplary embodiment of the present disclosure, constructing a second data set according to the second historical protein and the first predicted association value includes:

[0029] Determining whether the first predicted association relationship value between the second historical proteins is greater than a preset threshold;

[0030] When it is determined that the first predicted association value is greater than the preset threshold, forming a second historical protein pair according to a second historical protein corresponding to the first predicted association value;

[0031] A second data set is constructed based on the second historical protein pairs and the first predicted association relationship values.

[0032] In an exemplary embodiment of the present disclosure, a global feature extraction model is trained according to the first data set and the second data set to obtain a second interaction relationship prediction model, including:

[0033] Constructing an original feature map according to the first data set and the second data set, and inputting the original feature map into a global feature extraction model to obtain a second prediction result;

[0034] A second target loss function is constructed according to the second prediction result and the actual association relationship value and / or the first predicted association relationship value, and the second model parameters in the global feature extraction model are adjusted based on the second target loss function to obtain a second interaction relationship prediction model.

[0035] In an exemplary embodiment of the present disclosure, constructing an original feature map according to the first data set and the second data set includes:

[0036] Obtaining first historical protein pairs and actual association relationship values ​​between the first historical protein pairs in the first data set, and second historical protein pairs and first predicted association relationship values ​​between the second historical protein pairs in the second data set;

[0037] The first subprotein and the second subprotein in the first historical protein pair are the first vertex and the second vertex, the actual association relationship is the first connecting edge, and the actual association relationship value is the first edge weight of the first connecting edge;

[0038] The third subprotein and the fourth subprotein in the second historical protein pair are the third vertex and the fourth vertex, the first predicted association relationship is the second connecting edge, and the first predicted association relationship value is the second edge weight of the second connecting edge;

[0039] An original feature graph is constructed according to the first vertex, the second vertex, the third vertex, the fourth vertex, the first connecting edge, the second connecting edge, the first edge weight, and the second edge weight.

[0040] In an exemplary embodiment of the present disclosure, the global feature extraction model includes a graph convolutional neural network and a second classification layer;

[0041] The original feature map is input into the global feature extraction model to obtain a second prediction result, including:

[0042] Inputting the original feature map into the graph convolutional neural network, updating the first connecting edge and / or the second connecting edge in the original feature map to obtain a target feature map;

[0043] The target feature map is input into the second classification layer to obtain the second prediction result.

[0044] In an exemplary embodiment of the present disclosure, a target relationship prediction model is constructed according to the first interaction relationship prediction model and the second interaction relationship prediction model, including:

[0045] Inputting the second historical protein into the second interaction relationship prediction model to obtain a second predicted association relationship value, and constructing a third data set based on the second historical protein and the second predicted association relationship value;

[0046] Training the local feature extraction model according to the first data set and the third data set to obtain an updated first interaction relationship prediction model;

[0047] A target relationship prediction model is constructed based on the updated first interaction relationship prediction model and the second interaction relationship prediction model.

[0048] According to one aspect of the present disclosure, a method for predicting an interaction relationship is provided, comprising:

[0049] Acquire a first protein to be predicted and a second protein to be predicted, and construct a protein pair to be predicted based on the first protein to be predicted and the second protein to be predicted;

[0050] Inputting the protein pair to be predicted into a target relationship prediction model to obtain a relationship prediction result; wherein the target relationship prediction model is trained by any one of the relationship prediction model training methods described above;

[0051] According to the relationship prediction result, the interaction relationship between the first protein to be predicted and the second protein to be predicted is determined.

[0052] In an exemplary embodiment of the present disclosure, the target relationship prediction model includes a first interaction relationship prediction model and a second interaction relationship prediction model;

[0053] The protein pair to be predicted is input into the target relationship prediction model to obtain a relationship prediction result, including:

[0054] Inputting the protein pair to be predicted into a first interaction relationship prediction model to obtain a first relationship prediction result, and inputting the protein pair to be predicted into a second interaction relationship prediction model to obtain a second relationship prediction result;

[0055] The first relationship prediction result and the second relationship prediction result are weightedly summed to obtain the relationship prediction result.

[0056] In an exemplary embodiment of the present disclosure, obtaining a first protein to be predicted and a second protein to be predicted includes:

[0057] In response to a touch operation on a first interactive control on the display interface, obtaining a user identifier of a user to be predicted; wherein the first interactive control includes a user identifier input box;

[0058] In response to a touch operation on a second interactive control on the display interface, a first protein to be predicted associated with the user to be predicted and a second protein to be predicted associated with the first protein to be predicted are obtained based on the user identifier; wherein the second interactive control includes a relationship prediction interactive control.

[0059] In an exemplary embodiment of the present disclosure, after determining the interaction relationship between the first protein to be predicted and the second protein to be predicted, the method for predicting the interaction relationship further includes:

[0060] A relationship prediction report associated with the user to be predicted is generated based on the interaction relationship, and the relationship prediction report is displayed on the display interface.

[0061] According to one aspect of the present disclosure, a training device for a relationship prediction model is provided, comprising:

[0062] A first data set construction module is used to extract first historical proteins with correlation relationship and second historical proteins without correlation relationship in historical biological data, and construct a first data set according to the first historical proteins and the actual correlation relationship values ​​between the first historical proteins;

[0063] A first model training module is used to train a local feature extraction model according to a first data set to obtain a first interaction relationship prediction model, and input the second historical protein into the first interaction relationship prediction model to obtain a first predicted association relationship value;

[0064] A second model training module is used to construct a second data set according to the second historical protein and the first predicted association relationship value, and train a global feature extraction model according to the first data set and the second data set to obtain a second interaction relationship prediction model;

[0065] The target relationship prediction model construction module is used to construct a target relationship prediction model according to the first interaction relationship prediction model and the second interaction relationship prediction model.

[0066] According to one aspect of the present disclosure, there is provided a prediction device for interaction relationship, comprising:

[0067] A protein pair construction module to be predicted, used to obtain a first protein to be predicted and a second protein to be predicted, and to construct a protein pair to be predicted according to the first protein to be predicted and the second protein to be predicted;

[0068] A relationship prediction result determination module, used for inputting the protein pair to be predicted into a target relationship prediction model to obtain a relationship prediction result; wherein the target relationship prediction model is trained by any one of the relationship prediction model training methods described above;

[0069] The interaction relationship determination module is used to determine the interaction relationship between the first protein to be predicted and the second protein to be predicted according to the relationship prediction result.

[0070] According to one aspect of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the training method of the relationship prediction model described in any one of the above and the prediction method of the interaction relationship described in any one of the above are implemented.

[0071] According to one aspect of the present disclosure, there is provided an electronic device, including:

[0072] Processor; and

[0073] A memory, configured to store executable instructions of the processor;

[0074] Wherein, the processor is configured to execute any one of the above-mentioned relationship prediction model training methods and any one of the above-mentioned interaction relationship prediction methods by executing the executable instructions.

[0075] The present disclosure discloses a method for training a relationship prediction model. On the one hand, a first historical protein with an association relationship and a second historical protein without an association relationship are extracted from historical biological data, and a first data set is constructed based on the actual association relationship value between the first historical protein and the first historical protein; then a local feature extraction model is trained based on the first data set to obtain a first interaction relationship prediction model, and the second historical protein is input into the first interaction relationship prediction model to obtain a first predicted association relationship value; then a second data set is constructed based on the second historical protein and the first predicted association relationship value, and a global feature extraction model is trained based on the first data set and the second data set to obtain a second interaction relationship prediction model; finally, a target relationship prediction model is constructed based on the first interaction relationship prediction model and the second interaction relationship prediction model; since the global features and local features of the protein are taken into account at the same time during the model training process, the problem of only using the local information of the protein and ignoring the global information in the prior art, which results in low model accuracy, is solved; on the other hand, the accuracy of the prediction results obtained by performing relationship prediction through the target relationship prediction model is also improved.

[0076] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] The accompanying drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present disclosure, and together with the specification are used to explain the principles of the present disclosure. Obviously, the accompanying drawings described below are only some embodiments of the present disclosure, and for ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without creative work.

[0078] Figure 1 A flowchart of a method for training a relationship prediction model according to an example embodiment of the present disclosure is schematically shown.

[0079] Figure 2 A diagram schematically shows an example structure of a local feature extraction model according to an example embodiment of the present disclosure.

[0080] Figure 3 A diagram schematically shows an example structure of a global feature extraction model according to an example embodiment of the present disclosure.

[0081] Figure 4 A diagram schematically shows an example structure of a target relationship prediction model according to an example embodiment of the present disclosure.

[0082] Figure 5A flowchart of a method for inputting a first historical protein pair in the first data set into a local feature extraction model to obtain a first prediction result between the first historical proteins in the first historical protein pair according to an exemplary embodiment of the present disclosure is schematically shown.

[0083] Figure 6 A scene example diagram schematically illustrates a calculation process of a first feature vector according to an example embodiment of the present disclosure.

[0084] Figure 7 An example diagram of an original feature map according to an example embodiment of the present disclosure is schematically shown.

[0085] Figure 8 A diagram schematically shows an example of a calculation scenario of a second prediction result according to an example embodiment of the present disclosure.

[0086] Fig. 9 The flowchart schematically shows a method for predicting an interaction relationship according to an exemplary embodiment of the present disclosure.

[0087] Fig.10 A block diagram schematically illustrates a training device for a relationship prediction model according to an example embodiment of the present disclosure.

[0088] Fig.11 A block diagram schematically illustrates a device for predicting an interaction relationship according to an exemplary embodiment of the present disclosure.

[0089] Fig.12 An electronic device for implementing the above-mentioned relationship prediction model training method and / or interaction relationship prediction method according to an exemplary embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0090] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as being limited to the examples set forth herein; on the contrary, these embodiments are provided so that the present disclosure will be more comprehensive and complete, and the concepts of the example embodiments are fully conveyed to those skilled in the art. The described features, structures, or characteristics may be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present disclosure. However, those skilled in the art will appreciate that the technical solutions of the present disclosure may be practiced while omitting one or more of the specific details, or other methods, components, devices, steps, etc. may be adopted. In other cases, known technical solutions are not shown or described in detail to avoid obscuring various aspects of the present disclosure.

[0091] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.

[0092] Protein interaction studies can reveal the functions of proteins at the molecular level and help reveal the laws of cell activities such as growth, development, metabolism, differentiation and apoptosis. Identifying protein interaction pairs on a genome-wide scale is an important step in explaining cell regulatory mechanisms. At the same time, with the development of protein interaction experimental technology, people can obtain a large amount of protein interaction data and even analyze protein interactions on a genome-wide scale. However, due to the limitations of experimental technology, traditional experimental methods are not suitable for large-scale detection. In order to solve this problem, existing protein interaction data can be used to achieve the purpose of large-scale detection through machine learning methods.

[0093] In practical applications, the traditional method of analyzing and detecting protein interactions through machine learning can be achieved in the following two ways:

[0094] One is that the amino acid sequence of a protein can be represented as an amino acid sequence vector, and then the interaction between proteins can be predicted in the vector space of the amino acid sequence vector; wherein the amino acid sequence vector can be, for example, an amino acid k-mer vector; wherein the k-mer vector recorded here can be a small fragment in the amino acid sequence, and k is the length of the small fragment in the amino acid sequence; for example, for the amino acid sequence MTAQDD…SYS, it can be represented as a 3-mer vector [MTA, QDD,…,SYS]; at the same time, each dimension of the vector represents a different small amino acid fragment, and the vector elements are independent of each other, that is, the relationship between each small amino acid fragment can be ignored.

[0095] Another method is to predict through protein-protein interaction networks; in the protein-protein interaction network, vertices represent proteins, and edges represent interactions between corresponding proteins. The topological structure of the graph is then used through graph neural networks or random walks to complete the prediction task.

[0096] However, both of the above two implementation methods have their own shortcomings. The first implementation method uses the local information of the protein and ignores the global information. The second implementation method uses the global information and ignores the local information. Both methods have the problem of low accuracy of prediction results.

[0097] Based on this, the example embodiments of the present disclosure first provide a training method for a relationship prediction model. Different from traditional methods, the present disclosure comprehensively utilizes the global information provided by the topological structure of the protein interaction network and the local information provided by the amino acid sequence of the protein to predict the interaction between proteins, thereby improving the accuracy of the prediction.

[0098] In an exemplary embodiment, the training method of the relationship prediction model can be run on a server, a server cluster or a cloud server, etc. Of course, those skilled in the art can also run the method disclosed in this disclosure on other platforms as required, and this exemplary embodiment does not specifically limit this. Figure 1 As shown, the training method of the relationship prediction model may include the following steps:

[0099] Step S110. extracting first historical proteins with correlation and second historical proteins without correlation in the historical biological data, and constructing a first data set according to the first historical proteins and the actual correlation values ​​between the first historical proteins;

[0100] Step S120. Training the local feature extraction model according to the first data set to obtain a first interaction relationship prediction model, and inputting the second historical protein into the first interaction relationship prediction model to obtain a first predicted association relationship value;

[0101] Step S130. constructing a second data set according to the second historical protein and the first predicted association relationship value, and training a global feature extraction model according to the first data set and the second data set to obtain a second interaction relationship prediction model;

[0102] Step S140: Construct a target relationship prediction model based on the first interaction relationship prediction model and the second interaction relationship prediction model.

[0103] In the training method of the above-mentioned relationship prediction model, on the one hand, a first historical protein with an associated relationship and a second historical protein without an associated relationship are extracted from the historical biological data, and a first data set is constructed based on the actual association relationship value between the first historical protein and the first historical protein; then the local feature extraction model is trained based on the first data set to obtain a first interaction relationship prediction model, and the second historical protein is input into the first interaction relationship prediction model to obtain a first predicted association relationship value; then a second data set is constructed based on the second historical protein and the first predicted association relationship value, and a global feature extraction model is trained based on the first data set and the second data set to obtain a second interaction relationship prediction model; finally, a target relationship prediction model is constructed based on the first interaction relationship prediction model and the second interaction relationship prediction model; since the global and local features of the protein are taken into account at the same time during the model training process, the problem that only the local information of the protein is used and the global information is ignored in the prior art, thereby solving the problem that the model has low accuracy; on the other hand, the accuracy of the prediction results obtained by relationship prediction through the target relationship prediction model is also improved.

[0104] Hereinafter, the training method of the relationship prediction model described in the exemplary embodiment of the present disclosure will be further explained and illustrated in conjunction with the accompanying drawings.

[0105] First, the technical implementation principle of the exemplary embodiment of the present disclosure is explained and illustrated. Specifically, the training method of the relationship prediction model recorded in the exemplary embodiment of the present disclosure can predict protein-protein interactions based on topological structure and sequence content, which can avoid the problem of low accuracy of the prediction results of the interaction relationship caused by using sequence content to represent proteins as vectors in traditional methods, and then predicting protein-protein interactions in vector space; or using protein-protein interaction networks to predict. The training method of the relationship prediction model provided by the present disclosure can comprehensively utilize the global information provided by the topological structure of the protein interaction network and the local information provided by the amino acid sequence of the protein to predict the interaction between proteins, thereby improving the accuracy of the prediction.

[0106] Secondly, the local feature extraction model and the global feature extraction model involved in the exemplary embodiments of the present disclosure are explained and illustrated.

[0107] In an exemplary embodiment, referring to Figure 2As shown, the local feature extraction model may include a first input layer 210, a first feature encoding layer 220, a first feature mapping layer 230, a first feature sequence extraction layer 240, a first classification layer 250 and a first output layer 260; wherein the first input layer, the first feature encoding layer, the first feature mapping layer, the first feature sequence extraction layer, the first classification layer and the first output layer are connected in sequence; at the same time, the specific use of each functional layer in the feature extraction process will be explained and illustrated in detail later, and will not be repeated here.

[0108] In an exemplary embodiment, referring to Figure 3 As shown, the global feature extraction model may include a second input layer 310, a graph convolutional neural network 320, a second classification layer 330 and a second output layer 340; wherein the second input layer, the graph convolutional neural network, the second classification layer and the second output layer are connected in sequence; at the same time, the specific use of each functional layer in the feature extraction process will be explained and illustrated in detail later, and will not be repeated here.

[0109] In an exemplary embodiment, referring to Figure 4 As shown, the target relationship prediction model recorded in the exemplary embodiment of the present disclosure may include a first interaction relationship prediction model 410 obtained by training a local feature extraction model, and a second interaction relationship prediction model 420 obtained by training a global feature extraction model, and also includes a weighted summation layer 430 and a third output layer 440; wherein the first interaction relationship prediction model and the second interaction relationship prediction model are respectively connected to the weighted summation layer, and the weighted summation layer is connected to the third output layer; at the same time, the specific use of each functional layer in the feature extraction process will be explained and illustrated in detail later, and will not be repeated here.

[0110] The following will be combined Figure 2-Figure 4 right Figure 1 The training method of the relationship prediction model shown in is further explained and illustrated. Specifically:

[0111] In step S110, first historical proteins with correlation and second historical proteins without correlation are extracted from the historical biological data, and a first data set is constructed according to the first historical proteins and the actual correlation values ​​between the first historical proteins.

[0112] In this example embodiment, first, historical biological data can be obtained from a corresponding database or cluster; wherein the historical biological data recorded herein can include biological data of a corresponding patient, and the biological data can include multiple groups of different protein data; in the historical biological data, the relationship between each historical protein can be obtained according to the specific condition of the patient, and then from the historical biological data, the historical proteins with associated relationship can be classified as first historical proteins, and the historical proteins without associated relationship can be classified as second historical proteins.

[0113] Secondly, a first data set is constructed based on the first historical protein and the actual correlation value between the first historical proteins. Specifically, this can be achieved in the following way: a first historical protein pair is formed based on the first historical proteins with correlation, and a first data set is constructed based on the first historical protein pair and the actual correlation value between the first historical proteins in the first historical protein pair. That is, in the actual application process, the specific presentation form of the obtained first data set can be as follows: {(p1, p2, y1), (p3, p4, y3), ..., (pN, pN+1, yN)}, and the first data set can be labeled as D; at the same time, in the first data set, p1 and p2 can be used to characterize the first historical protein pair, and y1 can be used to characterize the actual correlation value between p1 and p2, and the actual correlation value can be obtained from historical biological data.

[0114] In an exemplary embodiment, each protein can be represented by an amino acid sequence pair; for example, the amino acid sequence pair of p1 can be MTAQDD…SYS, and the amino acid sequence pair of p2 can be EAELCP…DRC; in actual application, it is necessary to use a target relationship prediction model obtained through training to predict the interaction relationship between two proteins whose relationship is unknown. It can also be considered as predicting the interaction relationship between two amino acid sequence pairs whose relationship is unknown through the target relationship prediction model.

[0115] In step S120, the local feature extraction model is trained according to the first data set to obtain a first interaction relationship prediction model, and the second historical protein is input into the first interaction relationship prediction model to obtain a first predicted association relationship value.

[0116] In this example embodiment, first, the local feature extraction model is trained according to the first data set to obtain a first interaction relationship prediction model; specifically, it can be achieved in the following way: first, the first historical protein pair in the first data set is input into the local feature extraction model to obtain a first prediction result between the first historical proteins in the first historical protein pair; secondly, according to the first prediction result and the actual correlation relationship value, a first target loss function is constructed, and the first model parameters in the local feature extraction model are adjusted based on the first target loss function to obtain a first interaction relationship prediction model; wherein, the first historical protein pair recorded here may include a first sub-protein and a second sub-protein, such as p1 and p2 mentioned above, or p3 and p4, etc., and this example does not impose any special restrictions on this.

[0117] In an exemplary embodiment, referring to Figure 5 As shown, inputting the first historical protein pair in the first data set into the local feature extraction model to obtain a first prediction result between the first historical proteins in the first historical protein pair may include the following steps:

[0118] Step S510, obtain the first amino acid sequence of the first sub-protein and the second amino acid sequence of the second sub-protein in the first historical protein pair in the first data set, and construct a first amino acid vector and a second amino acid vector according to the first amino acid fragment in the first amino acid sequence and the second amino acid fragment in the second amino acid sequence.

[0119] Specifically, assume that the first amino acid sequence of the first subprotein p1 in the first historical protein pair is MTAQDD…SYS, and the second amino acid sequence of the second subprotein p2 is EAELCP…DRC; at the same time, the first amino acid fragment in the first amino acid sequence may include MTA, QDD,…, SYS, etc., and the second amino acid fragment in the second amino acid sequence may include EAE, LCP,…, DRC, etc.; at the same time, the obtained first amino acid vector may be {MTA, QDD,…, SYS}, and the second amino acid vector may be {EAE, LCP,…, DRC}. It should be noted here that the first amino acid vector and the second amino acid vector recorded here can be considered as 3-mer vectors, and 3 is the length of the amino acid fragment; in the actual application process, the first amino acid vector and the second amino acid vector can also be represented as k-mer vectors, and the value of k can be selected according to actual needs, and this example does not impose any special restrictions on this.

[0120] Step S520, using the first feature encoding layer to perform one-hot encoding on the first amino acid vector and the second amino acid vector to obtain a first encoding result and a second encoding result, and using the first feature mapping layer to perform dense mapping on the first encoding result and the second encoding result to obtain a first dense mapping result and a second dense mapping result.

[0121] Specifically, the first feature encoding layer recorded here can be a one-hot encoding layer, and the first feature mapping layer recorded here can be an embedding layer; in the actual application process, since there are 20 kinds of amino acids that make up proteins, there are 203 kinds (8000 kinds) of 3-mer vectors of amino acids, so 8000-dimensional one-hot vectors are used here to represent the 3-mer of protein amino acids; that is, the vector dimension of the first encoding result and the second encoding result is 8000 dimensions; then, the 3-mer one-hot of amino acids (that is, the first encoding result and the second encoding result) is mapped into a dense embedding vector (that is, the first dense mapping result and the second dense mapping result) through the mapping matrix W1. Among them, the specific calculation process of the first dense mapping result and the second dense mapping result can be shown in the following formula (1):

[0122]

[0123] in, Represents the one-hot vector of the j-th 3-mer of the amino acid sequence, W1 is a matrix of size 256*8000; represents the embedding vector of the j-th 3-mer of the amino acid sequence; that is, in the first encoding result and the second encoding result, the encoding vector of each 3-mer fragment is independent; at the same time, in the first dense mapping result and the second dense mapping result, the embedding vector of each 3-mer is also independent.

[0124] Step S530, using the first feature sequence extraction layer to perform temporal feature extraction on the first dense mapping result and the second dense mapping result to obtain a first feature vector of the first sub-protein based on the first amino acid sequence and a second feature vector of the second sub-protein based on the second amino acid sequence.

[0125] Specifically, the first feature sequence extraction layer recorded here may include multiple feature sequence prediction networks and a first splicing layer, and the number of the feature sequence prediction networks is consistent with the number of first amino acid fragments in the first amino acid vector; that is, each 3-mer embedding vector may correspond to a feature sequence prediction network; at the same time, the feature sequence prediction network recorded here may be a long short-term memory network (Long Short-Term Memory, LSTM). p 1 h p 2 h p M .

[0126] In the actual application process, the first feature sequence extraction layer is used to extract the time series features of the first dense mapping result to obtain the first feature vector of the first sub-protein based on the first amino acid sequence. This can be achieved in the following way: First, according to the fragment position of the first amino acid fragment in the first amino acid vector, the first sub-mapping result corresponding to each first amino acid fragment in the first dense mapping result is input into the feature sequence prediction network corresponding to the fragment position to obtain the first sub-sequence vector; secondly, the first weight value of the first sub-sequence vector is calculated, and the first sub-sequence vector and the first weight value are weighted and summed through the first concatenation layer to obtain the first feature vector of the first sub-protein based on the first amino acid sequence. That is, the embedding vector (the first dense mapping result and the second dense mapping result) can be used as the input of the LSTM, and the output obtained is the vector corresponding to each 3-mer. (i.e., the first subsequence vector, whose dimension size can be 128 dimensions); among them, the specific calculation scenario example diagram can be referred to Figure 6 As shown; further, assuming that the vector representation of the entire protein sequence is composed of a linear combination of the amino acid 3-mer vectors that constitute the protein with different weights, the specific calculation process of the first eigenvector can be shown in the following formula (2):

[0127]

[0128] Among them, a i represents the weight value of the ith amino acid 3-mer (i.e., the first weight value). Meanwhile, the specific calculation process of the first weight value can be shown in the following formula (3):

[0129]

[0130] Among them, θ is a 128-dimensional constant parameter vector.

[0131] So far, the specific calculation process of the first eigenvector has been fully realized; meanwhile, the specific calculation process of the second eigenvector is similar to that of the first eigenvector, and will not be further described here.

[0132] Step S540: input the first feature vector and the second feature vector into a first classification layer to obtain a first prediction result between the first sub-protein and the second sub-protein.

[0133] Specifically, in the actual application process, assuming that proteins p1 and p2 are given, the sequence model module can be used to obtain vector representations s1 (i.e., the first eigenvector) and s2 (i.e., the second eigenvector) of p1 and p2 based on the amino acid sequence, and then the first classification layer (also referred to as the first classifier) ​​is used to predict the relationship between p1 and p2; wherein, the first classification layer C1 can be implemented by the following formula (3) in the process of classifying s1 and s2:

[0134] y1=σ(W s [s1,s2]+b s ); Formula (3)

[0135] Where σ is the sigmoid function, [s1,s2] means concatenating s1 and s2; W s and b s Indicates the first classification parameter corresponding to the first classification layer.

[0136] At this point, the specific calculation process of the first prediction result has been fully realized. At the same time, after obtaining the first prediction result, the first target loss function can be constructed based on the first prediction result and the actual correlation relationship value; wherein, the first target loss function recorded here can use the minimization loss function corresponding to the stochastic gradient descent method; wherein, the specific presentation form of the first target loss function can be shown in the following formula (4):

[0137] l1=-∑ i y i logy i '; Formula (4)

[0138] Among them, i=1,3,5,…,N,y i Indicates the actual association value, y i ' is the interaction relationship between pi and pi+1 predicted by classifier C1, that is, the first prediction result.

[0139] Furthermore, after obtaining the first target loss function, the first model parameters in the local feature extraction model can be adjusted based on the first target loss function to obtain a first interaction relationship prediction model; wherein the first model parameters recorded here may include a first encoding parameter corresponding to the first feature encoding layer, a first mapping parameter corresponding to the first feature mapping layer, a first feature extraction parameter corresponding to the first feature sequence extraction layer, and a first classification parameter corresponding to the first classification layer, etc.; it can be known that in the process of adjusting the first model parameters in the local feature extraction model based on the first target loss function to obtain the first interaction relationship prediction model, it can be achieved in the following way: based on the first target loss function, the first encoding parameter, the first mapping parameter, the first feature extraction parameter and the first classification parameter in the local feature extraction model are adjusted to obtain the first interaction relationship prediction model.

[0140] At this point, the specific training process of the first interaction relationship prediction model has been fully realized. After obtaining the first interaction relationship prediction model, the first predicted association relationship value between the second historical proteins that do not have an association relationship can be predicted based on the first interaction relationship prediction model; wherein, the specific calculation process of the first predicted association relationship value is similar to the specific calculation process of the first prediction result, and no further elaboration is given here; at the same time, in the calculation process of the first predicted association relationship value, it is first necessary to construct a second protein pair based on the second historical protein; wherein, the second protein pair may include: {(up1, up2),…,(upK, upK+1)} and so on; after obtaining the second protein pair, the first predicted association relationship value usyi between (up1, up2),…,(upK, upK+1) can be calculated based on the first interaction relationship prediction model.

[0141] In step S130, a second data set is constructed according to the second historical protein and the first predicted association relationship value, and a global feature extraction model is trained according to the first data set and the second data set to obtain a second interaction relationship prediction model.

[0142] In this example embodiment, first, a second data set is constructed according to the second historical protein and the first predicted association relationship value. Specifically, it can be implemented in the following manner: first, determine whether the first predicted association relationship value between the second historical proteins is greater than a preset threshold; second, when it is determined that the first predicted association relationship value is greater than the preset threshold, form a second historical protein pair according to the second historical protein corresponding to the first predicted association relationship value; then, construct a second data set according to the second historical protein pair and the first predicted association relationship value. Among them, the obtained second data set may include: {(up1, up2, usy1), ..., (upK, upK+1, usyK)}, usyi (i = 1, 3, ..., K) is the predicted label, that is, the first predicted association relationship value; at the same time, in the second data set, it is necessary to select the second historical protein pair with usyi ≥ 0.8 to construct the second data set; the obtained second data set can be labeled as D1.

[0143] Secondly, after obtaining the second data set, the global feature extraction model can be trained according to the first data set and the second data set to obtain a second interaction relationship prediction model. Specifically, the specific training process of the global feature extraction model can be implemented in the following ways: first, construct an original feature map according to the first data set and the second data set, and input the original feature map into the global feature extraction model to obtain a second prediction result; secondly, construct a second target loss function according to the second prediction result and the actual association relationship value and / or the first predicted association relationship value, and adjust the second model parameters in the global feature extraction model based on the second target loss function to obtain a second interaction relationship prediction model.

[0144] In an exemplary embodiment, constructing the original feature graph based on the first data set and the second data set can be achieved in the following manner: first, obtaining the actual correlation relationship value between the first historical protein pair and the first historical protein pair in the first data set, and the first predicted correlation relationship value between the second historical protein pair and the second historical protein pair in the second data set; secondly, taking the first sub-protein and the second sub-protein in the first historical protein pair as the first vertex and the second vertex, taking the actual correlation relationship as the first connecting edge, and taking the actual correlation relationship value as the first edge weight of the first connecting edge; then, taking the third sub-protein and the fourth sub-protein in the second historical protein pair as the third vertex and the fourth vertex, taking the first predicted correlation relationship as the second connecting edge, and taking the first predicted correlation relationship value as the second edge weight of the second connecting edge; finally, constructing the original feature graph based on the first vertex, the second vertex, the third vertex, the fourth vertex, the first connecting edge, the second connecting edge, the first edge weight and the second edge weight. That is, in the actual application process, each sub-protein can be abstracted as a node, and the correlation relationship value between each sub-protein can be abstracted as an edge, so as to construct the original feature graph. Among them, the obtained original feature graph can refer to Figure 7 shown.

[0145] In an exemplary embodiment, the original feature map is input into the global feature extraction model to obtain the second prediction result, which can be achieved in the following manner: first, the original feature map is input into the graph convolutional neural network, and the first connection edge and / or the second connection edge in the original feature map is updated to obtain the target feature map; secondly, the target feature map is input into the second classification layer to obtain the second prediction result. That is, in the actual application process, assuming that there are N1 nodes {e i ',i=1,...,N1}, M1 edges {r j ',j=1,...,M1}; at the same time, h t (e i ') indicates that in step t, node e i ' (the vector dimension is 128); under this premise, the first connection edge and / or the second connection edge in the original feature graph is updated in the graph convolutional neural network, which can be shown as follows by formula (5) and formula (6):

[0146]

[0147]

[0148] Where σ is the Leaky ReLU activation function; N(e i ') represents node e i'The set of neighbor nodes, <h t (e i '),h t (e k ')> indicates h t (e i ') and h t (e k ') inner product between; is the parameter of the graph neural network; at the same time, the attention weight Indicates the strength of the connection between nodes i and k; in this example embodiment, first, the parameters of the graph neural network and the initial vector of each node are randomly initialized; secondly, the graph convolutional neural network is iterated through the original feature map, assuming that the maximum value of t is 5, that is, the graph neural network obtains the final vector representation (that is, the target feature map) of each node (that is, protein) after 5 iterations; wherein, the target feature map recorded here can be represented by g1 and g2.

[0149] In an exemplary embodiment, after obtaining the target feature map, the target feature map can be input into the second classification layer to obtain the second prediction result; wherein the specific calculation process of the second prediction result can be implemented by the following formula (7):

[0150] y2=σ(W g [g1,g2]+b g ); Formula (7)

[0151] σ is the sigmoid function, [g1,g2] means concatenating g1 and g2; W g and b g It is used to represent the parameters of the second classification layer, that is, the second classification parameters. For specific calculation scenario example diagrams, please refer to Figure 8 shown.

[0152] At the same time, after obtaining the second prediction result, the second target loss function can be constructed according to the second prediction result and the actual association relationship value and / or the first predicted association relationship value; that is, the loss function (second target loss function) can be minimized based on D∪D1 using the stochastic gradient descent method; wherein the specific presentation form of the second target loss function can be shown in the following formula (8):

[0153] l2=-∑ i y i logy i "; Formula (8)

[0154] Where i = 1, 3, ..., L, where L = the number of elements in the training data set + the number of elements in D1, y i” is the interaction relationship between pi and pi+1 and upK and upK+1 predicted by the second classification layer C2, that is, the second prediction result; y i It can be used to represent the actual association relationship value and / or the first predicted association relationship value; at the same time, in the process of adjusting the second model parameters in the global feature extraction model through the second objective loss function, the second model parameters may include parameters corresponding to the graph convolutional neural network, and also include second classification parameters corresponding to the second classification layer.

[0155] In step S140, a target relationship prediction model is constructed according to the first interaction relationship prediction model and the second interaction relationship prediction model.

[0156] Specifically, the target relationship prediction model is constructed according to the first interaction relationship prediction model and the second interaction relationship prediction model, which can be realized in the following manner: first, the second historical protein is input into the second interaction relationship prediction model to obtain the second predicted association relationship value, and the third data set is constructed according to the second historical protein and the second predicted association relationship value; second, the local feature extraction model is trained according to the first data set and the third data set to obtain the updated first interaction relationship prediction model; then, the target relationship prediction model is constructed according to the updated first interaction relationship prediction model and the second interaction relationship prediction model. That is, in the actual application process, the second predicted association relationship value ugyi of the second historical protein can be predicted based on the second interaction relationship prediction model, and then the third data set {(up1, up2, ugy1), ..., (upK, upK + 1, ugyK)} is obtained, and the third data set can be recorded as D1; ​​at the same time, in the process of constructing the third data set, it is necessary to select the second historical protein pair corresponding to ugyi ≥ 0.8 to construct the third data set; then, the updated first interaction relationship prediction model and the second interaction relationship prediction model can be obtained by repeating the iteration in sequence, and then the target relationship prediction model is obtained.

[0157] In the following, the specific generation process of the target relationship prediction model will be further explained and illustrated in combination with specific embodiments. First, given the training data {(p1, p2, y1), (p3, p4, y3), ..., (pN, pN+1, yN)}, denoted as D, and the unlabeled data {(up1, up2), ..., (upK, upK+1)}, let Secondly, based on the training data set D∪D2, the stochastic gradient descent method is used to minimize the loss function l1=-∑ i y i logy i'(i=1,3,…,N) to learn the parameters of the embedding layer, sequence model, and classifier C1; where y i ' is the interaction relationship between pi and pi+1 predicted by classifier C1; further, the trained parameters are used to predict the unlabeled data set to obtain {(up1, up2, usy1),…, (upK, upK+1, usyK)}, where usyi (i=1,3,…,K) is the predicted label; the items with usyi≥0.8 are selected to obtain a new data set, denoted as D1; ​​further, based on D∪D1, the stochastic gradient descent method is used to minimize the loss function l2=-∑ i y i logy i ”, (i=1,3,…,L where L=the number of elements in the training data set + the number of elements in D1) to learn the parameters of the graph neural network and classifier C2; where y i ” is the interaction relationship between pi and pi+1 predicted by classifier C2; further, the trained parameters are used to predict the unlabeled data set to obtain {(up1, up2, ugy1),…, (upK, upK+1, ugyK)}, where ugyi (i = 1, 3,…, K) is the predicted label; the item with ugyi ≥ 0.8 is selected to obtain a new data set, recorded as D2; finally, the above steps are repeated until the model converges or the maximum number of cycles is reached (the specific value can be 100, or other values, and this example does not impose special restrictions on this. It should be noted here that, in the process of model training, the global and local features of the protein are taken into account at the same time, thereby solving the problem that only the local information of the protein is used and the global information is ignored in the prior art, thereby making the model less accurate; at the same time, in the process of model training, the model can be iteratively trained based on the unlabeled data, thereby achieving the improvement of the accuracy of the obtained relationship prediction model on the basis of expanding the data volume of the training data set.

[0158] So far, the training method of the relationship prediction model described in the exemplary embodiments of the present disclosure has been fully implemented.

[0159] The exemplary embodiment of the present disclosure also provides a method for predicting an interaction relationship, which can be run on a server, a server cluster, or a cloud server, etc. Of course, those skilled in the art can also run the method of the present disclosure on other platforms as required, and this exemplary embodiment does not specifically limit this. Fig. 9 As shown, the prediction method of the interaction relationship may include the following steps:

[0160] Step S910. Obtain a first protein to be predicted and a second protein to be predicted, and construct a protein pair to be predicted based on the first protein to be predicted and the second protein to be predicted;

[0161] Step S920. Input the protein pair to be predicted into a target relationship prediction model to obtain a relationship prediction result; wherein the target relationship prediction model is trained by any of the relationship prediction model training methods described above; the target relationship prediction model includes a first interaction relationship prediction model and a second interaction relationship prediction model;

[0162] Step S930: Determine the interaction relationship between the first protein to be predicted and the second protein to be predicted according to the relationship prediction result.

[0163] In an exemplary embodiment, the protein pair to be predicted is input into the target relationship prediction model to obtain the relationship prediction result, which can be achieved in the following manner: first, the protein pair to be predicted is input into the first interaction relationship prediction model to obtain the first relationship prediction result, and the protein pair to be predicted is input into the second interaction relationship prediction model to obtain the second relationship prediction result; secondly, the first relationship prediction result and the second relationship prediction result are weighted summed to obtain the relationship prediction result. That is, assuming that new protein pi (the first protein to be predicted) and protein pj (the second protein to be predicted) are given, protein pi (the first protein to be predicted) and protein pj (the second protein to be predicted) can first be input into the first interaction relationship prediction model to obtain the first relationship prediction result y1, and then protein pi (the first protein to be predicted) and protein pj (the second protein to be predicted) can be input into the second interaction relationship prediction model to obtain the second relationship prediction result y2, and finally the relationship prediction result is calculated; wherein the specific calculation process of the relationship prediction result can be implemented by the following formula (9):

[0164]

[0165] At the same time, if y>0.5, it is determined that there is an interaction relationship between pi and pj, otherwise there is no interaction relationship. It should be noted here that in the actual application process, the weights of the first relationship prediction result y1 and the second relationship prediction result y2 can be determined according to actual needs. It only needs to ensure that the sum of the weights is 1. This example does not impose any special restrictions on this.

[0166] In an example embodiment, in the process of actual application, obtaining the first protein to be predicted and the second protein to be predicted can be achieved in the following manner: in response to a touch operation on a first interactive control on a display interface, obtaining a user identifier of a user to be predicted; wherein the first interactive control includes a user identifier input box; in response to a touch operation on a second interactive control on a display interface, obtaining a first protein to be predicted associated with the user to be predicted and a second protein to be predicted associated with the first protein to be predicted based on the user identifier; wherein the second interactive control includes a relationship prediction interactive control. That is, in the scenario of disease auxiliary diagnosis and treatment or disease prediction, when a doctor needs to predict whether the patient (user to be predicted) has the possibility of suffering from a certain disease or whether the patient has suffered from a certain disease based on the first protein to be predicted of the patient (user to be predicted), the user identifier of the patient can be input based on the display interface of the hospital information system; wherein the user identifier can be the patient's user code in the hospital, or the user's ID number or telephone number, etc., and this example does not impose special restrictions on this; then, the first protein to be predicted can be obtained from the database based on the user identifier; then, if a certain disease is to be predicted, the second protein to be predicted associated with the disease can be obtained from the database.

[0167] In an exemplary embodiment, in the process of actual application, after determining the interaction relationship between the first protein to be predicted and the second protein to be predicted, the prediction method of the interaction relationship may also include: generating a relationship prediction report associated with the user to be predicted based on the interaction relationship, and displaying the relationship prediction report on the display interface. That is, the relationship prediction report can be displayed on the display interface, so that the doctor can make a comprehensive consideration of whether the patient has a certain disease based on the relationship prediction report. In this way, not only the patient's protein data (the first protein to be predicted) is fully utilized, but also the patient's disease can be comprehensively considered based on multiple different dimensions, thereby improving the accuracy of the obtained disease diagnosis and treatment results.

[0168] In the above-mentioned interaction relationship prediction method, on the one hand, since the global and local characteristics of the protein are taken into account at the same time during the model training process, the problem of low model accuracy in the prior art that only local information of the protein is used and global information is ignored is solved; on the other hand, the accuracy of the prediction results obtained by relationship prediction through the target relationship prediction model is also improved, that is, the accuracy of the obtained relationship prediction results is improved.

[0169] It should be further explained here that, in actual application, auxiliary treatment can be performed based on whether there is an interaction between the first protein to be predicted and the second protein to be predicted, such as determining whether a certain disease exists, determining whether an allergic reaction will occur, determining the severity of the symptoms of a certain disease, etc.

[0170] The following are embodiments of the device disclosed herein, which can be used to execute the method embodiments disclosed herein. For details not disclosed in the device embodiments disclosed herein, please refer to the method embodiments disclosed herein.

[0171] The exemplary embodiment of the present disclosure also provides a training device for a relationship prediction model. Specifically, refer to Fig.10 As shown, the training device of the relationship prediction model may include a first data set construction module 1010, a first model training module 1020, a second model training module 1030 and a target relationship prediction model construction module 1040. Among them:

[0172] The first data set construction module 1010 may be used to extract first historical proteins with correlation and second historical proteins without correlation in the historical biological data, and construct a first data set according to the first historical proteins and the actual correlation values ​​between the first historical proteins;

[0173] The first model training module 1020 can be used to train the local feature extraction model according to the first data set to obtain a first interaction relationship prediction model, and input the second historical protein into the first interaction relationship prediction model to obtain a first predicted association relationship value;

[0174] A second model training module 1030 can be used to construct a second data set according to the second historical protein and the first predicted association relationship value, and train a global feature extraction model according to the first data set and the second data set to obtain a second interaction relationship prediction model;

[0175] The target relationship prediction model construction module 1040 can be used to construct a target relationship prediction model according to the first interaction relationship prediction model and the second interaction relationship prediction model.

[0176] In an exemplary embodiment of the present disclosure, a first data set is constructed based on a first historical protein and an actual correlation relationship value between the first historical proteins, including: forming a first historical protein pair based on the first historical proteins with a correlation relationship, and constructing the first data set based on the first historical protein pair and the actual correlation relationship value between the first historical proteins in the first historical protein pair.

[0177] In an exemplary embodiment of the present disclosure, a local feature extraction model is trained according to a first data set to obtain a first interaction relationship prediction model, including: inputting a first historical protein pair in the first data set into the local feature extraction model to obtain a first prediction result between the first historical proteins in the first historical protein pair; constructing a first target loss function according to the first prediction result and the actual correlation relationship value, and adjusting the first model parameters in the local feature extraction model based on the first target loss function to obtain a first interaction relationship prediction model.

[0178] In an exemplary embodiment of the present disclosure, the local feature extraction model includes a first feature encoding layer, a first feature mapping layer, a first feature sequence extraction layer and a first classification layer; the first historical protein pair includes a first sub-protein and a second sub-protein; wherein the first historical protein pair in the first data set is input into the local feature extraction model to obtain a first prediction result between the first historical proteins in the first historical protein pair, including: obtaining a first amino acid sequence of the first sub-protein and a second amino acid sequence of the second sub-protein in the first historical protein pair in the first data set, and constructing a first amino acid vector according to a first amino acid fragment in the first amino acid sequence and a second amino acid fragment in the second amino acid sequence. The first subprotein comprises a protein having a first amino acid sequence and a second amino acid vector; the first feature encoding layer performs one-hot encoding on the first amino acid vector and the second amino acid vector to obtain a first encoding result and a second encoding result, and the first feature mapping layer performs dense mapping on the first encoding result and the second encoding result to obtain a first dense mapping result and a second dense mapping result; the first feature sequence extraction layer performs temporal feature extraction on the first dense mapping result and the second dense mapping result to obtain a first feature vector of the first subprotein based on the first amino acid sequence and a second feature vector of the second subprotein based on the second amino acid sequence; the first feature vector and the second feature vector are input into the first classification layer to obtain a first prediction result between the first subprotein and the second subprotein.

[0179] In an exemplary embodiment of the present disclosure, the first feature sequence extraction layer includes multiple feature sequence prediction networks and a first splicing layer, and the number of the feature sequence prediction networks is consistent with the number of first amino acid fragments in the first amino acid vector; wherein, the first feature sequence extraction layer is used to perform temporal feature extraction on the first dense mapping result to obtain a first feature vector of the first sub-protein based on the first amino acid sequence, including: according to the fragment position of the first amino acid fragment in the first amino acid vector, the first sub-mapping result corresponding to each first amino acid fragment in the first dense mapping result is input into the feature sequence prediction network corresponding to the fragment position to obtain a first sub-sequence vector; the first weight value of the first sub-sequence vector is calculated, and the first sub-sequence vector and the first weight value are weighted summed through the first splicing layer to obtain the first feature vector of the first sub-protein based on the first amino acid sequence.

[0180] In an exemplary embodiment of the present disclosure, the first model parameters include a first encoding parameter corresponding to a first feature encoding layer, a first mapping parameter corresponding to a first feature mapping layer, a first feature extraction parameter corresponding to a first feature sequence extraction layer, and a first classification parameter corresponding to a first classification layer; the first model parameters in the local feature extraction model are adjusted based on a first target loss function to obtain a first interaction relationship prediction model, including: adjusting the first encoding parameter, the first mapping parameter, the first feature extraction parameter and the first classification parameter in the local feature extraction model based on the first target loss function to obtain a first interaction relationship prediction model.

[0181] In an exemplary embodiment of the present disclosure, a second data set is constructed based on the second historical proteins and the first predicted association relationship value, including: determining whether the first predicted association relationship value between the second historical proteins is greater than a preset threshold; when it is determined that the first predicted association relationship value is greater than the preset threshold, forming a second historical protein pair based on the second historical proteins corresponding to the first predicted association relationship value; and constructing a second data set based on the second historical protein pairs and the first predicted association relationship value.

[0182] In an exemplary embodiment of the present disclosure, a global feature extraction model is trained according to a first data set and a second data set to obtain a second interaction relationship prediction model, including: constructing an original feature map according to the first data set and the second data set, and inputting the original feature map into the global feature extraction model to obtain a second prediction result; constructing a second target loss function according to the second prediction result and the actual association relationship value and / or the first predicted association relationship value, and adjusting the second model parameters in the global feature extraction model based on the second target loss function to obtain a second interaction relationship prediction model.

[0183] In an exemplary embodiment of the present disclosure, an original feature graph is constructed based on the first data set and the second data set, including: obtaining the first historical protein pair and the actual association relationship value between the first historical protein pair in the first data set, and the second historical protein pair and the first predicted association relationship value between the second historical protein pair in the second data set; taking the first sub-protein and the second sub-protein in the first historical protein pair as the first vertex and the second vertex, the actual association relationship as the first connecting edge, and the actual association relationship value as the first edge weight of the first connecting edge; taking the third sub-protein and the fourth sub-protein in the second historical protein pair as the third vertex and the fourth vertex, the first predicted association relationship as the second connecting edge, and the first predicted association relationship value as the second edge weight of the second connecting edge; constructing the original feature graph based on the first vertex, the second vertex, the third vertex, the fourth vertex, the first connecting edge, the second connecting edge, the first edge weight and the second edge weight.

[0184] In an exemplary embodiment of the present disclosure, the global feature extraction model includes a graph convolutional neural network and a second classification layer; wherein, the original feature map is input into the global feature extraction model to obtain a second prediction result, including: inputting the original feature map into the graph convolutional neural network, updating the first connecting edge and / or the second connecting edge in the original feature map to obtain a target feature map; inputting the target feature map into the second classification layer to obtain the second prediction result.

[0185] In an exemplary embodiment of the present disclosure, a target relationship prediction model is constructed based on the first interaction relationship prediction model and the second interaction relationship prediction model, including: inputting the second historical protein into the second interaction relationship prediction model to obtain a second predicted association relationship value, and constructing a third data set based on the second historical protein and the second predicted association relationship value; training a local feature extraction model based on the first data set and the third data set to obtain an updated first interaction relationship prediction model; and constructing a target relationship prediction model based on the updated first interaction relationship prediction model and the second interaction relationship prediction model.

[0186] The exemplary embodiment of the present disclosure also provides a prediction device for interaction relationships. Specifically, refer to Fig.11 As shown, the prediction device for interaction relationship may include a to-be-predicted protein pair construction module 1110, a relationship prediction result determination module 1120, and an interaction relationship determination module 1130. Among them:

[0187] The module 1110 for constructing a protein pair to be predicted may be used to obtain a first protein to be predicted and a second protein to be predicted, and construct a protein pair to be predicted according to the first protein to be predicted and the second protein to be predicted;

[0188] The relationship prediction result determination module 1120 can be used to input the protein pair to be predicted into a target relationship prediction model to obtain a relationship prediction result; wherein the target relationship prediction model is trained by any of the relationship prediction model training methods described above;

[0189] The interaction relationship determination module 1130 may be used to determine the interaction relationship between the first protein to be predicted and the second protein to be predicted according to the relationship prediction result.

[0190] In an exemplary embodiment of the present disclosure, the target relationship prediction model includes a first interaction relationship prediction model and a second interaction relationship prediction model; wherein, the protein pair to be predicted is input into the target relationship prediction model to obtain a relationship prediction result, including: inputting the protein pair to be predicted into the first interaction relationship prediction model to obtain a first relationship prediction result, and inputting the protein pair to be predicted into the second interaction relationship prediction model to obtain a second relationship prediction result; performing weighted summation of the first relationship prediction result and the second relationship prediction result to obtain the relationship prediction result.

[0191] In an exemplary embodiment of the present disclosure, obtaining a first protein to be predicted and a second protein to be predicted includes: in response to a touch operation on a first interactive control on a display interface, obtaining a user identifier of a user to be predicted; wherein the first interactive control includes a user identifier input box; in response to a touch operation on a second interactive control on the display interface, obtaining a first protein to be predicted associated with the user to be predicted and a second protein to be predicted associated with the first protein to be predicted based on the user identifier; wherein the second interactive control includes a relationship prediction interactive control.

[0192] In an exemplary embodiment of the present disclosure, the prediction device for the interaction relationship further includes:

[0193] The relationship prediction report display module can be used to generate a relationship prediction report associated with the user to be predicted based on the interaction relationship, and display the relationship prediction report on the display interface.

[0194] The specific details of each module in the above-mentioned training device for relationship prediction model and the prediction device for interaction relationship have been described in detail in the corresponding training method for relationship prediction model and the prediction method for interaction relationship, so they will not be repeated here.

[0195] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be embodied.

[0196] In addition, although the steps of the method in the present disclosure are described in a specific order in the drawings, this does not require or imply that the steps must be performed in this specific order, or that all the steps shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps, etc.

[0197] In an exemplary embodiment of the present disclosure, an electronic device capable of implementing the above method is also provided.

[0198] Those skilled in the art will appreciate that various aspects of the present disclosure may be implemented as systems, methods or program products. Therefore, various aspects of the present disclosure may be specifically implemented in the following forms, namely: complete hardware implementation, complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to herein as "circuits", "modules" or "systems".

[0199] Refer to the following Fig.12 1200 according to this embodiment of the present disclosure is described. Fig.12 The electronic device 1200 shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present disclosure.

[0200] like Fig.12 As shown, the electronic device 1200 is in the form of a general computing device. The components of the electronic device 1200 may include, but are not limited to: the at least one processing unit 1210, the at least one storage unit 1220, a bus 1230 connecting different system components (including the storage unit 1220 and the processing unit 1210), and a display unit 1240.

[0201] The storage unit stores a program code, which can be executed by the processing unit 1210, so that the processing unit 1210 performs the steps described in the "Exemplary Method" section of the present specification according to various exemplary embodiments of the present disclosure. For example, the processing unit 1210 can perform the following steps: Figure 1 Step S110 shown in: extracting a first historical protein with an associated relationship and a second historical protein without an associated relationship in the historical biological data, and constructing a first data set based on the actual association value between the first historical protein and the first historical protein; step S120: training a local feature extraction model based on the first data set to obtain a first interaction relationship prediction model, and inputting the second historical protein into the first interaction relationship prediction model to obtain a first predicted association relationship value; step S130: constructing a second data set based on the second historical protein and the first predicted association relationship value, and training a global feature extraction model based on the first data set and the second data set to obtain a second interaction relationship prediction model; step S140: constructing a target relationship prediction model based on the first interaction relationship prediction model and the second interaction relationship prediction model.

[0202] For another example, the processing unit 1210 may execute the following Fig. 9 Step S910 shown in: obtaining a first protein to be predicted and a second protein to be predicted, and constructing a protein pair to be predicted based on the first protein to be predicted and the second protein to be predicted; step S920: inputting the protein pair to be predicted into a target relationship prediction model to obtain a relationship prediction result; wherein the target relationship prediction model is trained by any of the training methods of the relationship prediction models described above; step S930: determining the interaction relationship between the first protein to be predicted and the second protein to be predicted based on the relationship prediction result.

[0203] The storage unit 1220 may include a readable medium in the form of a volatile storage unit, such as a random access storage unit (RAM) 12201 and / or a cache storage unit 12202 , and may further include a read-only storage unit (ROM) 12203 .

[0204] The storage unit 1220 may also include a program / utility 12204 having a set (at least one) of program modules 12205, such program modules 12205 including but not limited to: an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment.

[0205] Bus 1230 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.

[0206] The electronic device 1200 may also communicate with one or more external devices 1300 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 1200, and / or communicate with any device that enables the electronic device 1200 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 1250. In addition, the electronic device 1200 may also communicate with one or more networks (e.g., local area networks (LANs), wide area networks (WANs), and / or public networks, such as the Internet) via a network adapter 1260. As shown, the network adapter 1260 communicates with other modules of the electronic device 1200 via a bus 1230. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 1200, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0207] Through the description of the above implementation, it is easy for those skilled in the art to understand that the example implementation described here can be implemented by software, or by software combined with necessary hardware. Therefore, the technical solution according to the implementation of the present disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, including several instructions to enable a computing device (which can be a personal computer, a server, a terminal device, or a network device, etc.) to execute the method according to the implementation of the present disclosure.

[0208] In an exemplary embodiment of the present disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the above method of the present specification is stored. In some possible implementations, various aspects of the present disclosure may also be implemented in the form of a program product, which includes a program code, and when the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps according to various exemplary implementations of the present disclosure described in the above "Exemplary Method" section of the present specification.

[0209] refer to Fig.10As shown, a program product 800 for implementing the above method according to an embodiment of the present disclosure is described, which can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited thereto. In this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, an apparatus or a device.

[0210] The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0211] Computer readable signal media may include data signals propagated in baseband or as part of a carrier wave, in which readable program code is carried. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Readable signal media may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0212] The program code embodied on the readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the foregoing.

[0213] Program code for performing the operations of the present disclosure may be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java, C++, etc., and conventional procedural programming languages ​​such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., through the Internet using an Internet service provider).

[0214] In addition, the above-mentioned figures are only schematic illustrations of the processes included in the method according to the exemplary embodiments of the present disclosure, and are not intended to be limiting. It is easy to understand that the processes shown in the above-mentioned figures do not indicate or limit the time sequence of these processes. In addition, it is also easy to understand that these processes can be performed synchronously or asynchronously, for example, in multiple modules.

[0215] Other embodiments of the present disclosure will be readily apparent to those skilled in the art after considering the specification and practicing the inventions invented herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not invented by the present disclosure. The specification and examples are to be considered merely exemplary, and the true scope and spirit of the present disclosure are indicated by the claims.

Claims

1. A training method for a relationship prediction model, characterized in that: include: Extracting first historical proteins with correlation and second historical proteins without correlation in the historical biological data, and constructing a first data set according to the first historical proteins and the actual correlation values ​​between the first historical proteins; Training a local feature extraction model according to the first data set to obtain a first interaction relationship prediction model, and inputting the second historical protein into the first interaction relationship prediction model to obtain a first predicted association relationship value; constructing a second data set according to the second historical protein and the first predicted association relationship value, and training a global feature extraction model according to the first data set and the second data set to obtain a second interaction relationship prediction model; A target relationship prediction model is constructed based on the first interaction relationship prediction model and the second interaction relationship prediction model.

2. The training method of the relationship prediction model according to claim 1, characterized in that: According to the first historical protein and the actual association relationship values ​​between the first historical proteins, a first data set is constructed, including: A first historical protein pair is formed according to the first historical proteins having an association relationship, and a first data set is constructed according to the first historical protein pair and the actual association relationship values ​​between the first historical proteins in the first historical protein pair.

3. The training method of the relationship prediction model according to claim 1, characterized in that: The local feature extraction model is trained according to the first data set to obtain a first interaction relationship prediction model, including: Inputting the first historical protein pair in the first data set into the local feature extraction model to obtain a first prediction result between the first historical proteins in the first historical protein pair; According to the first prediction result and the actual association relationship value, a first target loss function is constructed, and based on the first target loss function, the first model parameters in the local feature extraction model are adjusted to obtain a first interaction relationship prediction model.

4. The training method of the relationship prediction model according to claim 3, characterized in that: The local feature extraction model includes a first feature encoding layer, a first feature mapping layer, a first feature sequence extraction layer and a first classification layer; the first historical protein pair includes a first sub-protein and a second sub-protein; The first historical protein pair in the first data set is input into the local feature extraction model to obtain a first prediction result between the first historical proteins in the first historical protein pair, including: Obtaining a first amino acid sequence of a first subprotein and a second amino acid sequence of a second subprotein in a first historical protein pair in a first data set, and constructing a first amino acid vector and a second amino acid vector according to a first amino acid fragment in the first amino acid sequence and a second amino acid fragment in the second amino acid sequence; Using a first feature encoding layer to perform one-hot encoding on the first amino acid vector and the second amino acid vector to obtain a first encoding result and a second encoding result, and using a first feature mapping layer to perform dense mapping on the first encoding result and the second encoding result to obtain a first dense mapping result and a second dense mapping result; Using a first feature sequence extraction layer to perform temporal feature extraction on the first dense mapping result and the second dense mapping result, to obtain a first feature vector of the first sub-protein based on the first amino acid sequence and a second feature vector of the second sub-protein based on the second amino acid sequence; The first feature vector and the second feature vector are input into a first classification layer to obtain a first prediction result between the first sub-protein and the second sub-protein.

5. The method for training a relationship prediction model according to claim 4, characterized in that: The first feature sequence extraction layer includes a plurality of feature sequence prediction networks and a first splicing layer, and the number of the feature sequence prediction networks is consistent with the number of first amino acid fragments in the first amino acid vector; The first dense mapping result is subjected to temporal feature extraction using the first feature sequence extraction layer to obtain a first feature vector of the first sub-protein based on the first amino acid sequence, including: According to the fragment position of the first amino acid fragment in the first amino acid vector, the first sub-mapping result corresponding to each first amino acid fragment in the first dense mapping result is input into the feature sequence prediction network corresponding to the fragment position to obtain a first sub-sequence vector; A first weight value of the first subsequence vector is calculated, and the first subsequence vector and the first weight value are weighted and summed through a first concatenation layer to obtain a first feature vector of the first subprotein based on the first amino acid sequence.

6. The method for training a relationship prediction model according to claim 4, characterized in that: The first model parameters include a first encoding parameter corresponding to the first feature encoding layer, a first mapping parameter corresponding to the first feature mapping layer, a first feature extraction parameter corresponding to the first feature sequence extraction layer, and a first classification parameter corresponding to the first classification layer; The first model parameter in the local feature extraction model is adjusted based on the first objective loss function to obtain a first interaction relationship prediction model, including: Based on the first objective loss function, the first encoding parameter, the first mapping parameter, the first feature extraction parameter and the first classification parameter in the local feature extraction model are adjusted to obtain a first interaction relationship prediction model.

7. The method for training a relationship prediction model according to claim 1, characterized in that: Constructing a second data set according to the second historical protein and the first predicted association value includes: Determining whether the first predicted association relationship value between the second historical proteins is greater than a preset threshold; When it is determined that the first predicted association value is greater than the preset threshold, forming a second historical protein pair according to a second historical protein corresponding to the first predicted association value; A second data set is constructed based on the second historical protein pairs and the first predicted association relationship values.

8. The method for training a relationship prediction model according to claim 1, characterized in that: The global feature extraction model is trained according to the first data set and the second data set to obtain a second interaction relationship prediction model, including: Constructing an original feature map according to the first data set and the second data set, and inputting the original feature map into a global feature extraction model to obtain a second prediction result; A second target loss function is constructed according to the second prediction result and the actual association relationship value and / or the first predicted association relationship value, and the second model parameters in the global feature extraction model are adjusted based on the second target loss function to obtain a second interaction relationship prediction model.

9. The method for training a relationship prediction model according to claim 8, characterized in that: Constructing an original feature map according to the first data set and the second data set includes: Obtaining first historical protein pairs and actual association relationship values ​​between the first historical protein pairs in the first data set, and second historical protein pairs and first predicted association relationship values ​​between the second historical protein pairs in the second data set; The first subprotein and the second subprotein in the first historical protein pair are the first vertex and the second vertex, the actual association relationship is the first connecting edge, and the actual association relationship value is the first edge weight of the first connecting edge; The third subprotein and the fourth subprotein in the second historical protein pair are the third vertex and the fourth vertex, the first predicted association relationship is the second connecting edge, and the first predicted association relationship value is the second edge weight of the second connecting edge; An original feature graph is constructed according to the first vertex, the second vertex, the third vertex, the fourth vertex, the first connecting edge, the second connecting edge, the first edge weight, and the second edge weight.

10. The method for training a relationship prediction model according to claim 8, characterized in that: The global feature extraction model includes a graph convolutional neural network and a second classification layer; The original feature map is input into the global feature extraction model to obtain a second prediction result, including: Inputting the original feature map into the graph convolutional neural network, updating the first connecting edge and / or the second connecting edge in the original feature map to obtain a target feature map; The target feature map is input into the second classification layer to obtain the second prediction result.

11. The method for training a relationship prediction model according to claim 1, characterized in that: Constructing a target relationship prediction model according to the first interaction relationship prediction model and the second interaction relationship prediction model, including: Inputting the second historical protein into the second interaction relationship prediction model to obtain a second predicted association relationship value, and constructing a third data set based on the second historical protein and the second predicted association relationship value; Training the local feature extraction model according to the first data set and the third data set to obtain an updated first interaction relationship prediction model; A target relationship prediction model is constructed based on the updated first interaction relationship prediction model and the second interaction relationship prediction model.

12. A method for predicting an interaction relationship, characterized in that: include: Acquire a first protein to be predicted and a second protein to be predicted, and construct a protein pair to be predicted based on the first protein to be predicted and the second protein to be predicted; Inputting the protein pair to be predicted into a target relationship prediction model to obtain a relationship prediction result; wherein the target relationship prediction model is trained by the relationship prediction model training method according to any one of claims 1 to 11; According to the relationship prediction result, the interaction relationship between the first protein to be predicted and the second protein to be predicted is determined.

13. The method for predicting an interaction relationship according to claim 12, characterized in that: The target relationship prediction model includes a first interaction relationship prediction model and a second interaction relationship prediction model; The protein pair to be predicted is input into the target relationship prediction model to obtain a relationship prediction result, including: Inputting the protein pair to be predicted into a first interaction relationship prediction model to obtain a first relationship prediction result, and inputting the protein pair to be predicted into a second interaction relationship prediction model to obtain a second relationship prediction result; The first relationship prediction result and the second relationship prediction result are weightedly summed to obtain the relationship prediction result.

14. The method for predicting interaction relationship according to claim 12, characterized in that: Obtaining a first protein to be predicted and a second protein to be predicted, including: In response to a touch operation on a first interactive control on the display interface, obtaining a user identification of a user to be predicted; wherein the first interactive control includes a user identification input box; In response to a touch operation on a second interactive control on the display interface, a first protein to be predicted associated with the user to be predicted and a second protein to be predicted associated with the first protein to be predicted are obtained based on the user identifier; wherein the second interactive control includes a relationship prediction interactive control.

15. The method for predicting interaction relationship according to claim 12, characterized in that: After determining the interaction relationship between the first protein to be predicted and the second protein to be predicted, the method for predicting the interaction relationship further comprises: A relationship prediction report associated with the user to be predicted is generated based on the interaction relationship, and the relationship prediction report is displayed on the display interface.

16. A training device for a relationship prediction model, characterized in that: include: A first data set construction module is used to extract first historical proteins with correlation relationship and second historical proteins without correlation relationship in historical biological data, and construct a first data set according to the first historical proteins and the actual correlation relationship values ​​between the first historical proteins; A first model training module is used to train a local feature extraction model according to a first data set to obtain a first interaction relationship prediction model, and input the second historical protein into the first interaction relationship prediction model to obtain a first predicted association relationship value; A second model training module is used to construct a second data set according to the second historical protein and the first predicted association relationship value, and train a global feature extraction model according to the first data set and the second data set to obtain a second interaction relationship prediction model; The target relationship prediction model construction module is used to construct a target relationship prediction model according to the first interaction relationship prediction model and the second interaction relationship prediction model.

17. A prediction device for interaction relationship, characterized in that: include: A protein pair construction module to be predicted, used to obtain a first protein to be predicted and a second protein to be predicted, and to construct a protein pair to be predicted according to the first protein to be predicted and the second protein to be predicted; A relationship prediction result determination module, used for inputting the protein pair to be predicted into a target relationship prediction model to obtain a relationship prediction result; wherein the target relationship prediction model is trained by the relationship prediction model training method according to any one of claims 1 to 11; The interaction relationship determination module is used to determine the interaction relationship between the first protein to be predicted and the second protein to be predicted according to the relationship prediction result.

18. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the training method of the relationship prediction model described in any one of claims 1 to 11 and the prediction method of the interaction relationship described in any one of claims 12 to 15 are implemented.

19. An electronic device, characterized in that: include: processor; as well as A memory, configured to store executable instructions of the processor; Wherein, the processor is configured to execute the training method of the relationship prediction model described in any one of claims 1-11 and the prediction method of the interaction relationship described in any one of claims 12-15 by executing the executable instructions.

Images