RNA-protein interaction prediction method, apparatus, medium, and electronic device

Abstract

The present disclosure provides an RNA-protein interaction prediction method, device, medium and electronic equipment; it relates to artificial intelligence technical field.The method comprises: obtaining an RNA sequence and a protein sequence to be predicted; encoding the RNA sequence to be predicted to obtain an RNA first vector sequence; encoding the protein sequence to be predicted to obtain a protein first vector sequence; obtaining the correlation vector sequence of the RNA first vector sequence and the correlation vector sequence of the protein first vector sequence through an attention mechanism model; determining the interaction between the RNA sequence to be predicted and the protein sequence according to the correlation vector sequence of the RNA first vector sequence and the correlation vector sequence of the protein first vector sequence.

Description

Technical Field

[0001] This disclosure relates to the field of artificial intelligence technology, and more specifically, to an RNA-protein interaction prediction method, an RNA-protein interaction prediction device, a computer-readable storage medium, and an electronic device. Background Technology

[0002] Noncoding RNAs (ncRNAs) participate in many complex cellular processes, playing crucial roles in life processes such as alternative splicing, chromatin modification, and epigenetics, and are closely linked to many diseases. Studies have shown that most ncRNAs exert their regulatory functions through interactions with proteins. Therefore, researching the interactions between ncRNAs and proteins is of great significance for elucidating the molecular mechanisms of ncRNA action in human diseases and life activities, and has become one of the important approaches to analyzing the functions of ncRNAs and proteins.

[0003] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0004] This disclosure provides a method for predicting RNA-protein interactions, an apparatus for predicting RNA-protein interactions, a computer-readable storage medium, and an electronic device.

[0005] This disclosure provides a method for predicting RNA-protein interactions, including:

[0006] Obtain the RNA and protein sequences to be predicted;

[0007] The RNA sequence to be predicted is encoded to obtain the first RNA vector sequence;

[0008] The protein sequence to be predicted is encoded to obtain the first protein vector sequence;

[0009] The association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence are obtained through an attention mechanism model;

[0010] The interaction between the RNA sequence and the protein sequence to be predicted is determined based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence.

[0011] In one exemplary embodiment of this disclosure, encoding the RNA sequence to be predicted to obtain a first RNA vector sequence includes:

[0012] The RNA sequence to be predicted is converted into an N-base k-mer subsequence;

[0013] Vectorize each base k-mer subsequence to obtain the first vector sequence of the RNA.

[0014] In one exemplary embodiment of this disclosure, the vectorization of each base k-mer subsequence to obtain the first RNA vector sequence includes:

[0015] Encode each base k-mer subsequence to obtain a first vector of N base k-mer subsequences;

[0016] The first vector of the N base k-mer subsequences is obtained by performing operations on the first vector of the N base k-mer subsequences using the first mapping matrix;

[0017] The second vectors of the N base k-mer subsequences are sequentially input into a pre-trained recurrent neural network, which outputs N base k-mer vectors, and the N base k-mer vectors form the first RNA vector sequence.

[0018] In one exemplary embodiment of this disclosure, encoding the protein sequence to be predicted to obtain a first protein vector sequence includes:

[0019] The protein sequence to be predicted is converted into an M-amino acid k-mer subsequence;

[0020] Vectorize each amino acid k-mer subsequence to obtain the first vector sequence of the protein.

[0021] In one exemplary embodiment of this disclosure, the vectorization of each amino acid k-mer subsequence to obtain the first vector sequence of the protein includes:

[0022] Encode each amino acid k-mer subsequence to obtain a first vector of M amino acid k-mer subsequences;

[0023] The second vector of the M amino acid k-mer subsequences is obtained by operating on the first vector of the M amino acid k-mer subsequences using the second mapping matrix;

[0024] The second vectors of the M amino acid k-mer subsequences are sequentially input into a pre-trained recurrent neural network, which outputs M amino acid k-mer vectors, and the M amino acid k-mer vectors form the first vector sequence of the protein.

[0025] In one exemplary embodiment of this disclosure, obtaining the association vector sequence of the RNA first vector sequence and the association vector sequence of the protein first vector sequence through an attention mechanism model includes:

[0026] Feature extraction is performed on the first RNA vector sequence to obtain the first hidden RNA vector;

[0027] Feature extraction is performed on the first vector sequence of the protein to obtain the first hidden vector of the protein;

[0028] The first hidden vector of RNA and the first hidden vector of protein are processed to obtain the correlation vector sequence of the first vector sequence of RNA and the correlation vector sequence of the first vector sequence of protein.

[0029] In an exemplary embodiment of this disclosure, the first RNA latent vector includes an RNA first vector, an RNA second vector, and an RNA third vector; the step of extracting features from the RNA first vector sequence to obtain the first RNA latent vector includes:

[0030] The first RNA vector is obtained by performing operations on the first RNA first vector sequence using the first query weight matrix;

[0031] The first RNA vector is obtained by performing operations on the first RNA vector sequence using the first bond weight matrix;

[0032] The first RNA vector sequence is processed using the first-value weight matrix to obtain the third RNA vector.

[0033] In an exemplary embodiment of this disclosure, the first latent vector of the protein includes a first protein vector, a second protein vector, and a third protein vector; the step of extracting features from the first protein vector sequence to obtain the first latent vector of the protein includes:

[0034] The protein first vector is obtained by performing operations on the protein first vector sequence using the second query weight matrix;

[0035] The second vector of the protein is obtained by operating on the first vector sequence of the protein using the second bond weight matrix;

[0036] The third vector of the protein is obtained by operating on the first vector sequence of the protein using the second-value weight matrix.

[0037] In one exemplary embodiment of this disclosure, the step of performing operations on the first hidden RNA vector and the first hidden protein vector to obtain the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence includes:

[0038] The similarity between the first RNA vector and the second protein vector is calculated to obtain the first RNA attention score;

[0039] The third vector of the protein is summed based on the first attention score of the RNA to obtain the associated vector sequence of the first RNA vector sequence;

[0040] Calculate the similarity between the first protein vector and the second RNA vector to obtain the first protein attention score;

[0041] The RNA third vector is summed based on the protein's first attention score to obtain the associated vector sequence of the protein's first vector sequence.

[0042] In one exemplary embodiment of this disclosure, determining the interaction between the RNA sequence and the protein sequence to be predicted based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence includes:

[0043] The second RNA vector sequence and the second protein vector sequence are obtained based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence.

[0044] The interaction between the RNA sequence and the protein sequence to be predicted is determined based on the second vector sequence of the RNA and the second vector sequence of the protein.

[0045] In one exemplary embodiment of this disclosure, obtaining the second RNA vector sequence and the second protein vector sequence based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence includes:

[0046] The associated vector sequence of the first RNA vector sequence and the first RNA vector sequence are concatenated to obtain the RNA fusion vector sequence;

[0047] The RNA fusion vector sequence is input into a pre-trained recurrent neural network, which outputs the second RNA vector sequence.

[0048] The associated vector sequence of the protein first vector sequence and the protein first vector sequence are concatenated to obtain the protein fusion vector sequence;

[0049] The protein fusion vector sequence is input into a pre-trained recurrent neural network, which outputs the second protein vector sequence.

[0050] In one exemplary embodiment of this disclosure, obtaining the second RNA vector sequence and the second protein vector sequence based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence includes:

[0051] The associated vector sequence of the first RNA vector sequence and the first RNA vector sequence are concatenated to obtain the RNA fusion vector sequence;

[0052] The associated vector sequence of the protein first vector sequence and the protein first vector sequence are concatenated to obtain the protein fusion vector sequence;

[0053] The autocorrelation vector sequence of the RNA fusion vector sequence is obtained through an attention mechanism model, and the second RNA vector sequence is obtained based on the autocorrelation vector sequence of the RNA fusion vector sequence.

[0054] The autocorrelation vector sequence of the protein fusion vector sequence is obtained through an attention mechanism model, and the second vector sequence of the protein is obtained based on the autocorrelation vector sequence of the protein fusion vector sequence.

[0055] In one exemplary embodiment of this disclosure, the step of obtaining the autocorrelation vector sequence of the RNA fusion vector sequence through an attention mechanism model, and obtaining the second RNA vector sequence based on the autocorrelation vector sequence of the RNA fusion vector sequence, includes:

[0056] Feature extraction is performed on the RNA fusion vector sequence to obtain the second hidden RNA vector;

[0057] The second hidden vector of the RNA is processed to obtain the self-correlated vector sequence of the RNA fusion vector sequence;

[0058] The self-correlation vector sequence of the RNA fusion vector sequence is processed to obtain the second RNA vector sequence.

[0059] In one exemplary embodiment of this disclosure, the second RNA latent vector includes a fourth RNA vector, a fifth RNA vector, and a sixth RNA vector; the step of extracting features from the RNA fusion vector sequence to obtain the second RNA latent vector includes:

[0060] The RNA fusion vector sequence is processed using the third query weight matrix to obtain the fourth RNA vector;

[0061] The RNA fusion vector sequence is processed using the third bond weight matrix to obtain the fifth RNA vector;

[0062] The RNA fusion vector sequence is processed using a third-value weight matrix to obtain the sixth RNA vector.

[0063] In one exemplary embodiment of this disclosure, the step of performing operations on the second hidden vector of the RNA to obtain the self-correlated vector sequence of the RNA fusion vector sequence includes:

[0064] Calculate the similarity between the fourth and fifth RNA vectors to obtain the second RNA attention score.

[0065] The RNA sixth vector is summed based on the RNA second attention score to obtain the self-correlated vector sequence of the RNA fusion vector sequence.

[0066] In one exemplary embodiment of this disclosure, the step of obtaining the autocorrelation vector sequence of the protein fusion vector sequence through an attention mechanism model, and obtaining the second protein vector sequence based on the autocorrelation vector sequence of the protein fusion vector sequence, includes:

[0067] Feature extraction is performed on the protein fusion vector sequence to obtain the second hidden vector of the protein;

[0068] The second hidden vector of the protein is processed to obtain the self-correlated vector sequence of the protein fusion vector sequence;

[0069] The self-correlation vector sequence of the protein fusion vector sequence is processed to obtain the second vector sequence of the protein.

[0070] In an exemplary embodiment of this disclosure, the second hidden protein vector includes a fourth protein vector, a fifth protein vector, and a sixth protein vector; the step of extracting features from the protein fusion vector sequence to obtain the second hidden protein vector includes:

[0071] The protein fusion vector sequence is processed using the fourth query weight matrix to obtain the fourth protein vector;

[0072] The protein fusion vector sequence is processed using the fourth bond weight matrix to obtain the fifth protein vector;

[0073] The protein fusion vector sequence is processed using the fourth-value weight matrix to obtain the sixth vector of the protein.

[0074] In one exemplary embodiment of this disclosure, the step of operating on the second hidden vector of the protein to obtain the self-correlated vector sequence of the protein fusion vector sequence includes:

[0075] Calculate the similarity between the fourth and fifth vectors of the protein to obtain the second attention score of the protein;

[0076] The sixth vector of the protein is summed based on the second attention score of the protein to obtain the self-correlated vector sequence of the protein fusion vector sequence.

[0077] In one exemplary embodiment of this disclosure, determining the interaction between the RNA sequence and the protein sequence to be predicted based on the second RNA vector sequence and the second protein vector sequence includes:

[0078] The RNA second vector sequence and the protein second vector sequence are concatenated to obtain the feature vector to be predicted;

[0079] The predicted interaction values ​​between the RNA sequence and the protein sequence are obtained based on the feature vector to be predicted;

[0080] The interaction between the RNA sequence and the protein sequence to be predicted is determined based on the interaction prediction value.

[0081] In one exemplary embodiment of this disclosure, obtaining the predicted interaction value between the RNA sequence and the protein sequence based on the feature vector to be predicted includes:

[0082] The feature vector to be predicted is input into the classifier, which outputs the probability that there is an interaction between the RNA sequence and the protein sequence to be predicted.

[0083] In one exemplary embodiment of this disclosure, the method further includes:

[0084] Obtain a training dataset, which includes positive RNA-protein pairs and negative RNA-protein pairs;

[0085] The recurrent neural network and attention mechanism model are used to determine the predicted interaction values ​​for each RNA-protein pair in the training dataset;

[0086] The loss function is used to calculate the predicted interaction value and label value of each RNA-protein pair in the training dataset to obtain the corresponding loss value;

[0087] The model parameters of the recurrent neural network and attention mechanism model are adjusted based on the loss value.

[0088] In one exemplary embodiment of this disclosure, the method further includes:

[0089] Output the predicted results of the interaction between the RNA sequence and the protein sequence to be predicted.

[0090] This disclosure provides an RNA-protein interaction prediction device, comprising:

[0091] The data acquisition module is used to acquire the RNA and protein sequences to be predicted.

[0092] The data acquisition module is used to acquire the RNA and protein sequences to be predicted.

[0093] The first data encoding module is used to encode the RNA sequence to be predicted to obtain the first RNA vector sequence;

[0094] The second data encoding module is used to encode the protein sequence to be predicted to obtain the first protein vector sequence.

[0095] The association information acquisition module is used to acquire the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence through an attention mechanism model;

[0096] An interaction determination module is used to determine the interaction between the RNA sequence and the protein sequence to be predicted based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence.

[0097] This disclosure provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method described in any one of the above descriptions.

[0098] This disclosure provides an electronic device, including: a processor; and a memory for storing executable instructions of the processor; wherein the processor is configured to perform the method described in any of the preceding methods by executing the executable instructions.

[0099] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0100] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure. It is obvious that the drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0101] Figure 1 A schematic diagram of an exemplary system architecture for predicting RNA-protein interactions, to which embodiments of the present disclosure can be applied, is shown;

[0102] Figure 2 A flowchart illustrating an RNA-protein interaction prediction method according to an embodiment of the present disclosure is shown schematically.

[0103] Figure 3 A flowchart illustrating the process of obtaining a first RNA vector sequence according to an embodiment of the present disclosure is shown schematically.

[0104] Figure 4 A flowchart illustrating the process of obtaining a first protein vector sequence according to an embodiment of the present disclosure is shown schematically.

[0105] Figure 5 This schematically illustrates a flowchart of obtaining association information between RNA sequences and protein sequences according to an embodiment of the present disclosure;

[0106] Figure 6 A flowchart illustrating the process of obtaining a second RNA vector sequence and a second protein vector sequence according to an embodiment of the present disclosure is shown schematically.

[0107] Figure 7 A flowchart illustrating the process of obtaining a second RNA vector sequence according to another embodiment of this disclosure is shown schematically;

[0108] Figure 8 A flowchart illustrating the process of obtaining a second protein vector sequence according to another embodiment of the present disclosure is shown schematically;

[0109] Figure 9 A flowchart illustrating the determination of RNA sequence and protein sequence interactions according to an embodiment of the present disclosure is shown schematically.

[0110] Figure 10 A flowchart illustrating model training according to an embodiment of the present disclosure is shown schematically;

[0111] Figure 11 A flowchart illustrating an RNA-protein interaction prediction method according to another embodiment of the present disclosure is shown schematically;

[0112] Figure 12 A block diagram of an RNA-protein interaction prediction device according to an embodiment of the present disclosure is shown schematically;

[0113] Figure 13 A schematic diagram of the structure of a computer system suitable for implementing the embodiments of the present disclosure is shown. Detailed Implementation

[0114] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided to make this disclosure more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art. The described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a full understanding of embodiments of this disclosure. However, those skilled in the art will recognize that the technical solutions of this disclosure can be practiced with one or more of these specific details omitted, or other methods, components, apparatus, steps, etc., can be employed. In other instances, well-known technical solutions are not shown or described in detail to avoid obscuring various aspects of this disclosure.

[0115] Furthermore, the accompanying drawings are merely illustrative of this disclosure and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0116] Figure 1 A schematic diagram of a system architecture for an exemplary application environment in which an RNA-protein interaction prediction method and apparatus according to embodiments of the present disclosure can be applied is shown.

[0117] like Figure 1 As shown, the system architecture 100 of the interaction prediction system may include one or more of terminal devices 101, 102, and 103, a network 104, and a server 105. Network 104 serves as the medium for providing communication links between terminal devices 101, 102, and 103 and server 105. Network 104 may include various connection types, such as wired, wireless communication links, or fiber optic cables. Terminal devices 101, 102, and 103 may be various electronic devices, including but not limited to desktop computers, laptops, smartphones, and tablets. It should be understood that... Figure 1The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be used. For example, server 105 can be a single server, a server cluster consisting of multiple servers, a cloud computing platform, or a virtualization center. Specifically, server 105 can be used to perform the following: acquiring the RNA sequence and protein sequence to be predicted; encoding the RNA sequence to be predicted to obtain a first RNA vector sequence; encoding the protein sequence to be predicted to obtain a first protein vector sequence; acquiring the association vector sequences of the first RNA vector sequence and the first protein vector sequence through an attention mechanism model; and determining the interaction between the RNA sequence and the protein sequence to be predicted based on the association vector sequences of the first RNA vector sequence and the first protein vector sequence.

[0118] The RNA-protein interaction prediction method provided in this embodiment is generally executed by server 105. Correspondingly, the RNA-protein interaction prediction device is generally set in server 105. The server can send the prediction results of the interaction between the RNA sequence and the protein sequence to be predicted to the terminal device, and the terminal device can display them to the user. However, it is readily understood by those skilled in the art that the RNA-protein interaction prediction method provided in this embodiment can also be executed by one or more of terminal devices 101, 102, and 103. Correspondingly, the RNA-protein interaction prediction device can also be set in terminal devices 101, 102, and 103. For example, after execution by the terminal device, the prediction results can be directly displayed on the terminal device's screen, or the prediction results can be provided to the user through voice broadcast. This exemplary embodiment does not impose any special limitations on this.

[0119] The technical solutions of the embodiments of this disclosure are described in detail below:

[0120] This exemplary embodiment provides a method for predicting RNA-protein interactions. This method can be applied to the server 105 described above, or to one or more of the terminal devices 101, 102, and 103 described above; no special limitation is made in this exemplary embodiment. (See reference...) Figure 2 As shown, the RNA-protein interaction prediction method may include the following steps S210 to S250:

[0121] Step S210. Obtain the RNA sequence and protein sequence to be predicted;

[0122] Step S220. Encode the RNA sequence to be predicted to obtain the first RNA vector sequence;

[0123] Step S230. Encode the protein sequence to be predicted to obtain the first protein vector sequence;

[0124] Step S240. Obtain the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence through an attention mechanism model;

[0125] Step S250. Determine the interaction between the RNA sequence and the protein sequence to be predicted based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence.

[0126] In the RNA-protein interaction prediction method provided in the exemplary embodiments of this disclosure, the following steps are taken: First, an RNA sequence and a protein sequence to be predicted are obtained; the RNA sequence to be predicted is encoded to obtain a first RNA vector sequence; the protein sequence to be predicted is encoded to obtain a first protein vector sequence; an attention mechanism model is used to obtain the association vector sequences of the RNA first vector sequence and the protein first vector sequence; the interaction between the RNA sequence and the protein sequence to be predicted is determined based on the association vector sequences of the RNA first vector sequence and the protein first vector sequence. This disclosure utilizes an attention mechanism model to determine the association information between RNA and protein sequences. By fusing the RNA sequence, protein sequence, and association information, and introducing the fused sequence information to predict the interaction between RNA and protein sequences, the accuracy of RNA-protein interaction prediction can be improved.

[0127] The steps described above in this example implementation will now be explained in more detail.

[0128] In step S210, the RNA sequence and protein sequence to be predicted are obtained.

[0129] In this example embodiment, at least one RNA-protein pair to be predicted, consisting of an RNA sequence and a protein sequence, can be obtained, and the interaction between the RNA and protein sequences in each RNA-protein pair is unknown. For example, a user can input the RNA-protein pair to be predicted via a terminal device. For instance, the user can input the RNA-protein pair manually or via voice; this example does not specifically limit this. For instance, an RNA sequence can be input, followed by a protein sequence; the input order is not limited. For instance, the RNA and protein sequences can be input into different text boxes or into the same text box. For instance, after inputting the information, clicking the "Start Prediction" button will begin executing the prediction steps provided in some embodiments of this application.

[0130] The interaction between RNA and protein refers to the function of a protein manifested through its interactions with other proteins and RNA. For example, the interaction between protein and RNA plays a crucial role in protein synthesis. Simultaneously, the performance of many RNA functions depends on its interaction with proteins. Interactions can be regulatory, guiding, etc., and are not limited here. For instance, in the presence of an interaction, RNA can guide protein synthesis, or RNA can regulate protein function. The interaction between RNA and protein can also refer to their ability to regulate each other's life cycle and function through physical interactions. For example, RNA-coding sequences can guide protein synthesis, and correspondingly, proteins can regulate RNA expression and function.

[0131] After obtaining the RNA-protein pairs to be predicted, an interaction prediction system can be used to predict the interactions of each input RNA-protein pair and determine whether each pair interacts based on the prediction results. Simultaneously, the prediction results of the RNA-protein pair interactions can be output to a terminal device for user viewing. For example, the prediction results can be displayed directly on the terminal device's screen, or they can be provided to the user via voice announcement; this example does not specify any particular limitation.

[0132] In other examples, at least one RNA sequence to be predicted can be obtained, and protein sequences that interact with each input RNA sequence can be searched in a database. For instance, after a user inputs the RNA sequence to be predicted via a terminal device, they can select at least one protein sequence from the database. Multiple RNA-protein pairs are formed by the RNA sequence to be predicted and the various protein sequences. The interaction prediction system can then predict the interaction of each RNA-protein pair and output the protein sequences that can interact with the RNA sequence to be predicted based on the prediction results. Preferably, several types of protein sequences can be pre-stored in the database for easy retrieval when predicting RNA-protein pair interactions. For example, protein sequences can be stored in a Redis database or a MySQL database, allowing for real-time querying and selection of the protein sequence to be predicted. Redis is a key-value storage system; when stored in a Redis database, it can include key-value pairs formed by sequence identifiers (such as sequence numbers) and corresponding protein sequences, where the key is the sequence identifier and the value is the corresponding protein sequence. Redis, as a high-efficiency caching technology, supports read / write frequencies exceeding 100KB / s, offering advantages in data reading and storage speed. MySQL, on the other hand, is a relational database management system. Relational databases store data across different tables, rather than storing all data uniformly, increasing storage speed and flexibility. They also offer stability in data storage and can prevent data loss.

[0133] It is understandable that several types of RNA sequences can be pre-stored in a database for retrieval when predicting RNA-protein pair interactions. Therefore, at least one protein sequence to be predicted can be obtained, and RNA sequences that interact with each input protein sequence can be searched in the database. Similarly, after a user inputs a protein sequence via a terminal device, they can select at least one RNA sequence from the database, forming multiple RNA-protein pairs with the protein sequence to be predicted and various RNA sequences. The interaction prediction system can then predict the interaction of each RNA-protein pair and output the RNA sequences that can interact with the protein sequence to be predicted based on the prediction results. This disclosure does not specifically limit this approach.

[0134] In step S220, the RNA sequence to be predicted is encoded to obtain the first RNA vector sequence.

[0135] After obtaining the RNA sequence to be predicted, it can be encoded to obtain the first RNA vector sequence, thereby obtaining the association information between the first RNA vector sequence and the first protein vector sequence, and then predicting the interaction between the RNA sequence and the protein sequence based on the association information.

[0136] In the exemplary embodiments of this disclosure, RNA sequences can be represented by base sequences. For example, an RNA sequence can be represented as AGCAAGCACCU… An RNA sequence can include four bases: adenine (A), uracil (U), guanine (G), and cytosine (C). Correspondingly, RNA sequences can also be represented using base k-mer subsequences. A k-mer subsequence refers to a k-connector consisting of k bases or k amino acids as a group. Specifically, all base k-mer subsequences can be obtained by arranging and combining the four bases; for a given k value, four k-mer subsequences can be obtained. k There are 4 base k-mer subsequences. For example, when k is 3, there are 4 3 =64 base 3-mer subsequences, when k is 4, there are a total of 4 4 = 256 possible 4-mer subsequences. For example, AGC, AUA, GCA, and CCU are four different 3-mer subsequences, and AGCA, UAGC, and ACCU are three different 4-mer subsequences. Therefore, the RNA sequence AGCAUAGCACCU… can also be represented as {AGC, AUA, GCA, CCU, …}, or as {AGCA, UAGC, ACCU, …}. In other examples, the corresponding 3-mer or 4-mer subsequences can be obtained by reading the RNA sequence in an overlapping manner. Correspondingly, the 3-mer subsequences of the RNA sequence can also include AGC, GCA, CAU, AUA, etc., and the 4-mer subsequences of the RNA sequence can also include AGCA, GCAU, CAUA, etc., which are not specifically limited in this disclosure. In the example embodiments of this disclosure, k is a positive integer, such as 1, 2, 3…, and k can take one or more values. The specific value of k can be adjusted according to the actual situation and is not limited here.

[0137] When encoding the predicted RNA sequence, one can encode a portion of the bases of the predicted RNA sequence and use the encoding result as the first RNA vector sequence, or one can encode all the bases of the predicted RNA sequence and use all the encoded bases to form the first RNA vector sequence, or one can encode all the bases of the predicted RNA sequence and select a portion of the encoded bases to form the first RNA vector sequence. This disclosure does not make any specific limitations on this.

[0138] In this exemplary embodiment, the example described is encoding all bases of the RNA sequence to be predicted and forming a first RNA vector sequence from the encoded bases. The RNA sequence to be predicted can be converted into N base k-mer subsequences. For example, depending on the value of k, starting from the first base of the RNA sequence to be predicted, k consecutive bases can be taken to form a single base k-mer subsequence, until the last k bases of the RNA sequence are taken, resulting in all base k-mer subsequences of the RNA sequence to be predicted. Then, each base k-mer subsequence can be vectorized to obtain N base k-mer vectors, and these N base k-mer vectors can form the first RNA vector sequence. For example, the RNA sequence to be predicted can be divided into N non-overlapping base k-mer subsequences. For instance, if the RNA sequence to be predicted is AUCUGAAAU, it can be divided into three base k-mer subsequences: AUC, UGA, and AAU. Understandably, dividing the non-overlapping RNA sequence into multiple k-mer subsequences is to vectorize the bases in the RNA sequence as k-connectors. Similarly, in other examples, each base in the RNA sequence to be predicted can be vectorized to obtain multiple base vectors, which together form the first RNA vector sequence. Alternatively, the overlapping parts of the RNA sequence to be predicted can be divided into P k-mer subsequences, each k-mer subsequence vectorized to obtain P k-mer vectors, which together form the first RNA vector sequence. This disclosure does not impose specific limitations on this approach.

[0139] In one example implementation, after converting the RNA sequence to be predicted into N base k-mer subsequences, each base k-mer subsequence in the RNA sequence to be predicted can be encoded to obtain a first vector of N base k-mer subsequences, and the first vector of N base k-mer subsequences can be used to form the first RNA vector sequence.

[0140] In some embodiments of this disclosure, for a given k value, the base k-mer subsequence can have 4 k There are 64 possible 3-mer subsequences, each of which can be encoded using One-Hot encoding. One-Hot encoding, also known as one-bit valid encoding, uses an N-bit state register to encode N states, each with its own independent register bit. At any given time, only one bit in the register is valid. For example, when k=3, there are 64 possible 3-mer subsequences, and each of these can be One-Hot encoded to obtain the first vector of the k-mer subsequences.

[0141] For example, for the i-th 3-mer subsequence (i.e., the 3-mer subsequence with index i), a 64-dimensional One-Hot vector can be obtained through encoding. The i-th element of this vector is set to 1, and all other elements are set to 0, in the form [0, 1, 0, 0, ..., 0]. Similarly, each 3-mer subsequence can correspond to a 3-mer One-Hot vector. Furthermore, when k = 1, each base is a 1-mer subsequence, meaning each base in the RNA sequence to be predicted can be encoded to obtain a representation vector for each base. For instance, if the RNA sequence to be predicted contains L bases, for the j-th base (i.e., the base with index j), an L-dimensional One-Hot vector can be obtained through encoding. The j-th element of this vector is set to 1, and all other elements are set to 0, thus obtaining the One-Hot vector for the j-th base. In other examples, each base in the RNA sequence to be predicted can be encoded as a 4-dimensional One-Hot vector based on the base type. For example, base A can be represented by the One-Hot vector [1, 0, 0, 0], U by [0, 0, 0, 1], G by [0, 1, 0, 0], and C by [0, 0, 1, 0]. Correspondingly, a One-Hot vector for each base in the RNA sequence to be predicted can be obtained.

[0142] For example, the RNA sequence AUCUGAAAU to be predicted may include three 3-mer subsequences: AUC, UGA, and AAU. The corresponding three 3-mer One-Hot vectors are as follows: and The first RNA vector sequence can be composed of three bases: a 3-merOne-Hot vector. In the exemplary embodiments of this disclosure, by performing One-Hot encoding on the base k-mer subsequences, each base k-mer subsequence can be transformed into a binary feature, thereby compensating for the deficiencies of classifiers in processing attribute data, so that the classifier can more accurately predict the interaction between RNA sequences and protein sequences.

[0143] In some embodiments of this disclosure, dense vectors can be used to represent each k-mer subsequence. This involves sequentially embedding (vector mapping) encoding each k-mer subsequence, representing each k-mer subsequence with a low-dimensional vector, resulting in multiple corresponding k-mer embedding vectors. These multiple k-mer embedding vectors then form the first RNA vector sequence. For example, the Word2vec algorithm can be used to map each k-mer subsequence in the RNA sequence to a vector space, where each k-mer subsequence can be represented by a vector. Algorithms such as Doc2vec and Glove can also be used to convert the k-mer subsequence into embedding vectors. Alternatively, a BERT (Bidirectional Encoding Representation from Transformer) pre-trained model can be used to encode each k-mer subsequence, obtaining multiple corresponding k-mer embedding vectors. This disclosure does not specifically limit the specific implementation of these methods. In the exemplary embodiments of this disclosure, by embedding encoding the base k-mer subsequences, discrete base k-mer subsequences can be converted into a low-dimensional continuous vector. This continuous vector can better represent each base k-mer subsequence. Moreover, the embedding encoding process is learnable; through continuous training, similar base k-mer subsequences can become closer in the vector space, achieving class distinction while encoding the base k-mer subsequences. This facilitates more accurate prediction of interactions between RNA and protein sequences. Furthermore, it also improves the efficiency of interaction prediction to some extent.

[0144] In one example implementation, after converting the RNA sequence to be predicted into N base k-mer subsequences, each base k-mer subsequence can be encoded to obtain a first vector of N base k-mer subsequences. The first vectors of the N base k-mer subsequences are then sequentially input into a pre-trained recurrent neural network, which outputs N base k-mer vectors, and the N base k-mer vectors form the first RNA vector sequence.

[0145] For example, the first vector can be a One-Hot vector. It's understandable that there are relationships between the bases in an RNA sequence. In this example, all k-mer One-Hot vectors in the RNA sequence to be predicted can be considered as a temporal sequence, and a recurrent neural network can be used to operate on each k-mer One-Hot vector. For example, the 3-mer One-Hot vectors of all bases in the RNA sequence AUCUGAAAU to be predicted can be obtained (…). and After that, the three 3-mer One-Hot vectors can be input into the trained LSTM network, which outputs the corresponding 3-mer vectors for each base, respectively. and The first RNA vector sequence consists of three 3-mer bases. Among them, the LSTM network is a time recurrent neural network, which is suitable for processing and predicting important events with relatively long intervals and delays in time series.

[0146] In one example implementation, after converting the RNA sequence to be predicted into N base k-mer subsequences, each base k-mer subsequence can be encoded to obtain a first vector of N base k-mer subsequences. The first vector of N base k-mer subsequences is then processed (e.g., multiplied) using a first mapping matrix to obtain a second vector of N base k-mer subsequences, and this second vector of N base k-mer subsequences forms the first RNA vector sequence.

[0147] For example, the first vector can be a One-Hot vector, and the second vector can be an Embedding vector. For the RNA sequence AUCUGAAAU to be predicted, it can include three 3-mer subsequences: AUC, UGA, and AAU. Each 3-mer subsequence can be One-Hot encoded to obtain a 3-mer One-Hot vector, which is respectively... and Since the 3-mer One-Hot vector is a 64-dimensional sparse vector, it can be mapped to a dense embedding vector using the first mapping matrix W1, i.e., according to:

[0148]

[0149] Obtain the 3-mer embedding vector of the i-th base in the RNA sequence to be predicted. in, This represents the 3-mer One-Hot vector of the i-th base in the RNA sequence to be predicted. The first mapping matrix W1 is an A*64 parameter matrix; for example, A can be 128 or 256, and this disclosure does not specifically limit the value of A. Based on this, the 3-mer Embedding vectors corresponding to the three 3-mer subsequences can be obtained sequentially, respectively. and Furthermore, the first RNA vector sequence can be composed of three 3-mer embedding vectors.

[0150] In one example implementation, after converting the RNA sequence to be predicted into an N-base k-mer subsequence, reference is made... Figure 3 As shown, each base k-mer subsequence can be encoded according to steps S310 to S330 to obtain the first RNA vector sequence.

[0151] Step S310. Encode each base k-mer subsequence to obtain the first vector of N base k-mer subsequences.

[0152] For example, the first vector can be a One-Hot vector. For the RNA sequence AUCUGAAAU to be predicted, it can include three 3-mer subsequences: AUC, UGA, and AAU. Each 3-mer subsequence can be One-Hot encoded to obtain a 3-mer One-Hot vector, which is respectively... and

[0153] Step S320. Calculate the first vector of the N base k-mer subsequences using the first mapping matrix to obtain the second vector of the N base k-mer subsequences.

[0154] The second vector can be an embedding vector. Since the 3-mer One-Hot vector is a 64-dimensional sparse vector, it can be mapped to a dense embedding vector using the first mapping matrix W1, resulting in three 3-mer embedding vectors, namely... and

[0155] Step S330. Input the second vector of the N base k-mer subsequences into the pre-trained recurrent neural network in sequence, output N base k-mer vectors, and form the first vector sequence of RNA by the N base k-mer vectors.

[0156] It is understandable that there are relationships between the bases in an RNA sequence. In this example, the 3-mer embedding vectors of all bases in the RNA sequence to be predicted can be regarded as a temporal sequence, and then a recurrent neural network can be used to operate on each 3-mer embedding vector. For example, the 3-mer embedding vectors of all bases in the RNA sequence AUCUGAAAU to be predicted can be obtained ( and After that, the three base 3-mer embedding vectors can be sequentially input into the trained LSTM network, which outputs the corresponding 3-mer vector for each base, as follows: and The first RNA vector sequence consists of three 3-mer bases.

[0157] Specifically, we can first embed the vector corresponding to "AUC". Input into an LSTM network allows for the processing of data through the LSTM network. Extract the latent features and output the latent vector at time t. Then, the hidden vector at time t can be... The Embedding vector corresponding to "UGA" at time t+1 The concatenated vectors are then fed into an LSTM network, where their latent features are extracted, and the latent vector at time t+1 is output. Similarly, the current embedding vector can be concatenated with the hidden vector passed down from the previous time step, and features can be extracted from the concatenated vector using an LSTM network. Finally, the embedding vector corresponding to "AAU" can be... Input the hidden vector at time t+1 into the LSTM network. With Embedding vector The concatenated vectors are then processed, and their latent features are extracted using an LSTM network to output the latent vector at the final time step. In other examples, a GRU network can be used to operate on the 3-mer embedding vector for each base. The GRU network has a relatively simple structure and achieves the same effect as the LSTM network. Alternatively, the 3-mer One-Hot vector of each base in the RNA sequence to be predicted can be directly input into the GRU network to obtain the corresponding 3-mer vector. This disclosure does not impose any specific limitations on this method.

[0158] In this embodiment, when using an LSTM network to process multiple 3-mer embedding vectors in the RNA sequence to be predicted, the dependencies between the various 3-mer embedding vectors can be learned and memorized. Based on this, the association information between the RNA sequence and the protein sequence can be obtained more accurately, so as to fuse the association information to accurately predict the interaction between the RNA sequence and the protein sequence.

[0159] In step S230, the protein sequence to be predicted is encoded to obtain the first protein vector sequence.

[0160] After obtaining the protein sequence to be predicted, the protein sequence can be encoded to obtain the first vector sequence of the protein, so as to obtain the association information between the first vector sequence of RNA and the first vector sequence of protein, and then the interaction between the RNA sequence and the protein sequence can be predicted based on the association information.

[0161] In one example implementation, a protein sequence can be represented by an amino acid sequence. This can include 20 amino acids, which are sequentially encoded as A, G, V, I, L, F, P, Y, M, T, S, H, N, Q, W, R, K, D, E, C. For example, a protein sequence can be represented as MTAQDDSYS… Correspondingly, a protein sequence can also be represented using amino acid k-mer subsequences. Specifically, all amino acid k-mer subsequences can be obtained by permuting and combining the 20 amino acids; for a given k value, 20 k-mer subsequences can be obtained. k A set of 20 amino acid k-mer subsequences. For example, when k is 3, there are 20 k-mer subsequences. 3 =8000 amino acid 3-mer subsequences. For example, MTA, QDD, and SYS are three different amino acid 3-mer subsequences. Therefore, the protein sequence MTAQDDSYS… can also be represented as {MTA, QDD, SYS, …}. In other examples, the corresponding amino acid 3-mer subsequences can also be obtained by reading the protein sequence in an overlapping manner. Correspondingly, the amino acid 3-mer subsequences of this protein sequence can also include MTA, TAQ, AQD, etc. Furthermore, according to the physicochemical properties of amino acids, the 20 amino acids can be divided into 7 categories: {A, G, V}, {I, L, F, P}, {Y, M, T, S}, {H, N, Q, W}, {R, K}, {D, E}, and {C}, and each category of amino acids can be re-encoded, such as sequentially encoding them as 1, 2, 3, 4, 5, 6, and 7. For example, the protein sequence MTAQDDSYS… can be converted to 331466333…. Then, the seven classes of amino acids can be arranged and combined to obtain all the amino acid k-mer subsequences. For a certain k value, seven... k The k-mer subsequence of the 20 amino acids is not specifically limited in this disclosure. It is understood that dividing the 20 amino acids into 7 categories is merely illustrative; the 20 amino acids can also be classified according to their composition. Similarly, the four bases of the RNA sequence can also be classified according to actual needs.

[0162] When encoding the predicted protein sequence, one can encode a portion of the amino acids in the predicted protein sequence and use the encoded result as the first vector sequence of the protein, or one can encode all the amino acids in the predicted protein sequence and use the encoded amino acids to form the first vector sequence of the protein, or one can encode all the amino acids in the predicted protein sequence and select a portion of the encoded amino acids to form the first vector sequence of the protein. This disclosure does not make any specific limitations on this.

[0163] In this exemplary embodiment, the example described is the encoding of all amino acids in the protein sequence to be predicted, and the formation of a first protein vector sequence from the encoded amino acids. The protein sequence to be predicted can be converted into M amino acid k-mer subsequences. For example, depending on the value of k, starting from the first amino acid of the protein sequence to be predicted, k consecutive amino acids can be taken to form an amino acid k-mer subsequence, until the last k amino acids are taken, resulting in all amino acid k-mer subsequences of the protein sequence to be predicted. Then, each amino acid k-mer subsequence can be vectorized to obtain M amino acid k-mer vectors, and these M amino acid k-mer vectors can be used to form the first protein vector sequence. For example, the protein sequence to be predicted can be divided into M non-overlapping amino acid k-mer subsequences. For instance, if the protein sequence to be predicted is MTAQDDSYS, it can be divided into three amino acid k-mer subsequences: MTA, QDD, and SYS. Similarly, in other examples, each amino acid in the protein sequence to be predicted can be vectorized to obtain multiple amino acid vectors, which together form the first protein vector sequence. Alternatively, the overlapping parts of the protein sequence to be predicted can be divided into Q amino acid k-mer subsequences, each amino acid k-mer subsequence can be vectorized to obtain Q amino acid k-mer vectors, which together form the first protein vector sequence. This disclosure does not impose specific limitations on this approach.

[0164] In one example implementation, after converting the protein sequence to be predicted into M amino acid k-mer subsequences, each amino acid k-mer subsequence in the protein sequence to be predicted can be encoded to obtain a first vector of M amino acid k-mer subsequences. The first vectors of the M amino acid k-mer subsequences then form the first vector sequence of the protein. For example, when k=3, there can be 8000 possible amino acid 3-mer subsequences, and each amino acid 3-mer subsequence can be One-Hot encoded to obtain the first vector of amino acid k-mer subsequences.

[0165] For example, for the j-th amino acid 3-mer subsequence, i.e., the amino acid 3-mer subsequence with an integer index j, an 8000-dimensional One-Hot vector can be obtained through encoding. The j-th element of this vector is set to 1, and all other elements are set to 0, in the form [1, 0, 0, ..., 0]. Similarly, each amino acid 3-mer subsequence can correspond to an amino acid 3-mer One-Hot vector. For another example, when k = 1, each amino acid is a 1-mer subsequence, meaning that each amino acid in the protein sequence to be predicted can be encoded to obtain a representation vector corresponding to each amino acid. For instance, if the protein sequence to be predicted contains S amino acids, for the j-th amino acid, i.e., the amino acid with an integer index j, an S-dimensional One-Hot vector can be obtained through encoding. The j-th element of this vector is set to 1, and all other elements are set to 0, thus obtaining the One-Hot vector for the j-th amino acid. In other examples, each amino acid in the protein sequence to be predicted can be encoded into a 20-dimensional One-Hot vector based on its amino acid type, thus obtaining a One-Hot vector for each amino acid in the protein sequence to be predicted. Alternatively, the 20 amino acids can be classified, and each amino acid in the protein sequence to be predicted can be encoded into a One-Hot vector with the same vector dimension as the number of classification categories. For example, when the 20 amino acids are divided into 7 categories, each amino acid in the protein sequence to be predicted can be encoded into a 7-dimensional One-Hot vector; this disclosure does not impose specific limitations on this approach.

[0166] For example, the protein sequence MTAQDDSYS to be predicted may include three amino acid 3-mer subsequences: MTA, QDD, and SYS. The corresponding three amino acid 3-mer One-Hot vectors are as follows: and The first vector sequence of a protein can be composed of three amino acid 3-mer One-Hot vectors. In the exemplary embodiments of this disclosure, by performing One-Hot encoding on the amino acid k-mer subsequences, each amino acid k-mer subsequence can be transformed into a binary feature, thereby compensating for the deficiencies of classifiers in processing attribute data, so that the classifier can more accurately predict the interaction between RNA sequences and protein sequences.

[0167] In some embodiments of this disclosure, dense vectors can be used to represent each amino acid k-mer subsequence. This involves sequentially embedding each amino acid k-mer subsequence, representing each subsequence with a low-dimensional vector, resulting in multiple corresponding amino acid k-mer embedding vectors. These multiple embedding vectors then form the first vector sequence of the protein. For example, the Word2vec algorithm can be used to map each amino acid k-mer subsequence in the protein sequence into a vector space, where each subsequence can be represented by a vector. Algorithms such as Doc2vec and GloVe can also be used to convert the amino acid k-mer subsequence into embedding vectors. Alternatively, a BERT pre-trained model can be used to encode each subsequence, yielding multiple corresponding embedding vectors. This disclosure does not specifically limit the specific implementation of these methods. In the example embodiments of this disclosure, embedding encoding of the amino acid k-mer subsequences converts discrete subsequences into a low-dimensional continuous vector, which better represents each subsequence. Furthermore, the embedding encoding process is learnable. Through continuous training, similar amino acid k-mer subsequences can become closer in the vector space, enabling class distinction while encoding amino acid k-mer subsequences. This allows for more accurate prediction of interactions between RNA and protein sequences. Additionally, it improves the efficiency of interaction prediction to some extent.

[0168] In one example implementation, after converting the protein sequence to be predicted into M amino acid k-mer subsequences, each amino acid k-mer subsequence can be encoded to obtain a first vector of M amino acid k-mer subsequences. The first vectors of the M amino acid k-mer subsequences are then sequentially input into a pre-trained recurrent neural network, which outputs M amino acid k-mer vectors, and the M amino acid k-mer vectors form the first protein vector sequence.

[0169] For example, the first vector can be a One-Hot vector. It's understandable that there are relationships between the amino acids in a protein sequence. In this example, all the 3-mer One-Hot vectors of the amino acids in the protein sequence to be predicted can be considered as a temporal sequence, and then a recurrent neural network can be used to operate on each amino acid's 3-mer One-Hot vector. For example, the 3-mer One-Hot vectors of all amino acids in the protein sequence MTAQDDSYS to be predicted can be obtained (…). and After that, the three amino acid 3-mer One-Hot vectors can be sequentially input into the trained LSTM network, which outputs the corresponding 3-mer vector for each amino acid, as follows: and The protein's first vector sequence is composed of three amino acid 3-mer vectors.

[0170] In one example implementation, after converting the protein sequence to be predicted into M amino acid k-mer subsequences, each amino acid k-mer subsequence can be encoded to obtain a first vector of M amino acid k-mer subsequences. A second mapping matrix is ​​then used to perform operations (e.g., multiplication) on the first vector of the M amino acid k-mer subsequences to obtain a second vector of the M amino acid k-mer subsequences. The second vectors of the M amino acid k-mer subsequences then form the first vector sequence of the protein.

[0171] For example, the first vector can be a One-Hot vector, and the second vector can be an Embedding vector. For the protein sequence MTAQDDSYS to be predicted, it can include three amino acid 3-mer subsequences: MTA, QDD, and SYS. Each amino acid 3-mer subsequence can be One-Hot encoded to obtain an amino acid 3-mer One-Hot vector, which is respectively... and Since the amino acid 3-mer One-Hot vector is an 8000-dimensional sparse vector, it can be mapped to a dense embedding vector using the second mapping matrix W2, i.e., according to:

[0172]

[0173] Obtain the 3-mer embedding vector of the j-th amino acid in the protein sequence to be predicted. in, Let W1 represent the one-hot vector of the j-th amino acid 3-mer in the protein sequence to be predicted. The second mapping matrix W2 is a B*8000 parameter matrix. For example, B can be 256 or 128; this disclosure does not specify a particular value for B. Based on this, the 3-mer embedding vectors corresponding to the three amino acid 3-mer subsequences can be obtained sequentially, as follows: and Furthermore, the first vector sequence of a protein can be composed of three amino acid 3-mer embedding vectors.

[0174] In one example implementation, after converting the protein sequence to be predicted into an M-amino acid k-mer subsequence, reference is made... Figure 4 As shown, each amino acid k-mer subsequence can be encoded according to steps S410 to S430 to obtain the first vector sequence of the protein.

[0175] Step S410. Encode each amino acid k-mer subsequence to obtain the first vector of M amino acid k-mer subsequences.

[0176] For example, the first vector can be a One-Hot vector. For the protein sequence MTAQDDSYS to be predicted, it can include three amino acid 3-mer subsequences: MTA, QDD, and SYS. Each amino acid 3-mer subsequence can be One-Hot encoded to obtain an amino acid 3-mer One-Hot vector, which is respectively... and

[0177] Step S420. Calculate the first vector of the M amino acid k-mer subsequences using the second mapping matrix to obtain the second vector of the M amino acid k-mer subsequences.

[0178] The second vector can be an embedding vector. Since the amino acid 3-mer One-Hot vector is an 8000-dimensional sparse vector, it can be mapped to a dense embedding vector using the second mapping matrix W2, resulting in three amino acid 3-mer embedding vectors, respectively. and

[0179] Step S430. Input the second vector of the M amino acid k-mer subsequences into the pre-trained recurrent neural network in sequence, output the M amino acid k-mer vectors, and form the first vector sequence of the protein by the M amino acid k-mer vectors.

[0180] In this example, the 3-mer embedding vectors of all amino acids in the protein sequence to be predicted can be viewed as a temporal sequence, and then a recurrent neural network can be used to operate on each amino acid 3-mer embedding vector. For example, the 3-mer embedding vectors of all amino acids in the protein sequence MTAQDDSYS to be predicted can be obtained ( and After that, the three amino acid 3-mer embedding vectors can be sequentially input into the trained LSTM network, which outputs the corresponding 3-mer vector for each amino acid, as follows: and The protein's first vector sequence is composed of three amino acid 3-mer vectors.

[0181] Specifically, you can first embed the vector corresponding to "MTA". Input into an LSTM network allows for the processing of data through the LSTM network. Extract the latent features and output the latent vector at time t. Then, the hidden vector at time t can be... The Embedding vector corresponding to "QDD" at time t+1 The concatenated vectors are then fed into an LSTM network, where their latent features are extracted, and the latent vector at time t+1 is output. Finally, the embedding vector corresponding to "SYS" can be... Input the hidden vector at time t+1 into the LSTM network. With Embedding vector The concatenated vectors are then processed, and their latent features are extracted using an LSTM network to output the latent vector at the final time step. In other examples, the GRU network can also be used to calculate the 3-mer embedding vector for each amino acid. Alternatively, the 3-mer one-hot vector of each amino acid in the protein sequence to be predicted can be directly input into the GRU network to obtain the corresponding base 3-mer vector; this disclosure does not impose specific limitations on this method.

[0182] In this implementation, when using an LSTM network to process multiple amino acid 3-mer embedding vectors in the protein sequence to be predicted, the dependencies between these vectors can be learned and memorized. Based on this, the association information between the RNA and protein sequences can be obtained more accurately, allowing for the fusion of this association information to precisely predict the interactions between the RNA and protein sequences. Furthermore, it is understood that, to facilitate the prediction of interactions between RNA and protein, the LSTM network outputs base 3-mer vectors... and amino acid 3-mer vector The dimensions can be consistent, such as all being 64 dimensions or all being 128 dimensions. This disclosure does not make any specific limitation on this.

[0183] In step S240, the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence are obtained through an attention mechanism model.

[0184] In the exemplary embodiments of this disclosure, the association vector sequences of the first RNA vector sequence and the first protein vector sequence can be obtained through an attention mechanism model. Introducing the association vector sequences of the first RNA vector sequence and / or the first protein vector sequence when predicting the interaction between the input RNA and protein sequences can improve the accuracy of RNA-protein interaction prediction. The attention mechanism model is a machine learning model that simulates human visual attention behavior. When observing a painting, after a global scan of the image, a target area requiring attention is identified, and then more attention is paid to obtain detailed information about that target area. Therefore, human attention behavior can be abstracted and applied to machine learning models.

[0185] In attention-based models, each input vector requires three vector representations: query, key, and value. The attention mechanism maps the query and a series of key-value pairs to the output. The output is obtained by weighted summation of the values, and the weight corresponding to each value can be calculated from the query and key using a compatibility function. For example, the compatibility function could be the softmax function.

[0186] For example, refer to Figure 5 As shown, the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence can be obtained through the attention mechanism model according to steps S510 to S530.

[0187] Step S510. Perform feature extraction on the first RNA vector sequence to obtain the first hidden RNA vector.

[0188] When the RNA sequence to be predicted is converted into an N-base k-mer subsequence, the corresponding first RNA vector sequence can be: The first RNA vector sequence includes N base k-mer vectors. Before feature extraction from the first RNA vector sequence, three weight matrices can be initialized: a query weight matrix, a bond weight matrix, and a value weight matrix. Then, the query representation, bond representation, and value representation of each base k-mer vector in the first RNA vector sequence can be obtained from these three weight matrices. Correspondingly, the first implicit RNA vector obtained by feature extraction from each base k-mer vector in the first RNA vector sequence can include an RNA first vector, an RNA second vector, and an RNA third vector. For example, the first query weight matrix can be used to operate on the first RNA vector sequence to obtain the RNA first vector for each base k-mer subsequence; the first bond weight matrix can be used to operate on the first RNA vector sequence to obtain the RNA second vector for each base k-mer subsequence; and the first value weight matrix can be used to operate on the first RNA vector sequence to obtain the RNA third vector for each base k-mer subsequence. It can be understood that the RNA first vector, RNA second vector, and RNA third vector correspond to the query representation, bond representation, and value representation of the base k-mer vector, respectively.

[0189] For example, for the k-mer vector of the i-th base in the first vector sequence of RNA The first query weight matrix can be used First-key weight matrix and the first value weight matrix right Perform the operation to obtain the query representation queryR of the k-mer vector of the i-th base. i 、Key represents keyR i Sum value is represented by valueR i ,Right now:

[0190]

[0191] The first query weight matrix can be used First-key weight matrix and the first value weight matrix The operation is performed sequentially on each base k-mer vector in the first RNA vector sequence, or all base k-mer vectors in the first RNA vector sequence can be operated on simultaneously to improve the efficiency of interaction prediction; this disclosure does not specifically limit this approach. It should be noted that the first RNA hidden vector obtained by feature extraction from the first RNA vector sequence can include hidden vectors of multiple base k-mer subsequences, and the hidden vector of each base k-mer subsequence includes the first RNA vector, the second RNA vector, and the third RNA vector. For example, the hidden vector of the i-th base k-mer subsequence can include the first RNA vector queryR. i RNA second vector keyR i and RNA third vector valueR i .

[0192] Step S520. Perform feature extraction on the first vector sequence of the protein to obtain the first hidden vector of the protein.

[0193] Similarly, when converting the protein sequence to be predicted into an M-amino acid k-mer subsequence, the corresponding first vector sequence of the protein can be... The protein vector sequence includes M amino acid k-mer vectors. When extracting features from the first protein vector sequence using an attention mechanism model, the query representation, bond representation, and value representation of each amino acid k-mer vector in the first protein vector sequence can also be obtained based on the query weight matrix, bond weight matrix, and value weight matrix.

[0194] Correspondingly, the first latent vector of a protein obtained by feature extraction of each amino acid k-mer vector in the first protein vector sequence can include a first protein vector, a second protein vector, and a third protein vector. For example, the first protein vector sequence can be processed using a second query weight matrix to obtain the first protein vector for each amino acid k-mer subsequence; the second bond weight matrix can be processed using a second bond weight matrix to obtain the second protein vector for each amino acid k-mer subsequence; and the third protein vector can be processed using a second value weight matrix to obtain the third protein vector for each amino acid k-mer subsequence. It can be understood that the first protein vector, the second protein vector, and the third protein vector correspond to the query representation, bond representation, and value representation of the amino acid k-mer vector, respectively.

[0195] For example, for the k-mer vector of the j-th amino acid in the first vector sequence of a protein... The second query weight matrix can be used Second bond weight matrix Second-value weight matrix right Perform the calculation to obtain the query representation queryP of the k-mer vector of the j-th amino acid. j , key represents keyP j Sum value is represented by valueP j ,Right now:

[0196]

[0197] The second query weight matrix can be used Second bond weight matrix Second-value weight matrix The k-mer vectors of each amino acid in the first protein vector sequence are calculated sequentially, or all amino acid k-mer vectors in the first protein vector sequence can be calculated simultaneously to improve the efficiency of interaction prediction; this disclosure does not specifically limit this approach. It should be noted that the first hidden vector of the protein obtained by feature extraction from the first protein vector sequence can include hidden vectors of multiple amino acid k-mer subsequences, and the hidden vector of each amino acid k-mer subsequence includes a first protein vector, a second protein vector, and a third protein vector. For example, the hidden vector of the j-th amino acid k-mer subsequence can include the first protein vector queryP. j Protein second vector keyP j and the protein's third vector valueP j .

[0198] Step S530. Perform operations on the first hidden vector of RNA and the first hidden vector of protein to obtain the correlation vector sequence of the first vector sequence of RNA and the correlation vector sequence of the first vector sequence of protein.

[0199] After obtaining the first hidden vector of RNA for each k-mer subsequence and the first hidden vector of protein for each amino acid k-mer subsequence, calculations can be performed on these vectors to obtain the association vector sequences of the RNA and protein first vector sequences. The association vector sequences of the protein first vector sequences can be fused into the protein first vector sequence, and vice versa, to incorporate more sequence information for predicting RNA-protein interactions, thereby improving prediction accuracy. Alternatively, the association vector sequences of the protein first vector sequences can be fused into the protein first vector sequence alone, allowing for interaction prediction using both the protein and RNA first vector sequences with fused RNA sequence information. Interaction prediction can also be made directly using the association vector sequences of the protein first vector sequence and the RNA first vector sequence; this disclosure does not impose any specific limitations on this.

[0200] For example, the first hidden RNA vector of the i-th base k-mer subsequence is the first RNA vector (queryR). i ), RNA second vector (keyR) i ) and RNA third vector (valueR) i The first hidden vector of the protein for the j-th amino acid k-mer subsequence is the first protein vector (queryP). j ), protein second vector (keyP) j ) and protein third vector (valueP) j The similarity between the first RNA vector and the second protein vector can be calculated to obtain the first RNA attention score (i.e., weights). The third protein vector is then summed based on the first RNA attention score to obtain the associated vector sequence of the first RNA vector sequence. For example, an attention scoring function can be used to calculate the similarity between the first RNA vector and the second protein vector, and the calculated similarity value can be normalized. The normalized result is the first RNA attention score. The attention scoring function used can be a dot product operation, or it can use an additive model, a scaled-click model, or a bilinear model to calculate the attention score; this disclosure does not specifically limit this method.

[0201] Specifically, it can be based on:

[0202]

[0203] Normalization yields the first attention score for RNA. In some implementations, normalization can be performed using the softmax function. <queryR i keyP j > indicates the first vector queryR for the i-th RNA. i and the second vector keyP of the j-th protein j The dot product operation is performed, where k can take values ​​from [1, M], where M is the number of amino acid k-mer subsequences contained in the protein to be predicted. <queryR i keyP k > indicates the first vector queryR for the i-th RNA. i and the second vector keyP of the k-th protein j Perform the dot product operation. During normalization, Let represent the sum of the dot products of the first vector of the i-th RNA and the second vectors of all proteins. The queryR represents the first vector of the i-th RNA under the e-index. i and the second vector keyP of the j-th protein j The dot product results are used to amplify the differences between the various dot product results using the e-exponent. Furthermore, this example fuses two different types of information: RNA sequence and protein sequence. Therefore, an exponent can be used for smoothing in equation (5) to more accurately obtain the correlation information between the RNA sequence and the protein sequence. After obtaining the first attention score for RNA, it can be calculated according to:

[0204]

[0205] We perform a weighted summation of the third vectors of all proteins to obtain the correlation vector of the first vector sequence of the i-th RNA. Among them, RNA first attention score The valueP represents the third vector of the k-th protein. k The corresponding weights.

[0206] Similarly, the similarity between the first protein vector and the second RNA vector can be calculated to obtain the first protein attention score (i.e., weight). Then, the first protein attention score is summed with the third RNA vector to obtain the associated vector sequence of the first protein vector sequence. For example, an attention scoring function can be used to calculate the similarity between the first protein vector and the second RNA vector, and the calculated similarity value can be normalized. The normalized result is the first protein attention score. The attention scoring function used can be a dot product operation, or it can be calculated using an additive model, a scaled-click model, or a bilinear model, etc. This disclosure does not specifically limit the method used.

[0207] Specifically, it can be based on:

[0208]

[0209] Normalization yields the protein's first attention score. In some implementations, normalization can be performed using the softmax function. <queryP j keyR i > represents the first vector queryP for the j-th protein. j and the second vector keyR of the i-th RNA i The dot product operation is performed, where k can take values ​​from [1, N], where N is the number of k-mer subsequences in the RNA to be predicted. <queryP j keyR k > indicates a query P for the j-th protein. j and the second vector keyR of the k-th RNA k Perform the dot product operation. During normalization, Let represent the sum of the dot products of the first vector of the j-th protein and the second vectors of all RNAs. Let queryP represent the first vector of the i-th protein under the e-index. j and the second vector keyR of the j-th RNA i The dot product results are used to amplify the differences between the various dot product results using the e-exponent. After obtaining the protein's first attention score, it can be determined according to:

[0210]

[0211] For all RNA third vectors, valueR k By performing a weighted summation, we obtain the correlation vector of the first vector sequence of the j-th protein. Among them, the protein's first attention score Represents the third vector valueR of the k-th RNA.k The corresponding weights.

[0212] In this example, an attention mechanism model is used to determine the association information between RNA and protein sequences, thereby fusing the RNA sequence, protein sequence, and association information. Introducing the fused sequence information into the prediction of RNA-protein interaction can improve the accuracy of RNA-protein interaction prediction.

[0213] In step S250, the interaction between the RNA sequence and the protein sequence to be predicted is determined based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence.

[0214] In the exemplary embodiments of this disclosure, the interaction between the RNA sequence and the protein sequence to be predicted can be determined based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence. Alternatively, a second RNA vector sequence and a second protein vector sequence can be obtained based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence. For example, the association vector sequence of the first RNA vector sequence and the first RNA vector sequence can be concatenated to obtain an RNA fusion vector sequence. The RNA fusion vector sequence is then input into a pre-trained recurrent neural network to output the second RNA vector sequence. Similarly, the association vector sequence of the first protein vector sequence and the first protein vector sequence can be concatenated to obtain a protein fusion vector sequence, and the concatenated vector sequence is input into a pre-trained recurrent neural network to output the second protein vector sequence. Finally, the interaction between the RNA sequence and the protein sequence to be predicted can be determined based on the second RNA vector sequence and the second protein vector sequence. It is understood that the RNA fusion vector sequence and the protein fusion vector sequence can also be used to determine the interaction between the RNA sequence and the protein sequence to be predicted, and this disclosure does not limit this approach.

[0215] For example, for the k-mer vector of the i-th base in the first vector sequence of RNA The correlation vector of this base k-mer vector is The two can be concatenated by performing a dot product operation, i.e. The result of the operation is the RNA vector sequence after fusing protein sequence information, which can be denoted as: For the k-mer vector of the j-th amino acid in the first vector sequence of the protein The correlation vector of this amino acid k-mer vector is Alternatively, a dot product operation can be performed on the two to concatenate them, i.e. The result of the operation is the protein vector sequence after fusing RNA sequence information, which can be denoted as:

[0216] Taking the RNA sequence to be predicted as an example, in order to further capture the sequence... The association between the k-mer vectors of each base after fusing protein sequence information can reveal the sequence. As input to the LSTM network, it is processed by the LSTM network. The hidden features are extracted sequentially. In the exemplary implementation of this disclosure, the output of the LSTM network at the last time step can be used as the final vector representation of the RNA sequence to be predicted, or the average of the LSTM network outputs at all time steps can be used as the final vector representation of the RNA sequence to be predicted. For example, the hidden vector output at the last time step can be used as the final vector representation of the RNA sequence to be predicted. As the second vector sequence of RNA. Similarly, When used as input to an LSTM network, the output of the LSTM network at the last time step can be used as the final vector representation of the protein sequence to be predicted, or the average of the LSTM network outputs at all time steps can be used as the final vector representation of the protein sequence to be predicted. For example, the hidden vector output at the last time step can be used as the input. As a second vector sequence for proteins. In other examples, a GRU network can also be used for... The extraction of hidden features is not specifically limited in this disclosure.

[0217] In one example implementation, reference Figure 6 As shown, the second vector sequence of RNA and the second vector sequence of protein can also be obtained by information fusion through the attention mechanism model according to steps S610 to S640.

[0218] Step S610. Concatenate the associated vector sequence of the first RNA vector sequence with the first RNA vector sequence to obtain the RNA fusion vector sequence.

[0219] For example, for the k-mer vector of the i-th base in the first vector sequence of RNA The correlation vector of this base k-mer vector is The two can be concatenated by performing a dot product operation, i.e. Obtain RNA fusion vector sequence

[0220] Then, the RNA fusion vector sequence can be used as input to an LSTM network to extract hidden features, thereby capturing the correlation between the k-mer vectors of each base in the RNA fusion vector sequence. The output of the LSTM network at all time points can be denoted as...

[0221] Step S620. Concatenate the associated vector sequence of the first protein vector sequence with the first protein vector sequence to obtain the protein fusion vector sequence.

[0222] For example, for the k-mer vector of the j-th amino acid in the first vector sequence of a protein... The correlation vector of this amino acid k-mer vector is Alternatively, a dot product operation can be performed on the two to concatenate them, i.e. Obtain protein fusion vector sequence

[0223] The protein fusion vector sequence can be used as input to an LSTM network. The LSTM network extracts latent features to capture the correlations between the k-mer vectors of each amino acid in the protein fusion vector sequence. The output of the LSTM network at all time points can be denoted as...

[0224] Step S630. Obtain the autocorrelation vector sequence of the RNA fusion vector sequence through the attention mechanism model, and obtain the second RNA vector sequence based on the autocorrelation vector sequence of the RNA fusion vector sequence.

[0225] Obtain new RNA fusion vector sequences Subsequently, to obtain the correlation information between the k-mer vectors of each base in the RNA fusion vector sequence, an attention mechanism model can be used to obtain the self-correlation vector sequence of the RNA fusion vector sequence. (Reference) Figure 7 As shown, step S630 can be further derived from steps S710 to S730:

[0226] Step S710. Perform feature extraction on the RNA fusion vector sequence to obtain the second hidden RNA vector.

[0227] The second hidden RNA vector obtained by feature extraction of each k-mer vector in the RNA fusion vector sequence can include a fourth, fifth, and sixth RNA vector. For example, the RNA fusion vector sequence can be processed using a third query weight matrix to obtain the fourth RNA vector for each k-mer subsequence; the RNA fusion vector sequence can be processed using a third bond weight matrix to obtain the fifth RNA vector for each k-mer subsequence; and the RNA fusion vector sequence can be processed using a third-value weight matrix to obtain the sixth RNA vector for each k-mer subsequence.

[0228] For example, the RNA fusion vector sequence obtained after inputting the RNA fusion vector sequence into an LSTM network becomes For the k-mer vector of the i-th base The third query weight matrix can be used Third-bond weight matrix and the third-value weight matrix right Perform the operation to obtain the k-mer vector of the i-th base. The query represents Key representation Sum value representation Right now:

[0229]

[0230] The third query weight matrix can be used Third-bond weight matrix and the third-value weight matrix k-mer vector of each base in the RNA fusion vector sequence in turn The operation can also be performed on all base k-mer vectors in the RNA fusion vector sequence simultaneously to improve the efficiency of interaction prediction; however, this disclosure does not specifically limit this.

[0231] Step S720. Perform operations on the second hidden vector of the RNA to obtain the self-correlated vector sequence of the RNA fusion vector sequence;

[0232] After obtaining the second hidden vector of RNA for each k-mer subsequence, the fourth vector of RNA can be processed. RNA Fifth Vector and RNA sixth vector The calculations yield the self-correlation vector sequence of the RNA fusion vector sequence. For example, the similarity between the fourth and fifth RNA vectors can be calculated to obtain the second RNA attention score. Then, the sixth RNA vector is summed based on the second RNA attention score to obtain the self-correlation vector sequence of the RNA fusion vector sequence, which is to obtain the correlation information between the k-mer vectors of each base in the RNA fusion vector sequence.

[0233] Specifically, it can be based on:

[0234]

[0235] Normalization yields RNA second attention score in, This represents the fourth vector for the i-th RNA. and the fifth vector of the j-th RNA Perform dot product operation. This represents the sum of the dot product of the fourth vector of the i-th RNA and all the fifth vectors of the RNA. It can be seen that the attention scoring function used in equation (10) is a dot product operation. Attention scores can also be calculated using additive models, scaled-click models, and bilinear models, etc., and this disclosure does not specifically limit this. Furthermore, this example fuses information of the same type as the RNA sequence, so exponential smoothing is not required in equation (10) to accurately obtain the correlation information between the k-mer vectors of each base. After obtaining the second attention score of the RNA, it can be calculated according to:

[0236]

[0237] The weighted summation of all the sixth RNA vectors yields the k-mer vector of the i-th base in the RNA fusion vector sequence. self-correlated vectors That is, the k-mer vector of bases after incorporating information from other bases.

[0238] Step S730. Perform calculations on the self-correlation vector sequence of the RNA fusion vector sequence to obtain the second RNA vector sequence.

[0239] According to:

[0240]

[0241] The calculated second RNA vector sequence can be denoted as v. RNA .in, Let v be the k-mer vector of the i-th base obtained after incorporating information from other bases. The value v is obtained using the mean operation. RNA When representing a complete RNA sequence to be predicted, it is not necessary to consider whether the RNA and protein sequences have the same length; accurate prediction of interactions can still be performed. Understandably, this can also be... The summation result can be used as the second RNA vector sequence, or it can be directly used as... As a second vector sequence of RNA, this disclosure does not impose specific limitations on it.

[0242] It should be noted that, in some embodiments, after obtaining the RNA fusion vector sequence and the protein fusion vector sequence, the step of inputting them into the LSTM network can be omitted. Instead, the attention mechanism model can be used directly to obtain the autocorrelation vector sequence of the RNA fusion vector sequence and the autocorrelation vector sequence of the protein fusion vector sequence, and the interaction can be predicted using the autocorrelation vector sequence of the RNA fusion vector sequence and the autocorrelation vector sequence of the protein fusion vector sequence.

[0243] Step S640. Obtain the autocorrelation vector sequence of the protein fusion vector sequence through the attention mechanism model, and obtain the second vector sequence of the protein based on the autocorrelation vector sequence of the protein fusion vector sequence.

[0244] A new protein fusion vector sequence was obtained. Subsequently, to obtain the correlation information between the k-mer vectors of each amino acid in the protein fusion vector sequence, an attention mechanism model can be used to obtain the self-correlation vector sequence of the protein fusion vector sequence. (Reference) Figure 8 As shown, step S640 can be further derived from steps S810 to S830:

[0245] Step S810. Perform feature extraction on the protein fusion vector sequence to obtain the second hidden vector of the protein.

[0246] The second hidden protein vector obtained by feature extraction of each amino acid k-mer vector in the protein fusion vector sequence can include a fourth, fifth, and sixth protein vector. For example, the fourth query weight matrix can be used to operate on the protein fusion vector sequence to obtain the fourth protein vector for each amino acid k-mer subsequence; the fourth bond weight matrix can be used to operate on the protein fusion vector sequence to obtain the fifth protein vector for each amino acid k-mer subsequence; and the fourth value weight matrix can be used to operate on the protein fusion vector sequence to obtain the sixth protein vector for each amino acid k-mer subsequence.

[0247] For example, the protein fusion vector sequence obtained after inputting the protein fusion vector sequence into an LSTM network becomes... For the k-mer vector of the j-th base The fourth query weight matrix can be used Fourth bond weight matrix and the fourth-value weight matrix right Perform the operation to obtain the k-mer vector of the j-th base. The query represents Key representation Sum value representation Right now:

[0248]

[0249] The third query weight matrix can be used Third-bond weight matrix and the third-value weight matrix k-mer vector of each base in the RNA fusion vector sequence in turn The operation can also be performed on all base k-mer vectors in the RNA fusion vector sequence simultaneously to improve the efficiency of interaction prediction; however, this disclosure does not specifically limit this.

[0250] Step S820. Perform calculations on the second hidden vector of the protein to obtain the self-correlated vector sequence of the protein fusion vector sequence.

[0251] After obtaining the second hidden vector of the protein for each amino acid k-mer subsequence, the fourth hidden vector of the protein can be obtained. Protein Fifth Vector and the sixth vector of proteins The calculations yield a self-correlated vector sequence of the protein fusion vector sequence. For example, the similarity between the fourth and fifth protein vectors can be calculated to obtain the second protein attention score. The sixth protein vector is then summed based on the second protein attention score to obtain the self-correlated vector sequence of the protein fusion vector sequence, which is the association information between the k-mer vectors of each amino acid in the protein fusion vector sequence.

[0252] Specifically, it can be based on:

[0253]

[0254] Normalization yields the protein's second attention score. in, This represents the fourth vector for the j-th protein. and the fifth vector of the i-th protein Perform dot product operation. This represents the sum of the dot product of the fourth vector of the j-th protein and the fifth vectors of all proteins. After obtaining the protein's second attention score, it can be calculated based on:

[0255]

[0256] The weighted summation of the sixth vectors of all proteins yields the k-mer vector of the j-th amino acid in the protein fusion vector sequence. self-correlated vectors That is, the amino acid k-mer vector after incorporating information from other amino acids.

[0257] Step S830. Perform calculations on the self-correlation vector sequence of the protein fusion vector sequence to obtain the second vector sequence of the protein.

[0258] According to:

[0259]

[0260] The calculated second vector sequence of the protein can be denoted as v. P .in, Let v be the k-mer vector of the j-th amino acid obtained after incorporating information from other amino acids. The vector is obtained using the mean operation. P When representing the complete protein sequence to be predicted, it is not necessary to consider whether the sequence lengths of the protein sequence and the RNA sequence are the same; accurate prediction of interactions can still be performed. Understandably, it is also possible to directly represent... The summation result can be used as the second vector sequence of the protein, or it can be directly used as... As a second vector sequence of a protein, this disclosure does not specifically limit it.

[0261] In the exemplary embodiments of this disclosure, it is necessary to predict the interaction between the RNA sequence and the protein sequence to be predicted. The prediction result may be that there is an interaction between the RNA sequence and the protein sequence to be predicted, or there may be no interaction between the RNA sequence and the protein sequence to be predicted, i.e., binary classification prediction is performed.

[0262] After obtaining the second vector sequence of RNA and the second vector sequence of protein, refer to Figure 9 As shown, the interaction between the RNA sequence and the protein sequence to be predicted can be determined according to steps S910 to S930.

[0263] Step S910. The second RNA vector sequence and the second protein vector sequence are spliced ​​together to obtain the feature vector to be predicted.

[0264] For example, the resulting second RNA vector sequence is v RNA The second vector sequence of the protein is v P In this case, a dot product operation can be performed on the two to concatenate them, that is... <v RNA v P The original feature vector obtained by splicing can be denoted as v. In other examples, if the obtained second RNA vector sequence is... The second vector sequence of the protein is In this case, a dot product operation can be performed on the two to concatenate them, that is... The interaction between RNA and protein sequences can also be predicted based on the original feature vector obtained by splicing, and this disclosure does not specifically limit this.

[0265] To facilitate subsequent binary classification prediction, the original feature vector v can be mapped to a two-dimensional feature vector to be predicted using the third mapping matrix W3, i.e., according to:

[0266] c = W3 × v (17)

[0267] The feature vector to be predicted is obtained as c; where c is a 2-dimensional feature vector to be predicted [c0, c1], the third mapping matrix W3 is a 2*C parameter matrix, v is the original feature vector, and the value of C is consistent with the dimension of the original feature vector.

[0268] Step S920. Obtain the predicted interaction value between the RNA sequence and the protein sequence to be predicted based on the feature vector to be predicted.

[0269] After obtaining the feature vector to be predicted for interaction prediction, the predicted interaction value between the RNA sequence and the protein sequence can be obtained from the feature vector, and the interaction between the RNA sequence and the protein sequence can be determined based on the predicted interaction value.

[0270] For example, the feature vector to be predicted can be input into a classifier, which then classifies the interaction between the RNA and protein sequences to be predicted based on the feature vector. After classification, the predicted interaction value between the RNA and protein sequences is output. For instance, a Softmax classifier can be used to predict the interaction between the RNA and protein sequences. Specifically, the Softmax classifier can be used to transform the feature vector to be predicted, obtaining probability distributions for "interaction exists" and "interaction does not exist" for the interaction between the RNA and protein sequences.

[0271] For example, the probability of an interaction between the RNA sequence and the protein sequence to be predicted can be obtained using the Softmax classifier:

[0272]

[0273] The probability that there is no interaction between the RNA sequence and the protein sequence to be predicted is:

[0274]

[0275] Where r represents the RNA sequence to be predicted, p represents the protein sequence to be predicted, c0 is the first eigenvalue of the feature vector to be predicted, and c1 is the second eigenvalue of the feature vector to be predicted. When the feature vector to be predicted is a 2-dimensional vector, the vector is (c0, c1). In other examples, logistic regression classifiers and SVM (Support Vector Machine) classifiers can also be used for binary classification prediction to obtain the predicted interaction value between the RNA sequence and the protein sequence to be predicted based on the feature vector to be predicted. This disclosure does not specifically limit this.

[0276] Step S930. Determine the interaction between the RNA sequence and the protein sequence to be predicted based on the interaction prediction value.

[0277] After obtaining the predicted interaction values ​​between the RNA and protein sequences to be predicted using a classifier, the interaction between the sequences can be determined based on these predicted values. For example, if the predicted interaction values ​​meet a preset threshold condition, it can be determined that there is an interaction between the RNA and protein sequences to be predicted.

[0278] For example, the Softmax classifier can be used to obtain the probability P(1|r, p) of an interaction between the RNA and protein sequences to be predicted. Here, P(1|r, p) can be any value between 0 and 1. For instance, a threshold of 0.5 can be preset for the probability of interaction. When P(1|r, p) > 0.5, the prediction result can be marked as 1, indicating an interaction between the RNA and protein sequences. When P(1|r, p) ≤ 0.5, the prediction result can be marked as 0, indicating no interaction between the RNA and protein sequences. In other examples, it can also be set so that when P(1|r, p) ≥ 0.5, an interaction between the RNA and protein sequences is determined, and when P(1|r, p) < 0.5, no interaction is determined. Finally, the prediction result of the interaction between the RNA and protein sequences can be output to the terminal device for user viewing. It should be noted that the output may only show the probability that there is an interaction between the RNA sequence and the protein sequence to be predicted, or it may only show the probability that there is no interaction between the RNA sequence and the protein sequence to be predicted, or it may simultaneously show the probability that there is an interaction between the RNA sequence and the protein sequence to be predicted and the probability that there is no interaction. This disclosure does not make any specific limitations on this.

[0279] In the exemplary embodiments of this disclosure, when determining the interaction between the RNA sequence and the protein sequence to be predicted based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence, the association vector sequences of the first RNA vector sequence and the protein first vector sequence can be directly concatenated to obtain the feature vector to be predicted. This feature vector is then input into a classifier to obtain the predicted interaction value between the RNA sequence and the protein sequence, thereby determining the interaction between the RNA sequence and the protein sequence to be predicted. Alternatively, the RNA fusion vector sequence and the protein fusion vector sequence can be directly concatenated to obtain the feature vector to be predicted. This feature vector is then input into a classifier to obtain the predicted interaction value between the RNA sequence and the protein sequence, thereby determining the interaction between the RNA sequence and the protein sequence to be predicted. This disclosure does not limit the approach in this way.

[0280] In the exemplary embodiments of this disclosure, reference is made to Figure 10 As shown, the recurrent neural network and attention mechanism model can be pre-trained according to steps S1010 to S1040 to optimize all model parameters in each prediction model, and then the RNA and protein sequences with unknown interactions can be predicted according to the final model obtained from the training.

[0281] Step S1010. Obtain a training dataset, which includes positive RNA-protein pairs and negative RNA-protein pairs.

[0282] For example, various models can be trained based on the RPI1807 dataset. This dataset contains 3243 RNA-protein pairs, specifically 1807 positive pairs and 1436 negative pairs. Positive pairs indicate an interaction between the RNA and protein sequences in the RNA-protein pair, while negative pairs indicate no interaction. A training dataset of 1200 positive and 1000 negative pairs can be selected, or all RNA-protein pairs can be used. It's important to understand that the number of RNA-protein pairs in the training dataset is merely illustrative; any number of RNA-protein pairs can be used to train each model multiple times to improve its performance. Positive RNA-protein pairs can be labeled with a value of "1," indicating an interaction, while negative RNA-protein pairs can be labeled with a value of "0," indicating no interaction. It is understandable that other examples can also be based on datasets such as RPI2241 and RPI369, and this disclosure does not impose any specific limitations on them.

[0283] Step S1020. Use the recurrent neural network and attention mechanism model to determine the predicted interaction value of each RNA-protein pair in the training dataset.

[0284] Similarly, a recurrent neural network can be used to encode the RNA and protein sequences in each RNA-protein pair in the training dataset, obtaining the corresponding first RNA vector sequence and first protein vector sequence. An attention mechanism model can be used to obtain the association vector sequences of the first RNA and protein vector sequences, and based on these sequences, second RNA and second protein vector sequences can be obtained. These second RNA and second protein vector sequences are then concatenated and dimensionality-reduced to obtain the feature vector to be predicted. Finally, a classifier can be used to classify and predict the feature vector, obtaining the predicted interaction value for each RNA-protein pair.

[0285] Step S1030. Calculate the predicted interaction value and label value of each RNA-protein pair in the training dataset using the loss function to obtain the corresponding loss value.

[0286] Each RNA-protein pair in the training dataset has a label value, such as 1 for each positive pair and 0 for each negative pair. For example, the i-th RNA-protein pair is a positive example, and its label value is 1. The interaction value p(1|r) can be predicted based on this RNA-protein pair. i p i p(0|r) i p i The loss function is calculated using the label value 1 and the target value 1, yielding the corresponding loss value. During model training, the goal is to minimize the interaction predictions to be infinitely close to the label values, i.e., to minimize the objective function. In one example, the cross-entropy loss function can be chosen as the objective function. When calculating the cross-entropy loss function, if the label value is 1, p(1|r) = 1 / 2 * ... i p i The closer p(1|r) is to 1, the smaller the calculated loss value. i p i The closer p(0|r) is to 0, the larger the calculated loss value. At the same time, p(0|r) i p i The closer p(0|r) is to 1, the larger the calculated loss value. i p iThe closer the cross-entropy loss function is to 0, the smaller the calculated loss value. It's understandable that the cross-entropy loss function is a performance function in the prediction model, used to estimate the degree of discrepancy between the model's predicted values ​​and the labeled values. The smaller the calculated cross-entropy loss function value, the better the model's prediction performance.

[0287] Specifically, the cross-entropy loss function can be:

[0288]

[0289] Where, r i p represents the i-th RNA sequence in the training dataset. i Let y represent the i-th protein sequence in the training dataset. i p(1|r) represents the label value of the i-th RNA-protein pair in the training dataset. i p i ) represents the predicted value of the i-th RNA-protein pair in the training dataset that there is an interaction, p(0|r i p i ) represents the predicted value that the i-th RNA-protein pair in the training dataset does not interact, and K is the total number of RNA-protein pairs in the training dataset.

[0290] Step S1040. Adjust the model parameters of the recurrent neural network and attention mechanism model according to the loss value.

[0291] The model parameters can be weight parameters, bias parameters, and parameter matrices, such as mapping matrices W1, W2, and W3. For example, the model parameters of each model can be iteratively updated based on the calculated loss value. When the iteration termination condition is met, the training of model parameters for multiple interaction prediction models is complete. For example, the stochastic gradient descent algorithm can be used to update the model parameters. According to the backpropagation principle, the objective function, such as the cross-entropy loss function, is continuously calculated, and the model parameters of each model are updated simultaneously based on the calculated loss value. When the objective function converges to its minimum value, the training of all model parameters is complete. Alternatively, the model parameters can be updated iteratively in reverse order. When the preset number of iterations is met, the training of all model parameters is complete. After iteration, the optimized model parameters can be obtained. In other examples, the objective function can be minimized alternately using least squares, Adam optimization algorithms, etc., and the model parameters can be updated sequentially from back to front to optimize the parameters.

[0292] During the training process described above, the parameters of the recurrent neural network (RNN) and the attention mechanism model can be trained simultaneously. For example, using L as the objective function, the mapping matrix W3 in the fully connected layer can be adjusted first. Since the attention mechanism model is needed to extract features from the first vector sequences of RNA and protein before binary classification prediction, and the RNN is needed to encode the RNA and protein sequences to be predicted, backpropagation can be further performed on the attention mechanism model and the RNN to adjust the model parameters and mapping matrices W1 and W2. Through multiple backpropagations layer by layer, the parameters of each model can eventually converge, or training can terminate after a certain number of iterations. This training method allows for the simultaneous training of the RNN and the attention mechanism model, ensuring higher accuracy and precision for each model while improving training efficiency. After training, the final models can be used to predict the interactions between the RNA and protein sequences to be predicted.

[0293] In one specific example implementation, reference is made to Figure 11 As shown, the interaction between the RNA and protein sequences to be predicted can be predicted using the trained attention mechanism model, LSTM network and Softmax classifier according to steps S1101 to S1106.

[0294] Step S1101. The RNA sequence to be predicted, AGCAA…GCA, is converted into N 3-mer subsequences such as AGC and AUA. Embedding encoding can be performed on each 3-mer subsequence to obtain an N 3-mer embedding vector. The protein sequence to be predicted, MTAQDD…SYS, is converted into M amino acid 3-mer subsequences such as MTA and QDD. Embedding encoding can be performed on each amino acid 3-mer subsequence to obtain an M amino acid 3-mer embedding vector.

[0295] Step S1102. Input the obtained N base 3-mer embedding vectors and M amino acid 3-mer embedding vectors into the LSTM network respectively, and output the vector corresponding to each base 3-mer. and the vector corresponding to each amino acid 3-mer And a vector corresponding to N base 3-mers. Composition of the first RNA vector sequence Vectors corresponding to M amino acid 3-mers The first vector sequence of proteins

[0296] Step S1103. Obtain the association vector sequence of the first RNA vector sequence through the attention mechanism model. The associated vector sequence with the first vector sequence of the protein The associated vector sequence of the first RNA vector sequence The sequence is obtained by fusing with the first RNA vector sequence. The associated vector sequence of the first vector sequence of the protein. The sequence is obtained by fusing with the first vector sequence of the protein. Then, the sequence Input into an LSTM network to capture the correlation between the k-mer vectors of each base in the sequence, and output the sequence. Similarly, the sequence Input into an LSTM network to capture the correlation between the k-mer vectors of each amino acid in the sequence, and output the sequence.

[0297] Step S1104. Obtain the sequence through the attention mechanism model. self-correlated vector sequence The final representation v of the RNA sequence to be predicted is obtained by averaging the autocorrelation vector sequence. RNA Sequences are obtained through an attention mechanism model. self-correlated vector sequence The mean of the autocorrelation vector sequence is then calculated to obtain the final representation v of the protein sequence to be predicted. P v RNA and v P The features are concatenated to obtain the feature vector v to be predicted for interaction prediction.

[0298] Step S1105. Use the Softmax classifier to perform binary classification prediction on the feature vector v to be predicted, and obtain the predicted value of the interaction between the RNA sequence and the protein sequence to be predicted.

[0299] Step S1106. Output the predicted value of the interaction between the RNA sequence and the protein sequence to be predicted to the terminal device for the user to view.

[0300] In this exemplary embodiment of the present disclosure, at least one RNA sequence can be obtained, and protein sequences that interact with each input RNA sequence can be searched in a database. For example, after a user inputs at least one RNA sequence, each input RNA sequence can be combined with all protein sequences in the database to form several RNA-protein pairs. Further, the interaction of each RNA-protein pair can be predicted according to steps S220 to S250. Specifically, a recurrent neural network can be used to encode the RNA and protein sequences in each RNA-protein pair to obtain the corresponding first RNA vector sequence and first protein vector sequence. An attention mechanism model is used to obtain the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence. Based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence, a second RNA vector sequence and a second protein vector sequence are obtained. The second RNA vector sequence and the second protein vector sequence are then concatenated to obtain the feature vector to be predicted. Finally, a classifier can be used to classify and predict the feature vector to be predicted, obtaining the interaction prediction value for each RNA-protein pair. An interaction prediction value of 1 indicates that the RNA-protein pair interacts, and an interaction prediction value of 0 indicates that the RNA-protein pair does not interact. Then, all RNA-protein pairs with an interaction prediction value of 1 can be screened out, and the protein sequences in each RNA-protein pair can be output to the terminal device so that users can view the protein sequences that interact with the input RNA sequence.

[0301] Similarly, in the exemplary embodiments of this disclosure, at least one protein sequence can be obtained, and RNA sequences that interact with each input protein sequence can be searched in a database. For example, after a user inputs at least one protein sequence, each input protein sequence can be combined with all RNA sequences in the database to form several RNA-protein pairs. Further, the interaction of each RNA-protein pair can be predicted according to steps S220 to S250. Specifically, a recurrent neural network can be used to encode the RNA and protein sequences in each RNA-protein pair to obtain the corresponding first RNA vector sequence and first protein vector sequence. An attention mechanism model is used to obtain the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence. Based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence, a second RNA vector sequence and a second protein vector sequence are obtained. The second RNA vector sequence and the second protein vector sequence are then concatenated to obtain the feature vector to be predicted. Finally, a classifier can be used to classify and predict the feature vector to be predicted, obtaining the interaction prediction value for each RNA-protein pair. An interaction prediction value of 1 indicates that the RNA-protein pair interacts, and an interaction prediction value of 0 indicates that the RNA-protein pair does not interact. Then, all RNA-protein pairs with an interaction prediction value of 1 can be screened out, and the RNA sequence in each RNA-protein pair can be output to the terminal device so that the user can view the RNA sequences that interact with the input protein sequence.

[0302] In the RNA-protein interaction prediction method provided in the exemplary embodiments of this disclosure, the following steps are taken: First, an RNA sequence and a protein sequence to be predicted are obtained; the RNA sequence to be predicted is encoded to obtain a first RNA vector sequence; the protein sequence to be predicted is encoded to obtain a first protein vector sequence; an attention mechanism model is used to obtain the association vector sequences of the RNA first vector sequence and the protein first vector sequence; the interaction between the RNA sequence and the protein sequence to be predicted is determined based on the association vector sequences of the RNA first vector sequence and the protein first vector sequence. This disclosure utilizes an attention mechanism model to determine the association information between RNA and protein sequences. By fusing the RNA sequence, protein sequence, and association information, and introducing the fused sequence information to predict the interaction between RNA and protein sequences, the accuracy of RNA-protein interaction prediction can be improved.

[0303] It should be noted that although the steps of the method in this disclosure are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or a step may be broken down into multiple steps.

[0304] Furthermore, this example embodiment also provides an RNA-protein interaction prediction device. This device can be applied to a server or terminal device. (Reference) Figure 12 As shown, the RNA-protein interaction prediction device 1200 may include a data acquisition module 1210, a first data encoding module 1220, a second data encoding module 1230, an association information acquisition module 1240, and an interaction determination module 1250, wherein:

[0305] The data acquisition module 1210 is used to acquire RNA-protein pairs to be predicted;

[0306] The first data encoding module 1220 is used to encode the RNA sequence to be predicted to obtain a first RNA vector sequence;

[0307] The second data encoding module 1230 is used to encode the protein sequence to be predicted to obtain a first protein vector sequence.

[0308] The association information acquisition module 1240 is used to acquire the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence through an attention mechanism model;

[0309] The interaction determination module 1250 is used to determine the interaction between the RNA sequence and the protein sequence to be predicted based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence.

[0310] In one optional implementation, the first data encoding module 1220 includes:

[0311] The first sequence conversion module is used to convert the RNA sequence to be predicted into an N-base k-mer subsequence;

[0312] The first sequence encoding module is used to vectorize each base k-mer subsequence to obtain the first vector sequence of the RNA.

[0313] In one optional implementation, the first sequence encoding module includes:

[0314] The first sequence encoding unit is used to encode each base k-mer subsequence to obtain a first vector of N base k-mer subsequences;

[0315] The first vector operation unit is used to perform operations on the first vector of the N base k-mer subsequences using the first mapping matrix to obtain the second vector of the N base k-mer subsequences.

[0316] The first vector sequence determination unit is used to sequentially input the second vectors of the N base k-mer subsequences into a pre-trained recurrent neural network, output N base k-mer vectors, and form the first vector sequence of RNA from the N base k-mer vectors.

[0317] In one optional implementation, the second data encoding module 1230 includes:

[0318] The second sequence conversion module is used to convert the protein sequence to be predicted into an M-amino acid k-mer subsequence;

[0319] The second sequence encoding module is used to vectorize each amino acid k-mer subsequence to obtain the first vector sequence of the protein.

[0320] In one optional implementation, the second sequence encoding module includes:

[0321] The second sequence coding unit is used to encode each amino acid k-mer subsequence to obtain a first vector of M amino acid k-mer subsequences;

[0322] The second vector operation unit is used to perform operations on the first vector of the M amino acid k-mer subsequences using the second mapping matrix to obtain the second vector of the M amino acid k-mer subsequences.

[0323] The second vector sequence determination unit is used to sequentially input the second vectors of the M amino acid k-mer subsequences into a pre-trained recurrent neural network, output M amino acid k-mer vectors, and form the first vector sequence of the protein from the M amino acid k-mer vectors.

[0324] In one optional implementation, the association information acquisition module 1240 includes:

[0325] The RNA feature extraction unit is used to extract features from the first RNA vector sequence to obtain the first hidden RNA vector.

[0326] The protein feature extraction unit is used to extract features from the first vector sequence of the protein to obtain the first hidden vector of the protein.

[0327] The association information acquisition unit is used to perform operations on the first hidden vector of RNA and the first hidden vector of protein to obtain the association vector sequence of the first vector sequence of RNA and the association vector sequence of the first vector sequence of protein.

[0328] In one optional implementation, the first hidden RNA vector includes a first RNA vector, a second RNA vector, and a third RNA vector; the RNA feature extraction unit is used to perform operations on the first RNA vector sequence using a first query weight matrix to obtain the first RNA vector; to perform operations on the first RNA vector sequence using a first bond weight matrix to obtain the second RNA vector; and to perform operations on the first RNA vector sequence using a first value weight matrix to obtain the third RNA vector.

[0329] In one optional implementation, the first hidden protein vector includes a first protein vector, a second protein vector, and a third protein vector; the protein feature extraction unit is used to perform operations on the first protein vector sequence using a second query weight matrix to obtain the first protein vector; to perform operations on the first protein vector sequence using a second bond weight matrix to obtain the second protein vector; and to perform operations on the first protein vector sequence using a second value weight matrix to obtain the third protein vector.

[0330] In one optional implementation, the association information acquisition unit is used to calculate the similarity between the first RNA vector and the second protein vector to obtain a first RNA attention score; sum the first RNA attention score with the third protein vector to obtain an association vector sequence of the first RNA vector sequence; calculate the similarity between the first protein vector and the second RNA vector to obtain a first protein attention score; and sum the first protein attention score with the third RNA vector to obtain an association vector sequence of the first protein vector sequence.

[0331] In one alternative implementation, the interaction determination module 1250 includes:

[0332] The vector sequence determination unit is used to obtain the second RNA vector sequence and the second protein vector sequence based on the associated vector sequence of the first RNA vector sequence and the associated vector sequence of the first protein vector sequence.

[0333] An interaction determination unit is used to determine the interaction between the RNA sequence and the protein sequence to be predicted based on the second vector sequence of the RNA and the second vector sequence of the protein.

[0334] In one optional implementation, the vector sequence determination unit includes:

[0335] The first sequence splicing unit is used to splice the associated vector sequence of the first RNA vector sequence and the first RNA vector sequence to obtain the RNA fusion vector sequence;

[0336] The first RNA vector sequence determining unit is used to input the RNA fusion vector sequence into a pre-trained recurrent neural network and output the second RNA vector sequence.

[0337] The second sequence splicing unit is used to splice the associated vector sequence of the protein first vector sequence and the protein first vector sequence to obtain the protein fusion vector sequence.

[0338] The first protein vector sequence determining unit is used to input the protein fusion vector sequence into a pre-trained recurrent neural network and output the second protein vector sequence.

[0339] In one optional implementation, the vector sequence determination unit includes:

[0340] The first sequence splicing unit is used to splice the associated vector sequence of the first RNA vector sequence and the first RNA vector sequence to obtain the RNA fusion vector sequence;

[0341] The second sequence splicing unit is used to splice the associated vector sequence of the protein first vector sequence and the protein first vector sequence to obtain the protein fusion vector sequence.

[0342] The second RNA vector sequence determining unit is used to obtain the autocorrelation vector sequence of the RNA fusion vector sequence through an attention mechanism model, and to obtain the second RNA vector sequence based on the autocorrelation vector sequence of the RNA fusion vector sequence.

[0343] The second protein vector sequence determining unit is used to obtain the autocorrelation vector sequence of the protein fusion vector sequence through an attention mechanism model, and to obtain the second protein vector sequence based on the autocorrelation vector sequence of the protein fusion vector sequence.

[0344] In one optional implementation, the second RNA vector sequence determining unit includes:

[0345] The RNA feature extraction unit is used to extract features from the RNA fusion vector sequence to obtain the second hidden RNA vector.

[0346] The RNA self-association information determination unit is used to perform operations on the second hidden vector of the RNA to obtain the self-association vector sequence of the RNA fusion vector sequence;

[0347] The RNA self-association information processing unit is used to perform operations on the self-association vector sequence of the RNA fusion vector sequence to obtain the second RNA vector sequence.

[0348] In one optional implementation, the second hidden RNA vector includes a fourth RNA vector, a fifth RNA vector, and a sixth RNA vector; the RNA feature extraction unit is used to perform operations on the RNA fusion vector sequence using a third query weight matrix to obtain the fourth RNA vector; to perform operations on the RNA fusion vector sequence using a third bond weight matrix to obtain the fifth RNA vector; and to perform operations on the RNA fusion vector sequence using a third value weight matrix to obtain the sixth RNA vector.

[0349] In one optional implementation, the RNA self-association information determination unit is used to calculate the similarity between the fourth RNA vector and the fifth RNA vector to obtain the second RNA attention score; and to sum the sixth RNA vector based on the second RNA attention score to obtain the self-association vector sequence of the RNA fusion vector sequence.

[0350] In one optional implementation, the second protein vector sequence determining unit includes:

[0351] The protein feature extraction unit is used to extract features from the protein fusion vector sequence to obtain the second hidden vector of the protein.

[0352] The protein self-association information determination unit is used to perform calculations on the second hidden vector of the protein to obtain the self-association vector sequence of the protein fusion vector sequence;

[0353] The protein self-association information processing unit is used to perform calculations on the self-association vector sequence of the protein fusion vector sequence to obtain the second protein vector sequence.

[0354] In one optional implementation, the second hidden vector of the protein includes a fourth vector of the protein, a fifth vector of the protein, and a sixth vector of the protein; the protein feature extraction unit is used to perform operations on the protein fusion vector sequence using a fourth query weight matrix to obtain the fourth vector of the protein; to perform operations on the protein fusion vector sequence using a fourth bond weight matrix to obtain the fifth vector of the protein; and to perform operations on the protein fusion vector sequence using a fourth value weight matrix to obtain the sixth vector of the protein.

[0355] In one optional implementation, the protein self-association information determination unit is used to calculate the similarity between the fourth protein vector and the fifth protein vector to obtain a second protein attention score; and to sum the sixth protein vector based on the second protein attention score to obtain a self-association vector sequence of the protein fusion vector sequence.

[0356] In one alternative implementation, the interaction determination module 1250 includes:

[0357] The original feature vector determination unit is used to concatenate the RNA second vector sequence and the protein second vector sequence to obtain the feature vector to be predicted;

[0358] The prediction value determination unit is used to obtain the predicted interaction value between the RNA sequence and the protein sequence to be predicted based on the feature vector to be predicted;

[0359] An interaction determination unit is used to determine the interaction between the RNA sequence and the protein sequence to be predicted based on the interaction prediction value.

[0360] In one optional implementation, the interaction prediction value determination unit is used to input the feature vector to be predicted into the classifier and output the probability that there is an interaction between the RNA sequence and the protein sequence to be predicted.

[0361] In an optional embodiment, the RNA-protein interaction prediction device 1200 further includes:

[0362] The training data acquisition module is used to acquire a training dataset, which includes positive RNA-protein pairs and negative RNA-protein pairs.

[0363] The prediction output module is used to determine the interaction prediction value of each RNA-protein pair in the training dataset using the recurrent neural network and attention mechanism model;

[0364] The loss calculation module is used to calculate the interaction prediction value and label value of each RNA-protein pair in the training dataset using a loss function to obtain the corresponding loss value;

[0365] The model parameter adjustment module is used to adjust the model parameters of the recurrent neural network and the attention mechanism model according to the loss value.

[0366] In an optional embodiment, the RNA-protein interaction prediction device 1200 further includes:

[0367] The data output module is used to output the prediction results of the interaction between the RNA sequence and the protein sequence to be predicted.

[0368] The specific details of each module in the above-mentioned RNA-protein interaction prediction device have been described in detail in the corresponding RNA-protein interaction prediction methods, so they will not be repeated here.

[0369] The modules in the above-described device can be general-purpose processors, including central processing units (CPUs), network processors, etc.; they can also be digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. Each module can also be implemented using software, firmware, etc. The processors in the above-described device can be independent processors or integrated together.

[0370] Exemplary embodiments of this disclosure also provide a computer-readable storage medium having a program product stored thereon capable of implementing the methods described above in this specification. In some possible embodiments, various aspects of this disclosure may also be implemented as a program product including program code that, when run on an electronic device, causes the electronic device to perform the steps described in the "Exemplary Methods" section of this specification according to various exemplary embodiments of this disclosure. This program product may be a portable compact disc read-only memory (CD-ROM) including program code and may run on an electronic device, such as a personal computer. However, the program product of this disclosure is not limited thereto. In this document, the readable storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0371] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples (a non-exhaustive list) of readable storage media include: electrical connections having one or more wires, portable disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0372] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting programs for use by or in conjunction with an instruction execution system, apparatus, or device.

[0373] The program code contained on the readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0374] Program code for performing the operations of this disclosure can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing devices can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0375] Exemplary embodiments of this disclosure also provide an electronic device capable of implementing the above-described method. Referring below... Figure 13 To describe an electronic device 1300 according to such an exemplary embodiment of the present disclosure. Figure 13 The electronic device 1300 shown is merely an example and should not be construed as limiting the functionality and scope of the embodiments disclosed herein.

[0376] like Figure 13 As shown, the electronic device 1300 can be represented as a general-purpose computing device. The components of the electronic device 1300 may include, but are not limited to: at least one processing unit 1310, at least one storage unit 1320, a bus 1330 connecting different system components (including storage unit 1320 and processing unit 1310), and a display unit 1340.

[0377] Storage unit 1320 stores program code that can be executed by processing unit 1310, causing processing unit 1310 to perform the steps described in the "Exemplary Methods" section of this specification according to various exemplary embodiments of this disclosure. For example, processing unit 1310 can perform... Figures 2 to 11 Any one or more of the method steps.

[0378] Storage unit 1320 may include readable media in the form of volatile storage units, such as random access memory (RAM) 1321 and / or cache memory 1322, and may further include read-only memory (ROM) 1323.

[0379] Storage unit 1320 may also include a program / utility 1324 having a set (at least one) program module 1325, such program module 1325 including but not limited to: operating system, one or more application programs, other program modules and program data, each or some combination of these examples may include an implementation of a network environment.

[0380] Bus 1330 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.

[0381] Electronic device 1300 can also communicate with one or more external devices 1400 (e.g., keyboard, pointing device, Bluetooth device, etc.), one or more devices that enable a user to interact with electronic device 1300, and / or any device that enables electronic device 1300 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 1350. Furthermore, electronic device 1300 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 1360. As shown, network adapter 1360 communicates with other modules of electronic device 1300 via bus 1330. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 1300, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0382] In some embodiments, the RNA-protein interaction prediction method described herein can be executed by the processing unit 1310 of the electronic device. In some embodiments, the RNA sequence and protein sequence to be predicted, as well as the training dataset for training each model, can be input through the input interface 1350. For example, the RNA sequence and protein sequence to be predicted, as well as the training dataset for training each model, can be input through the user interface of the electronic device. In some embodiments, the prediction result of the interaction between the RNA sequence and protein sequence to be predicted can be output to an external device 1400 for user viewing through the output interface 1350.

[0383] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, terminal device, or network device, etc.) to execute the method according to the exemplary embodiments of this disclosure.

[0384] Furthermore, the above figures are merely illustrative representations of the processes included in the methods according to exemplary embodiments of this disclosure, and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Additionally, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0385] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0386] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A method for predicting RNA-protein interactions, characterized in that, include: Obtain the RNA and protein sequences to be predicted; The RNA sequence to be predicted is encoded to obtain the first RNA vector sequence; The protein sequence to be predicted is encoded to obtain the first protein vector sequence; The association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence are obtained through an attention mechanism model; The interaction between the RNA sequence and the protein sequence to be predicted is determined based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence. Wherein, obtaining the association vector sequence of the RNA first vector sequence and the association vector sequence of the protein first vector sequence through the attention mechanism model includes: Feature extraction is performed on the first RNA vector sequence to obtain the first hidden RNA vector; Feature extraction is performed on the first vector sequence of the protein to obtain the first hidden vector of the protein; The first hidden vector of RNA and the first hidden vector of protein are processed to obtain the correlation vector sequence of the first vector sequence of RNA and the correlation vector sequence of the first vector sequence of protein.

2. The method for predicting RNA-protein interactions according to claim 1, characterized in that, The process of encoding the RNA sequence to be predicted to obtain the first RNA vector sequence includes: The RNA sequence to be predicted is converted into an N-base k-mer subsequence; Vectorize each base k-mer subsequence to obtain the first vector sequence of the RNA.

3. The method for predicting RNA-protein interactions according to claim 2, characterized in that, The vectorization of each base k-mer subsequence yields the first vector sequence of the RNA, including: Encode each base k-mer subsequence to obtain a first vector of N base k-mer subsequences; The first vector of the N base k-mer subsequences is obtained by performing operations on the first vector of the N base k-mer subsequences using the first mapping matrix; The second vector of the N base k-mer subsequences is input into a pre-trained recurrent neural network, which outputs N base k-mer vectors, and the N base k-mer vectors form the first RNA vector sequence.

4. The method for predicting RNA-protein interactions according to claim 1, characterized in that, The process of encoding the protein sequence to be predicted to obtain a first protein vector sequence includes: The protein sequence to be predicted is converted into an M-amino acid k-mer subsequence; Vectorize each amino acid k-mer subsequence to obtain the first vector sequence of the protein.

5. The method for predicting RNA-protein interactions according to claim 4, characterized in that, The vectorization of each amino acid k-mer subsequence yields the first vector sequence of the protein, including: Encode each amino acid k-mer subsequence to obtain a first vector of M amino acid k-mer subsequences; The second vector of the M amino acid k-mer subsequences is obtained by operating on the first vector of the M amino acid k-mer subsequences using the second mapping matrix; The second vectors of the M amino acid k-mer subsequences are sequentially input into a pre-trained recurrent neural network, which outputs M amino acid k-mer vectors, and the M amino acid k-mer vectors form the first vector sequence of the protein.

6. The method for predicting RNA-protein interactions according to claim 1, characterized in that, The first hidden RNA vector includes a first RNA vector, a second RNA vector, and a third RNA vector; the feature extraction of the first RNA vector sequence to obtain the first hidden RNA vector includes: The first RNA vector is obtained by performing operations on the first RNA first vector sequence using the first query weight matrix; The first RNA vector is obtained by performing operations on the first RNA vector sequence using the first bond weight matrix; The first RNA vector sequence is processed using the first-value weight matrix to obtain the third RNA vector.

7. The method for predicting RNA-protein interactions according to claim 1, characterized in that, The first latent vector of the protein includes a first protein vector, a second protein vector, and a third protein vector; the feature extraction of the first protein vector sequence to obtain the first latent vector of the protein includes: The protein first vector is obtained by performing operations on the protein first vector sequence using the second query weight matrix; The second vector of the protein is obtained by operating on the first vector sequence of the protein using the second bond weight matrix; The third vector of the protein is obtained by operating on the first vector sequence of the protein using the second-value weight matrix.

8. The method for predicting RNA-protein interactions according to claim 1, characterized in that, The first hidden RNA vector includes a first RNA vector, a second RNA vector, and a third RNA vector; the feature extraction of the first RNA vector sequence to obtain the first hidden RNA vector includes: The first RNA vector is obtained by performing operations on the first RNA first vector sequence using the first query weight matrix; The first RNA vector is obtained by performing operations on the first RNA vector sequence using the first bond weight matrix; The first RNA vector sequence is processed using the first-value weight matrix to obtain the third RNA vector; The first latent vector of the protein includes a first protein vector, a second protein vector, and a third protein vector; the feature extraction of the first protein vector sequence to obtain the first latent vector of the protein includes: The protein first vector is obtained by performing operations on the protein first vector sequence using the second query weight matrix; The second vector of the protein is obtained by operating on the first vector sequence of the protein using the second bond weight matrix; The third vector of the protein is obtained by operating on the first vector sequence of the protein using the second-value weight matrix; The step of performing operations on the first hidden vector of RNA and the first hidden vector of protein to obtain the association vector sequence of the first vector sequence of RNA and the association vector sequence of the first vector sequence of protein includes: The similarity between the first RNA vector and the second protein vector is calculated to obtain the first RNA attention score; The third vector of the protein is summed based on the first attention score of the RNA to obtain the associated vector sequence of the first RNA vector sequence; Calculate the similarity between the first protein vector and the second RNA vector to obtain the first protein attention score; The RNA third vector is summed based on the protein's first attention score to obtain the associated vector sequence of the protein's first vector sequence.

9. The method for predicting RNA-protein interactions according to claim 1, characterized in that, Determining the interaction between the RNA sequence and the protein sequence to be predicted based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence includes: The second RNA vector sequence and the second protein vector sequence are obtained based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence. The interaction between the RNA sequence and the protein sequence to be predicted is determined based on the second vector sequence of the RNA and the second vector sequence of the protein.

10. The method for predicting RNA-protein interactions according to claim 9, characterized in that, The step of obtaining the second RNA vector sequence and the second protein vector sequence based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence includes: The associated vector sequence of the first RNA vector sequence and the first RNA vector sequence are concatenated to obtain the RNA fusion vector sequence; The RNA fusion vector sequence is input into a pre-trained recurrent neural network, which outputs the second RNA vector sequence. The associated vector sequence of the protein first vector sequence and the protein first vector sequence are concatenated to obtain the protein fusion vector sequence; The protein fusion vector sequence is input into a pre-trained recurrent neural network, which outputs the second protein vector sequence.

11. The method for predicting RNA-protein interactions according to claim 9, characterized in that, The step of obtaining the second RNA vector sequence and the second protein vector sequence based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence includes: The associated vector sequence of the first RNA vector sequence and the first RNA vector sequence are concatenated to obtain the RNA fusion vector sequence; The associated vector sequence of the protein first vector sequence and the protein first vector sequence are concatenated to obtain the protein fusion vector sequence; The autocorrelation vector sequence of the RNA fusion vector sequence is obtained through an attention mechanism model, and the second RNA vector sequence is obtained based on the autocorrelation vector sequence of the RNA fusion vector sequence. The autocorrelation vector sequence of the protein fusion vector sequence is obtained through an attention mechanism model, and the second vector sequence of the protein is obtained based on the autocorrelation vector sequence of the protein fusion vector sequence.

12. The method for predicting RNA-protein interactions according to claim 11, characterized in that, The step of obtaining the autocorrelation vector sequence of the RNA fusion vector sequence through an attention mechanism model, and obtaining the second RNA vector sequence based on the autocorrelation vector sequence of the RNA fusion vector sequence, includes: Feature extraction is performed on the RNA fusion vector sequence to obtain the second hidden RNA vector; The second hidden vector of the RNA is processed to obtain the self-correlated vector sequence of the RNA fusion vector sequence; The self-correlation vector sequence of the RNA fusion vector sequence is processed to obtain the second RNA vector sequence.

13. The method for predicting RNA-protein interactions according to claim 12, characterized in that, The second hidden RNA vector includes a fourth RNA vector, a fifth RNA vector, and a sixth RNA vector; the feature extraction of the RNA fusion vector sequence to obtain the second hidden RNA vector includes: The RNA fusion vector sequence is processed using the third query weight matrix to obtain the fourth RNA vector; The RNA fusion vector sequence is processed using the third bond weight matrix to obtain the fifth RNA vector; The RNA fusion vector sequence is processed using a third-value weight matrix to obtain the sixth RNA vector.

14. The method for predicting RNA-protein interactions according to claim 13, characterized in that, The step of performing operations on the second hidden vector of the RNA to obtain the autocorrelation vector sequence of the RNA fusion vector sequence includes: Calculate the similarity between the fourth and fifth RNA vectors to obtain the second RNA attention score. The RNA sixth vector is summed based on the RNA second attention score to obtain the self-correlated vector sequence of the RNA fusion vector sequence.

15. The method for predicting RNA-protein interactions according to claim 11, characterized in that, The step of obtaining the autocorrelation vector sequence of the protein fusion vector sequence through an attention mechanism model, and obtaining the second protein vector sequence based on the autocorrelation vector sequence of the protein fusion vector sequence, includes: Feature extraction is performed on the protein fusion vector sequence to obtain the second hidden vector of the protein; The second hidden vector of the protein is processed to obtain the self-correlated vector sequence of the protein fusion vector sequence; The self-correlation vector sequence of the protein fusion vector sequence is processed to obtain the second vector sequence of the protein.

16. The method for predicting RNA-protein interactions according to claim 15, characterized in that, The second hidden vector of the protein includes a fourth vector, a fifth vector, and a sixth vector; the feature extraction of the protein fusion vector sequence to obtain the second hidden vector of the protein includes: The protein fusion vector sequence is processed using the fourth query weight matrix to obtain the fourth protein vector; The protein fusion vector sequence is processed using the fourth bond weight matrix to obtain the fifth protein vector; The protein fusion vector sequence is processed using the fourth-value weight matrix to obtain the sixth vector of the protein.

17. The method for predicting RNA-protein interactions according to claim 16, characterized in that, The step of performing operations on the second hidden vector of the protein to obtain the self-correlated vector sequence of the protein fusion vector sequence includes: Calculate the similarity between the fourth and fifth vectors of the protein to obtain the second attention score of the protein; The sixth vector of the protein is summed based on the second attention score of the protein to obtain the self-correlated vector sequence of the protein fusion vector sequence.

18. The method for predicting RNA-protein interactions according to claim 9, characterized in that, The step of determining the interaction between the RNA sequence and the protein sequence to be predicted based on the second vector sequence of the RNA and the second vector sequence of the protein includes: The RNA second vector sequence and the protein second vector sequence are concatenated to obtain the feature vector to be predicted; The predicted interaction values ​​between the RNA sequence and the protein sequence are obtained based on the feature vector to be predicted. The interaction between the RNA sequence and the protein sequence to be predicted is determined based on the interaction prediction value.

19. The method for predicting RNA-protein interactions according to claim 18, characterized in that, The step of obtaining the predicted interaction value between the RNA sequence and the protein sequence based on the predicted feature vector includes: The feature vector to be predicted is input into the classifier, which outputs the probability that there is an interaction between the RNA sequence and the protein sequence to be predicted.

20. The method for predicting RNA-protein interactions according to any one of claims 1 to 19, characterized in that, The method further includes: Obtain a training dataset, which includes positive RNA-protein pairs and negative RNA-protein pairs; The interaction prediction values ​​for each RNA-protein pair in the training dataset were determined using a recurrent neural network and attention mechanism model. The loss function is used to calculate the predicted interaction value and label value of each RNA-protein pair in the training dataset to obtain the corresponding loss value; The model parameters of the recurrent neural network and attention mechanism model are adjusted based on the loss value.

21. The method for predicting RNA-protein interactions according to claim 1, characterized in that, The method further includes: Output the predicted results of the interaction between the RNA sequence and the protein sequence to be predicted.

22. An RNA-protein interaction prediction device, characterized in that, include: The data acquisition module is used to acquire the RNA and protein sequences to be predicted. The first data encoding module is used to encode the RNA sequence to be predicted to obtain the first RNA vector sequence; The second data encoding module is used to encode the protein sequence to be predicted to obtain the first protein vector sequence. The association information acquisition module is used to acquire the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence through an attention mechanism model; wherein, the association information acquisition module is configured to: extract features from the first RNA vector sequence to obtain a first RNA hidden vector; extract features from the first protein vector sequence to obtain a first protein hidden vector; and perform operations on the first RNA hidden vector and the first protein hidden vector to obtain the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence; An interaction determination module is used to determine the interaction between the RNA sequence and the protein sequence to be predicted based on the association vector sequence of the first RNA vector sequence and the association vector sequence of the first protein vector sequence.

23. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method according to any one of claims 1-21.

24. An electronic device, characterized in that, include: processor; as well as Memory for storing the executable instructions of the processor; The processor is configured to execute the method of any one of claims 1-21 by executing the executable instructions.

Images