Method and device for searching ribonucleic acid (RNA) structure and electronic equipment

Abstract

The invention provides a ribonucleic acid (RNA) structure searching method and device and electronic equipment, and relates to the technical field of artificial intelligence, biological calculation and the like. According to the specific implementation scheme, the method comprises the steps that a first discretization sequence of a first RNA structure to be searched is obtained and serves as a query sequence; based on the query sequence, performing index search in a pre-constructed RNA structure database, and determining a hit index sequence; and according to the query sequence and the hit index sequence, carrying out structure search in the RNA structure database, and determining at least one target RNA structure corresponding to the first RNA structure. Therefore, the target RNA structure is determined by performing discretization processing on the RNA structure and performing structure search in the pre-constructed RNA structure database, so that the complexity of obtaining the RNA structure is reduced, and the efficiency and accuracy of obtaining the RNA structure are improved.

Description

Technical Field

[0001] The present disclosure relates to the field of deep learning technology, and more specifically to the field of artificial intelligence and biological computing, and in particular to a method, device and electronic device for searching RNA structure. Background Art

[0002] Ribonucleic Acid (RNA) structure search technology has important value and broad application prospects, but it also faces certain challenges. In related technologies, structure-based search and comparison are often performed directly to determine the search results of RNA structure. However, the above methods are often highly complex and often lead to low efficiency and accuracy in searching for RNA structure. Therefore, how to reduce the complexity of obtaining RNA structure while improving the efficiency and accuracy of obtaining RNA structure has become a problem that needs to be solved urgently. Summary of the invention

[0003] The present invention provides a method, device, electronic device, storage medium and program product for searching RNA structure.

[0004] According to one aspect of the present disclosure, a method for searching for ribonucleic acid (RNA) structure is provided, comprising: determining a reconstruction loss based on the sample RNA structure and the reduced RNA structure; determining a quantization loss based on the sample latent space representation and the sample discretization representation; and adjusting model parameters of the VQ-VAE model based on the reconstruction loss and the quantization loss.

[0005] According to another aspect of the present disclosure, a device for searching for ribonucleic acid (RNA) structures is provided, comprising: an acquisition module for acquiring a first discretized sequence of a first RNA structure to be searched as a query sequence; a determination module for performing an index search in a pre-constructed RNA structure database based on the query sequence to determine a hit index sequence; and a search module for performing a structure search in the RNA structure database based on the query sequence and the hit index sequence to determine at least one target RNA structure corresponding to the first RNA structure.

[0006] According to another aspect of the present disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method for searching for ribonucleic acid RNA structure described in one aspect of the present disclosure.

[0007] According to another aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to enable the computer to execute the method for searching for ribonucleic acid RNA structure according to one aspect of the present disclosure.

[0008] According to another aspect of the present disclosure, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the steps of the method for searching for ribonucleic acid RNA structure according to one aspect of the present disclosure.

[0009] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure.

[0011] Figure 1 is a schematic flow chart of a method for searching for RNA structure according to the first embodiment of the present disclosure;

[0012] Figure 2 A schematic diagram of a flow chart of a method for searching for RNA structures according to a second embodiment of the present disclosure;

[0013] Figure 3 A schematic diagram of the training process of the VQ-VAE model in the method for searching for RNA structure according to the third embodiment of the present disclosure;

[0014] Figure 4 A block diagram of a device for searching RNA structures for RNA according to an embodiment of the present disclosure;

[0015] Figure 5 A block diagram of an electronic device used to implement an embodiment of the present disclosure. DETAILED DESCRIPTION

[0016] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for the sake of clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description.

[0017] Artificial Intelligence (AI) is a technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. At present, AI technology has the advantages of high automation, high precision and low cost, and has been widely used.

[0018] Deep Learning (DL) is a new research direction in the field of Machine Learning (ML). It is a science that studies the inherent laws and representation levels of sample data, enabling machines to have analytical learning capabilities like humans and to recognize data such as text, images, and sounds. It is widely used in speech and image recognition.

[0019] Biocomputing refers to a new computing model developed by using the inherent information processing mechanism of biological systems. Biocomputing research includes two aspects: devices and systems. The ordered system formed by organic (or biological) materials at the molecular scale provides the basic unit for information detection, processing, transmission and storage through physical and chemical processes at the molecular level. It is called a molecular device. The structure and computing principle of the biocomputing system are different from those of traditional computing systems. Its structure is generally parallel and distributed. Information storage is often a combination of short-term memory and long-term memory, and is completed through learning.

[0020] The following describes the method for searching for RNA structure in an embodiment of the present disclosure in conjunction with the accompanying drawings.

[0021] Figure 1 It is a flowchart of the method for searching RNA structure according to the first embodiment of the present disclosure.

[0022] It should be noted that the execution subject of the RNA structure search method of this embodiment is the RNA structure search device, which can be a hardware device or software in a hardware device, etc. The hardware device is, for example, a terminal device, a server, etc.

[0023] like Figure 1 As shown, the method for searching RNA structure provided by the embodiment of the present disclosure may include the following steps:

[0024] S101, obtaining a first discretized sequence of a first RNA structure to be searched as a query sequence.

[0025] The first RNA structure is any RNA structure that needs to be searched.

[0026] It should be noted that the present disclosure does not limit the specific method for obtaining the first discretized sequence of the first RNA structure to be searched, and it can be selected according to actual conditions.

[0027] Optionally, first structural information of the first RNA structure may be acquired, and a first discretized sequence of the first RNA structure may be determined according to the first structural information of the first RNA structure.

[0028] It should be noted that the first structural information includes but is not limited to the first coordinate information of the first RNA structure, the first conformation information between each nucleotide in the RNA structure and the upstream nucleotide, and the first position relationship and first pairing relationship between each nucleotide in the RNA structure and the nearest adjacent nucleotide in space.

[0029] For example, the first structural information of the first RNA structure can be input into a pre-trained first vector quantized variational autoencoder (VQ-VAE) model, and the encoder in the first VQ-VAE model encodes the first structural information to obtain a first latent space representation. The quantization layer in the first VQ-VAE model quantizes the first latent space representation, and matches the first quantized vector in a discretized vocabulary to obtain a first discretized sequence of the first RNA structure.

[0030] The first discretization sequence of the first RNA structure includes state codes corresponding to the codebook vector matching the first quantization vector in the discretization vocabulary.

[0031] It should be noted that, through the pre-trained first VQ-VAE model, the structure of the first RNA structure can be discretized to obtain a first discretized sequence of the first RNA structure.

[0032] It should be noted that the present disclosure does not limit the training process of the VQ-VAE model, which can be selected according to actual conditions.

[0033] Optionally, the sample RNA structure and sample structure information of the sample RNA structure can be obtained, and based on the sample structure information, a deep learning algorithm is used to iteratively train the initial VQ-VAE model to obtain a pre-trained VQ-VAE model.

[0034] S102, based on the query sequence, an index search is performed in a pre-constructed RNA structure database to determine a hit index sequence.

[0035] In the disclosed embodiment, a known RNA structure set can be obtained, and an RNA structure database can be constructed based on multiple second RNA structures included in the RNA structure set, and a second discretized sequence of the second RNA structure can be obtained, and index information of the RNA structure database can be constructed according to the second discretized sequence.

[0036] Optionally, index information of the RNA structure database can be obtained, the index information includes an index sequence of each cluster in the cluster set, a first similarity between the query sequence and the index sequence of the cluster is obtained, and a hit index sequence is determined from the index sequence of the cluster based on the first similarity.

[0037] In the embodiment of the present disclosure, after the first similarity is obtained, a hit index sequence may be determined from the index sequence of the cluster according to the first similarity.

[0038] Optionally, the first similarities may be sorted in descending order, and a hit index sequence may be determined from the index sequence of the cluster according to the sorting result.

[0039] Optionally, a preset first similarity threshold may be obtained, and a hit index sequence may be determined from the index sequence of the cluster according to the first similarity and the first similarity threshold.

[0040] S103, performing a structure search in an RNA structure database according to the query sequence and the hit index sequence to determine at least one target RNA structure corresponding to the first RNA structure.

[0041] In an embodiment of the present disclosure, at least one target cluster is determined from a cluster set of an RNA structure database based on a hit index sequence, a second similarity between a query sequence and a second discretized sequence in the target cluster is obtained, and based on the second similarity, at least one target RNA structure corresponding to the first RNA structure is determined from the second RNA structure within the target RNA structure group associated with the target cluster.

[0042] Optionally, the second similarities between the query sequence and the second discretized sequence in the target cluster may be sorted in descending order, and at least one target RNA structure corresponding to the first RNA structure may be determined from the second RNA structures in the target RNA structure group associated with the target cluster according to the sorting result.

[0043] For example, the second similarities between the query sequence and the second discretized sequence in the target cluster can be sorted in descending order, and the second RNA structure in the target RNA structure group associated with the top N target clusters can be determined as the target RNA structure, where N is a positive integer.

[0044] Optionally, a preset second similarity threshold may be obtained, and at least one target RNA structure corresponding to the first RNA structure may be determined from the second RNA structure according to the second similarity between the query sequence and the second discretized sequence in the target cluster and the second similarity threshold.

[0045] For example, if the second similarity between the query sequence and the second discretized sequence in the target cluster is greater than the second similarity threshold, the second RNA structure in the target RNA structure group associated with the target cluster greater than the second similarity threshold is determined as the target RNA structure corresponding to the first RNA structure.

[0046] In summary, the method for searching for ribonucleic acid RNA structure of the embodiment of the present invention obtains a first discretized sequence of the first RNA structure to be searched as a query sequence, performs an index search in a pre-constructed RNA structure database based on the query sequence, determines a hit index sequence, and performs a structure search in the RNA structure database according to the query sequence and the hit index sequence to determine at least one target RNA structure corresponding to the first RNA structure. Therefore, the present invention reduces the complexity of obtaining the RNA structure while improving the efficiency and accuracy of obtaining the RNA structure by discretizing the RNA structure and performing a structure search in a pre-constructed RNA structure database to determine the target RNA structure.

[0047] Figure 2 FIG. 1 is a flow chart of a method for searching for RNA structure according to the second embodiment of the present disclosure. Figure 2 As shown, based on the above embodiment, the method for searching RNA structure of the embodiment of the present disclosure may specifically include the following steps:

[0048] S201, obtaining a known RNA structure set, wherein the RNA structure set includes a plurality of second RNA structures, and constructing an RNA structure database based on the plurality of second RNA structures.

[0049] S202, obtaining a second discretized sequence of a second RNA structure.

[0050] It should be noted that the present disclosure does not limit the specific method for obtaining the second discretized sequence of the second RNA structure, which can be selected according to actual conditions.

[0051] Optionally, second structural information of the second RNA structure may be acquired, and a second discretized sequence of the second RNA structure may be determined according to the second structural information of the second RNA structure.

[0052] In the embodiments of the present disclosure, the second coordinate information of the second RNA structure can be obtained, the upstream nucleotide and the spatially nearest adjacent nucleotide of each nucleotide in the second RNA structure can be determined, the second conformation information between the nucleotide and the upstream nucleotide can be obtained, the second positional relationship between the nucleotide and the adjacent nucleotide can be obtained, and the second pairing relationship between the nucleotide and the adjacent nucleotide can be obtained. According to the second coordinate information, the second conformation information, the second positional relationship and the second pairing relationship, the second structural information of the second RNA structure is generated.

[0053] The specific process of obtaining the second coordinate information, the second conformation information, the second position relationship and the second pairing relationship proposed in the present disclosure is explained below.

[0054] For example, the second coordinate information can be obtained from the Protein Data Bank (PDB), where the coordinate information of the second RNA structure includes but is not limited to the three-dimensional coordinate information, bond length, bond angle, dihedral angle and other coordinate information of each nucleotide in the second.

[0055] For example, the second conformational information includes, but is not limited to, conformational information such as the bond length, bond angle, and dihedral angle between the nucleotide and the upstream nucleotide, wherein the bond length involves the bond length between the phosphate group and the two connected ribose sugars, i.e., the bond length between O5'i+1 (the 5' oxygen atom of the i+1 nucleotide) and C5'i+1 (the 5' carbon atom of the i+1 nucleotide), and the distance between O5'i+1 and C5'i+1 can be calculated based on the three-dimensional coordinate information to obtain the bond length; the bond angle involves the bond angle between the phosphate group and the two connected ribose sugars, and the bond angle θ POC is the bond angle between Pi+1 (phosphorus atom of phosphate group)-O5'+1-C5'i+1, bond angle θ OCC It is the bond angle between O5'+1-C5'i+1-C4'+1; for dihedral angles, it involves the dihedral angles of the phosphate group and the adjacent ribose or the rotation of the base, that is, the dihedral angle α is the angle of O3'-P-O5'-C5', the dihedral angle β is the angle of C3'-O3'-P-O5', the dihedral angle γ is the angle of C2'-C3'-O3'-P, the dihedral angle ε is the angle of C2'-C3'-O3'-P, the dihedral angle ζ is the angle of C4'-C3'-O3'-P, the dihedral angle ν is the angle of C4'-C5'-O5'-P, and the rotation of χ involving the base is N1 / N9-C1'-N9 / C4-C5.

[0056] For example, for the second position relationship, the distance r ij and the rotation angle ω ij And two direction vectors rij and rji are defined, where the distance r ijis the straight-line distance between the centers of two bases (or reference points, such as the C1' position), and the rotation angle ω ij The surrounding distance r ij The rotation angle describes the rotation of the two base planes relative to each other. The direction vector rij is the direction vector pointing to the center of the second base (base j) in the local coordinate system of the first base (base i). The direction vector rji is the direction vector corresponding to the direction vector rij, which is the direction pointing to the center of the first base (base i) in the local coordinate system of the second base (base j).

[0057] For example, for the second pairing relationship, it can be determined whether there is a pairing relationship between the nucleotide and the adjacent nucleotide, wherein the pairing relationship includes Watson-Crick pairing, for example: adenine (A) and uracil (U) pairing, i.e. AU pairing, guanine (G) and cytosine (C) pairing, i.e. GC pairing, and the pairing relationship also includes atypical pairing, for example: wobble pairing, i.e. Wobble pairing.

[0058] In an embodiment of the present disclosure, after obtaining the second structural information, a second latent space representation of the second RNA structure can be obtained based on the second structural information of the second RNA structure, the second latent space representation can be quantized to obtain a second quantization vector, and based on the second quantization vector, a second discretized sequence of the second RNA structure can be obtained from a pre-constructed discretization vocabulary.

[0059] Optionally, the second structural information can be input into a pre-trained first vector quantization variational autoencoder VQ-VAE model, and the encoder in the first VQ-VAE model encodes the second structural information to obtain a second latent space representation. The quantization layer in the first VQ-VAE model quantizes the second latent space representation, and the second quantized vector is matched in a discretized vocabulary to obtain a second discretized sequence of the second RNA structure.

[0060] For example, the second structural information of the second RNA structure can be input into a pre-trained first vector quantization variational autoencoder VQ-VAE model, and the encoder in the first VQ-VAE model encodes the second structural information to obtain a second latent space representation. The quantization layer in the first VQ-VAE model quantizes the second latent space representation, and matches the second quantized vector in the discretized vocabulary to obtain a second discretized sequence of the second RNA structure.

[0061] The discretized vocabulary includes a plurality of codebook vectors and state codes corresponding to the codebook vectors.

[0062] The second discretization sequence of the second RNA structure includes state codes corresponding to the codebook vector matching the second quantization vector in the discretization vocabulary.

[0063] S203: construct index information of an RNA structure database according to the second discretized sequence.

[0064] In the disclosed embodiment, after the second discretization sequence is obtained, index information of the RNA structure database can be constructed according to the second discretization sequence.

[0065] Optionally, a similarity matrix between the second discretized sequences of the multiple second RNA structures can be determined, and the second discretized sequences of the multiple second RNA structures can be clustered according to the similarity matrix to obtain a cluster set of the RNA structure database, and a central discretized sequence can be determined from the second discretized sequences included in each cluster in the cluster set as the index sequence of the cluster, and index information of the RNA structure database can be constructed based on the index sequence of the cluster.

[0066] For example, a similarity algorithm can be used to calculate a similarity matrix between second discretized sequences of multiple second RNA structures, wherein each element in the similarity matrix is ​​the distance (similarity) between every two discretized sequences in the feature space. The similarity matrix can be used as input, and the second discretized sequences can be clustered using a clustering algorithm, that is, the second discretized sequences are grouped into clusters to obtain a cluster set, and the sequence with the highest average similarity between the discrete sequences in each cluster and the other remaining discrete sequences is determined as the central discretized sequence, the central discretized sequence is used as the index sequence of the cluster, and based on the index sequence of the cluster, the index information of the RNA structure database is constructed.

[0067] S204, obtaining a first discretized sequence of a first RNA structure to be searched as a query sequence.

[0068] It should be noted that the present disclosure does not limit the specific method for obtaining the first discretized sequence of the first RNA structure, which can be selected according to actual conditions.

[0069] Optionally, first structural information of the first RNA structure may be acquired, and a first discretized sequence of the first RNA structure may be determined according to the first structural information of the first RNA structure.

[0070] In the embodiments of the present disclosure, the first coordinate information of the first RNA structure can be obtained, the upstream nucleotide and the spatially nearest adjacent nucleotide of each nucleotide in the first RNA structure can be determined, the first conformation information between the nucleotide and the upstream nucleotide can be obtained, the first positional relationship between the nucleotide and the adjacent nucleotide can be obtained, and the first pairing relationship between the nucleotide and the adjacent nucleotide can be obtained. According to the first coordinate information, the first conformation information, the first positional relationship and the first pairing relationship, the first structural information of the first RNA structure is generated.

[0071] It should be noted that the specific process of obtaining the first coordinate information, the first conformation information, the first position relationship and the first pairing relationship can be found in the above embodiments and will not be described in detail here.

[0072] In an embodiment of the present disclosure, after obtaining the first structural information of the first RNA structure, a first latent space representation of the first RNA structure can be obtained based on the first structural information, the first latent space representation can be quantized to obtain a first quantization vector, and based on the first quantization vector, a first discretization sequence of the first RNA structure can be obtained from a pre-constructed discretization vocabulary.

[0073] Optionally, the first structural information can be input into a pre-trained first vector quantization variational autoencoder VQ-VAE model, and the encoder in the first VQ-VAE model encodes the first structural information to obtain a first latent space representation. The quantization layer in the first VQ-VAE model quantizes the first latent space representation, and the first quantized vector is matched in a discretized vocabulary to obtain a first discretized sequence of the first RNA structure.

[0074] For example, the first structural information of the first RNA structure can be input into a pre-trained first vector quantization variational autoencoder VQ-VAE model, and the encoder in the first VQ-VAE model encodes the first structural information to obtain a first latent space representation. The quantization layer in the first VQ-VAE model quantizes the first latent space representation, and matches the first quantized vector in a discretized vocabulary to obtain a first discretized sequence of the first RNA structure.

[0075] In the above embodiment, step S102 of “searching in a pre-constructed RNA structure database based on the query sequence to obtain a hit index sequence” may specifically include the following steps S205 and S207.

[0076] S205, obtaining index information of the RNA structure database, where the index information includes an index sequence of each cluster in the cluster set.

[0077] S206: Obtain a first similarity between the query sequence and the index sequence of the cluster.

[0078] It should be noted that the present disclosure does not limit the specific method for obtaining the first similarity between the query sequence and the index sequence of the cluster, and it can be selected according to actual conditions.

[0079] Optionally, the first similarity between the query sequence and the index sequence of the cluster may be calculated by a comparison algorithm to obtain the first similarity between the query sequence and the index sequence of the cluster.

[0080] S207: Determine a hit index sequence from the index sequences of the cluster according to the first similarity.

[0081] In the embodiment of the present disclosure, after the first similarity is obtained, a hit index sequence may be determined from the index sequence of the cluster according to the first similarity.

[0082] For example, the first similarities may be sorted in descending order, and the index sequences corresponding to the first M first similarities in the sorting result are taken as hit index sequences, where M is a positive integer. The present disclosure does not limit the setting of M, and it may be set according to actual conditions.

[0083] In the above embodiment, step S103 of "determining at least one target RNA structure corresponding to the first RNA structure from an RNA structure database according to the query sequence and the hit index sequence" may specifically include the following steps S208 and S2011.

[0084] S208, obtaining a cluster set of the RNA structure database.

[0085] S209: Determine at least one target cluster from the cluster set based on the hit index sequence.

[0086] It should be noted that after the hit index sequence is obtained, a query may be performed from the cluster set based on the hit index sequence to determine at least one target cluster.

[0087] S2010, obtaining a second similarity between the query sequence and a second discretized sequence in the target cluster.

[0088] It should be noted that the present disclosure does not limit the specific method for obtaining the second similarity between the query sequence and the second discretized sequence in the target cluster, which can be selected according to actual conditions.

[0089] Optionally, the second similarity between the query sequence and the second discretized sequence in the target cluster may be calculated by a comparison algorithm to obtain the second similarity between the query sequence and the second discretized sequence in the target cluster.

[0090] S2011, determining a target RNA structure group associated with the target cluster, and determining at least one target RNA structure corresponding to the first RNA structure from the second RNA structures in the target RNA structure group based on the second similarity.

[0091] In the embodiment of the present disclosure, after the second similarity is obtained, at least one target RNA structure corresponding to the first RNA structure can be determined from the second RNA structures in the target RNA structure group based on the second similarity.

[0092] For example, the second similarities can be sorted in descending order, and the second RNA structures corresponding to the first N second similarities in the sorting results are taken as the target RNA structure, where N is a positive integer. The present disclosure does not limit the setting of N, and it can be set according to actual conditions.

[0093] In summary, the method for searching RNA structure of the ribonucleic acid according to the embodiment of the present invention obtains a known RNA structure set, wherein the RNA structure set includes multiple second RNA structures, and based on the multiple second RNA structures, constructs an RNA structure database, obtains a second discretized sequence of the second RNA structure, and constructs index information of the RNA structure database according to the second discretized sequence, obtains a first discretized sequence of the first RNA structure to be searched as a query sequence, obtains index information of the RNA structure database, the index information includes an index sequence of each cluster in the cluster set, obtains a first similarity between the query sequence and the index sequence of the cluster, determines a hit index sequence from the index sequence of the cluster according to the first similarity, obtains a cluster set of the RNA structure database, determines at least one target cluster from the cluster set based on the hit index sequence, obtains a second similarity between the query sequence and the second discretized sequence in the target cluster, and determines a target RNA associated with the target cluster. Structure grouping, and based on the second similarity, determining at least one target RNA structure corresponding to the first RNA structure from the second RNA structure in the target RNA structure grouping, thereby, the present disclosure pre-constructs the index information of the RNA structure database, based on the pre-trained first VQ-VAE model, obtains the first discretized sequence of the first RNA structure, determines the hit index sequence from the index sequence of the cluster according to the first similarity between the query sequence and the index sequence of the cluster, and determines at least one target RNA structure corresponding to the first RNA structure from the second RNA structure in the target RNA structure grouping based on the second similarity between the query sequence and the second discretized sequence in the target cluster, thereby reducing the complexity of RNA structure search, realizing large-scale RNA structure search, ensuring the accuracy, reliability and efficiency of RNA structure search, and by obtaining the search results of the RNA structure, it can be applied in various scenarios such as assisting the research and development of RNA drugs.

[0094] The training process of the VQ-VAE model proposed in the embodiment of the present disclosure is explained below.

[0095] like Figure 3 As shown, based on the above embodiment, the training process of the VQ-VAE model of the embodiment of the present disclosure may specifically include the following steps:

[0096] S301, obtaining a sample RNA structure, and obtaining sample structure information of the sample RNA structure.

[0097] It should be noted that the present disclosure does not limit the specific method for obtaining sample structure information of RNA structure, which can be selected according to actual conditions.

[0098] Optionally, sample coordinate information of the sample RNA structure can be obtained, the upstream nucleotide and the spatially nearest adjacent nucleotide of each nucleotide in the sample RNA structure can be determined, sample conformation information between the nucleotide and the upstream nucleotide can be obtained, the sample position relationship between the nucleotide and the adjacent nucleotide can be obtained, and the sample pairing relationship between the nucleotide and the adjacent nucleotide can be obtained. Based on the sample coordinate information, sample conformation information, sample position relationship and sample pairing relationship, sample structure information of the sample RNA structure is generated.

[0099] It should be noted that the specific process of obtaining sample coordinate information, sample conformation information, sample position relationship and sample pairing relationship can be found in the above embodiments and will not be described in detail here.

[0100] S302, input the sample structure information into the initial VQ-VAE model, and the encoder in the VQ-VAE model encodes the sample structure information to obtain the sample latent space representation.

[0101] It should be noted that the sample structure information is used as the input of the encoder in the VQ-VAE model. The encoder can map the sample structure information into a latent space representation by encoding the sample structure information to obtain the sample latent space representation.

[0102] S303, the quantization layer in the VQ-VAE model quantizes the sample latent space representation, and matches the sample quantization vector in the discretized vocabulary obtained by the current iteration to obtain the sample discretized representation of the sample RNA structure.

[0103] Among them, the core of VQ-VAE is the codebook, which can randomly initialize multiple discretized vocabularies as the initial values ​​of the codebook. The codebook contains multiple discretized vocabularies, and the discretized vocabularies include multiple codebook vectors and state codes corresponding to the codebook vectors.

[0104] It should be noted that after obtaining the sample latent space representation, the sample latent space representation is quantized through the quantization layer, and the sample quantization vector is matched in the discretized vocabulary of the current iteration. The sample quantization vector is quantized to the closest code book vector, that is, it is mapped to a discretized state to obtain the sample discretized representation of the sample RNA structure.

[0105] S304, the decoder in the VQ-VAE model decodes the discretized representation of the sample to obtain the restored RNA structure.

[0106] It should be noted that the sample discretization representation is decoded by the decoder to reconstruct the input structure from the sample discretization representation, that is, the decoder attempts to reconstruct (restore) a structure as similar as possible to the sample structure information to ensure the validity of the discretization representation, and then obtain the restored RNA structure.

[0107] S305, adjusting the model parameters of the VQ-VAE model according to the sample RNA structure and the reduced RNA structure, as well as the sample latent space representation and the sample discretization representation, and continuing training until the end condition is met to obtain a second VQ-VAE model, where the first VQ-VAE model only includes the encoder and quantization layer in the second VQ-VAE model.

[0108] In the disclosed embodiments, the reconstruction loss can be determined based on the sample RNA structure and the reduced RNA structure, the quantization loss can be determined based on the sample latent space representation and the sample discretization representation, and the model parameters of the VQ-VAE model can be adjusted based on the reconstruction loss and the quantization loss.

[0109] Among them, the reconstruction loss is used to measure the difference between the sample RNA structure and the restored RNA structure, and the mean square error (MSE) can be used as the loss function.

[0110] Among them, the quantization loss is used to measure the difference between the sample latent space representation and the sample discretization representation.

[0111] Optionally, a codebook loss may also be determined, and model parameters of the VQ-VAE model may be adjusted according to the reconstruction loss, the quantization loss, and the codebook loss.

[0112] For example, the codebook is updated by exponential moving average (EMA) and other methods to determine the codebook loss, ensuring that the discretized vocabulary in the codebook can effectively cover the latent space and avoid "unused" discretized vocabulary.

[0113] Optionally, the above loss can be optimized using a back-propagation algorithm by iteratively adjusting the model parameters of the VQ-VAE model to ensure that the model converts the sample RNA structure into a discretized sequence.

[0114] It should be noted that the present disclosure does not limit the setting of the termination condition, which can be selected according to the actual situation. Optionally, the training termination condition can be set as the loss function value is less than a preset loss threshold; optionally, the training termination condition can also be set as the number of adjustments of the model parameters of the VQ-VAE model reaches a preset number of thresholds.

[0115] In an embodiment of the present disclosure, after the training of the VQ-VAE model is completed, the structural information of any RNA structure can be input into a pre-trained first VQ-VAE model, and the encoder in the first VQ-VAE model encodes the structural information to obtain a latent space representation. The quantization layer in the first VQ-VAE model quantizes the latent space representation, and matches the quantized vector in the discretized vocabulary to obtain a discretized sequence of any RNA structure.

[0116] In summary, the RNA structure search method of the disclosed embodiment obtains the sample RNA structure and the sample structure information of the sample RNA structure, inputs the sample structure information into the initial VQ-VAE model, and the encoder in the VQ-VAE model encodes the sample structure information to obtain the sample latent space representation, and the quantization layer in the VQ-VAE model quantizes the sample latent space representation, and matches the sample quantization vector in the discretized vocabulary obtained by the current iteration to obtain the sample discretized representation of the sample RNA structure, and the decoder in the VQ-VAE model decodes the sample discretized representation to obtain the restored RNA structure, and according to the sample RNA structure and Restore the RNA structure, as well as the sample latent space representation and the sample discretization representation, adjust the model parameters of the VQ-VAE model, and continue training until the end condition is met to obtain a second VQ-VAE model. The first VQ-VAE model only includes the encoder and quantization layer in the second VQ-VAE model. Therefore, the present invention trains the VQ-VAE model through sample structure information. The trained VQ-VAE model can process the structural information of any input RNA structure to obtain a discretized sequence of any RNA structure, thereby improving the accuracy and reliability of obtaining the discretized sequence of the RNA structure and laying a foundation for subsequent RNA structure search.

[0117] Figure 4 The RNA structure search device 400 proposed in the present disclosure is applied to a computing node in a distributed system, and includes: an acquisition module 401, a determination module 402 and a search module 403.

[0118] An acquisition module 401 is used to acquire a first discretized sequence of a first RNA structure to be searched as a query sequence;

[0119] A determination module 402 is used to perform an index search in a pre-constructed RNA structure database based on the query sequence to determine a hit index sequence;

[0120] The search module 403 is used to perform a structure search in the RNA structure database according to the query sequence and the hit index sequence, and determine at least one target RNA structure corresponding to the first RNA structure.

[0121] Furthermore, before obtaining the first discretized sequence of the first RNA structure to be searched as a query sequence, the device 400 is also used to: obtain a known RNA structure set, the RNA structure set includes multiple second RNA structures, and construct the RNA structure database based on the multiple second RNA structures; obtain a second discretized sequence of the second RNA structure; and construct index information of the RNA structure database according to the second discretized sequence.

[0122] Furthermore, the device 400 is also used to: determine a similarity matrix between multiple second discretized sequences of the second RNA structure; cluster the multiple second discretized sequences of the second RNA structure according to the similarity matrix to obtain a cluster set of the RNA structure database; determine a central discretized sequence from the second discretized sequences included in each cluster in the cluster set as an index sequence of the cluster; and construct index information of the RNA structure database based on the index sequence of the cluster.

[0123] Furthermore, the determination module 402 is also used to: obtain index information of the RNA structure database, the index information including the index sequence of each cluster in the cluster set; obtain a first similarity between the query sequence and the index sequence of the cluster; and determine the hit index sequence from the index sequence of the cluster according to the first similarity.

[0124] Furthermore, the determination module 402 is also used to: obtain index information of the RNA structure database, the index information including the index sequence of each cluster in the cluster set; obtain a first similarity between the query sequence and the index sequence of the cluster; and determine the hit index sequence from the index sequence of the cluster according to the first similarity.

[0125] Furthermore, the device 400 is also used to: obtain structural information of any RNA structure among the first RNA structure and the second RNA structure; and determine a discretized sequence of any RNA structure based on the structural information of any RNA structure.

[0126] Furthermore, the device 400 is also used to: obtain coordinate information of any RNA structure; determine the upstream nucleotide and the spatially nearest adjacent nucleotide of each nucleotide in any RNA structure; obtain conformation information between the nucleotide and the upstream nucleotide; obtain the positional relationship between the nucleotide and the adjacent nucleotide; obtain the pairing relationship between the nucleotide and the adjacent nucleotide; and generate the structural information of any RNA structure based on the coordinate information, the conformation information, the positional relationship and the pairing relationship.

[0127] Furthermore, the device 400 is also used to: obtain a latent space representation of any RNA structure based on the structural information of any RNA structure; quantize the latent space representation to obtain a quantization vector, and obtain a discretized sequence of any RNA structure from a pre-constructed discretization vocabulary based on the quantization vector.

[0128] Furthermore, the device 400 is also used to: input the structural information of any RNA structure into a pre-trained first vector quantization variational autoencoder VQ-VAE model, and the encoder in the first VQ-VAE model encodes the structural information to obtain the latent space representation; the quantization layer in the first VQ-VAE model quantizes the latent space representation, and matches the quantized vector in the discretized vocabulary to obtain a discretized sequence of any RNA structure.

[0129] Further, the discretized vocabulary includes multiple codebook vectors and state codes corresponding to the codebook vectors, wherein the discretized sequence of the RNA structure includes state codes corresponding to the codebook vectors in the discretized vocabulary that match the quantization vector.

[0130] Furthermore, the training process of the VQ-VAE model includes: obtaining a sample RNA structure and obtaining sample structure information of the sample RNA structure; inputting the sample structure information into the initial VQ-VAE model, and encoding the sample structure information by the encoder in the VQ-VAE model to obtain a sample latent space representation; quantizing the sample latent space representation by the quantization layer in the VQ-VAE model, and matching the sample quantization vector in the discretized vocabulary obtained by the current iteration to obtain a sample discretized representation of the sample RNA structure; decoding the sample discretized representation by the decoder in the VQ-VAE model to obtain a restored RNA structure; adjusting the model parameters of the VQ-VAE model according to the sample RNA structure and the restored RNA structure, as well as the sample latent space representation and the sample discretized representation, and continuing training until the end condition is met to obtain a second VQ-VAE model, wherein the first VQ-VAE model only includes the encoder and quantization layer in the second VQ-VAE model.

[0131] Furthermore, the device 400 is also used to: determine the reconstruction loss based on the sample RNA structure and the reduced RNA structure; determine the quantization loss based on the sample latent space representation and the sample discretization representation; and adjust the model parameters of the VQ-VAE model based on the reconstruction loss and the quantization loss.

[0132] In summary, the searching device for RNA structure of the ribonucleic acid (RNA) of the embodiment of the present disclosure obtains a first discretized sequence of the first RNA structure to be searched as a query sequence, performs an index search in a pre-constructed RNA structure database based on the query sequence, determines a hit index sequence, and performs a structure search in the RNA structure database according to the query sequence and the hit index sequence to determine at least one target RNA structure corresponding to the first RNA structure. Thus, the present disclosure reduces the complexity of obtaining the RNA structure while improving the efficiency and accuracy of obtaining the RNA structure by discretizing the RNA structure and performing a structure search in a pre-constructed RNA structure database to determine the target RNA structure.

[0133] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0134] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium and a computer program product.

[0135] Figure 5A schematic block diagram of an example electronic device 500 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0136] like Figure 5 As shown, the electronic device 500 includes a computing unit 501, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 502 or a computer program loaded from a storage unit 508 into a random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the device 500 can also be stored. The computing unit 501, the ROM 502, and the RAM 503 are connected to each other via a bus 504. An input / output (I / O) interface 505 is also connected to the bus 504.

[0137] A number of components in the device 500 are connected to the I / O interface 505, including: an input unit 506, such as a keyboard, a mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a disk, an optical disk, etc.; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 509 allows the device 500 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0138] The computing unit 501 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 501 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The computing unit 501 performs the various methods and processes described above, such as Figures 1 to 3For example, in some embodiments, the method for searching for RNA structures of RNA can be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 508. In some embodiments, part or all of the computer program can be loaded and / or installed on the device 500 via the ROM 502 and / or the communication unit 509. When the computer program is loaded into the RAM 503 and executed by the computing unit 501, one or more steps of the method for searching for RNA structures of RNA described above can be performed. Alternatively, in other embodiments, the computing unit 501 can be configured to perform the method for searching for RNA structures of RNA by any other appropriate means (e.g., by means of firmware).

[0139] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0140] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable RNA structure search device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine as a stand-alone software package and partially on a remote machine, or entirely on a remote machine or server.

[0141] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0142] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0143] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), the Internet, and a blockchain network.

[0144] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0145] According to an embodiment of the present disclosure, the present disclosure also provides a computer program product, including a computer program, wherein the computer program, when executed by a processor, implements the steps of the method for searching the ribonucleic acid RNA structure shown in the above embodiment of the present disclosure.

[0146] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0147] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for searching for RNA structure, wherein: The method comprises: Obtaining a first discretized sequence of a first RNA structure to be searched as a query sequence; Based on the query sequence, an index search is performed in a pre-constructed RNA structure database to determine a hit index sequence; According to the query sequence and the hit index sequence, a structure search is performed in the RNA structure database to determine at least one target RNA structure corresponding to the first RNA structure.

2. The method according to claim 1, wherein: Before obtaining the first discretized sequence of the first RNA structure to be searched as a query sequence, the method further includes: Acquire a known RNA structure set, wherein the RNA structure set includes a plurality of second RNA structures, and construct the RNA structure database based on the plurality of second RNA structures; Obtaining a second discretized sequence of the second RNA structure; According to the second discretized sequence, index information of the RNA structure database is constructed.

3. The method according to claim 2, wherein: The step of constructing index information of the RNA structure database according to the second discretized sequence includes: determining a similarity matrix between a plurality of second discretized sequences of the second RNA structure; Clustering the second discretized sequences of the plurality of second RNA structures according to the similarity matrix to obtain a cluster set of the RNA structure database; Determine a central discretization sequence from the second discretization sequences included in each cluster in the cluster set as an index sequence of the cluster; Based on the index sequence of the cluster, the index information of the RNA structure database is constructed.

4. The method according to any one of claims 1 to 3, wherein: The method of searching a pre-constructed RNA structure database based on the query sequence to obtain a hit index sequence comprises: Obtaining index information of the RNA structure database, wherein the index information includes an index sequence of each cluster in the cluster set; Obtaining a first similarity between the query sequence and the index sequence of the cluster; The hit index sequence is determined from the index sequences of the clusters according to the first similarity.

5. The method according to claim 4, wherein: Determining at least one target RNA structure corresponding to the first RNA structure from the RNA structure database according to the query sequence and the hit index sequence comprises: Obtaining a cluster set of the RNA structure database; Based on the hit index sequence, determining at least one target cluster from the cluster set; Obtaining a second similarity between the query sequence and a second discretized sequence in the target cluster; A target RNA structure group associated with the target cluster is determined, and based on the second similarity, at least one target RNA structure corresponding to the first RNA structure is determined from the second RNA structures in the target RNA structure group.

6. The method according to any one of claims 1 to 3, wherein: The method comprises: For any RNA structure of the first RNA structure and the second RNA structure, obtaining structural information of the any RNA structure; According to the structural information of any RNA structure, a discretized sequence of any RNA structure is determined.

7. The method according to claim 6, wherein: The obtaining of structural information of any RNA structure comprises: Obtaining coordinate information of any RNA structure; Determining the upstream nucleotide and the spatially nearest neighboring nucleotide of each nucleotide in any of the RNA structures; obtaining conformational information between the nucleotide and the upstream nucleotide; Obtaining the positional relationship between the nucleotide and the adjacent nucleotide; Obtaining a pairing relationship between the nucleotide and the adjacent nucleotide; The structural information of any RNA structure is generated according to the coordinate information, the conformation information, the positional relationship and the pairing relationship.

8. The method according to claim 6, wherein: Determining the discretized sequence of any RNA structure according to the structural information of any RNA structure includes: Obtaining a potential space representation of any RNA structure according to the structural information of any RNA structure; The latent space representation is quantized to obtain a quantized vector, and a discretized sequence of any RNA structure is obtained from a pre-constructed discretized vocabulary according to the quantized vector.

9. The method according to claim 8, wherein: The method comprises: Inputting the structural information of any RNA structure into a pre-trained first vector quantized variational autoencoder VQ-VAE model, and encoding the structural information by an encoder in the first VQ-VAE model to obtain the latent space representation; The latent space representation is quantized by the quantization layer in the first VQ-VAE model, and the quantized vector is matched in the discretized vocabulary to obtain a discretized sequence of any RNA structure.

10. The method according to claim 9, wherein: The discretization vocabulary includes multiple codebook vectors and state codes corresponding to the codebook vectors, wherein the discretization sequence of the RNA structure includes state code components corresponding to the codebook vectors in the discretization vocabulary that match the quantization vector.

11. The method according to claim 9, wherein: The training process of the VQ-VAE model includes: Obtaining a sample RNA structure and obtaining sample structure information of the sample RNA structure; Inputting the sample structure information into an initial VQ-VAE model, and encoding the sample structure information by an encoder in the VQ-VAE model to obtain a sample latent space representation; The sample latent space representation is quantized by a quantization layer in the VQ-VAE model, and the sample quantization vector is matched in the discretization vocabulary obtained by the current iteration to obtain a sample discretization representation of the sample RNA structure; The decoder in the VQ-VAE model decodes the discretized representation of the sample to obtain a restored RNA structure; According to the sample RNA structure and the reduced RNA structure, as well as the sample latent space representation and the sample discretization representation, the model parameters of the VQ-VAE model are adjusted, and training is continued until the termination condition is met to obtain a second VQ-VAE model, wherein the first VQ-VAE model only includes the encoder and quantization layer in the second VQ-VAE model.

12. The method according to claim 11, wherein: The adjusting of the model parameters of the VQ-VAE model according to the sample RNA structure and the reduced RNA structure, the sample latent space representation and the sample discretization representation includes: determining a reconstruction loss based on the sample RNA structure and the reduced RNA structure; Determining a quantization loss according to the sample latent space representation and the sample discretization representation; According to the reconstruction loss and the quantization loss, the model parameters of the VQ-VAE model are adjusted.

13. A device for searching RNA structure, comprising: An acquisition module, used for acquiring a first discretized sequence of a first RNA structure to be searched as a query sequence; A determination module, for performing an index search in a pre-constructed RNA structure database based on the query sequence to determine a hit index sequence; A search module is used to perform a structure search in the RNA structure database according to the query sequence and the hit index sequence to determine at least one target RNA structure corresponding to the first RNA structure.

14. The RNA structure search device according to claim 13, wherein: Before obtaining the first discretized sequence of the first RNA structure to be searched as a query sequence, the device is further used to: Acquire a known RNA structure set, wherein the RNA structure set includes a plurality of second RNA structures, and construct the RNA structure database based on the plurality of second RNA structures; Obtaining a second discretized sequence of the second RNA structure; According to the second discretized sequence, index information of the RNA structure database is constructed.

15. The RNA structure search device according to claim 14, wherein: The device is also used for: determining a similarity matrix between a plurality of second discretized sequences of the second RNA structure; Clustering the second discretized sequences of the plurality of second RNA structures according to the similarity matrix to obtain a cluster set of the RNA structure database; Determine a central discretization sequence from the second discretization sequences included in each cluster in the cluster set as an index sequence of the cluster; Based on the index sequence of the cluster, the index information of the RNA structure database is constructed.

16. The RNA structure search device according to claims 13-15, wherein: The determining module is further used for: Obtaining index information of the RNA structure database, wherein the index information includes an index sequence of each cluster in the cluster set; Obtaining a first similarity between the query sequence and the index sequence of the cluster; The hit index sequence is determined from the index sequences of the clusters according to the first similarity.

17. The RNA structure search device according to claim 16, wherein: The determining module is further used for: Obtaining a cluster set of the RNA structure database; Based on the hit index sequence, determining at least one target cluster from the cluster set; Obtaining a second similarity between the query sequence and a second discretized sequence in the target cluster; A target RNA structure group associated with the target cluster is determined, and based on the second similarity, at least one target RNA structure corresponding to the first RNA structure is determined from the second RNA structures in the target RNA structure group.

18. The RNA structure search device according to claim 18, wherein: The device is also used for: For any RNA structure of the first RNA structure and the second RNA structure, obtaining structural information of the any RNA structure; According to the structural information of any RNA structure, a discretized sequence of any RNA structure is determined.

19. The RNA structure search device according to claim 18, wherein: The device is also used for: Obtaining coordinate information of any RNA structure; Determining the upstream nucleotide and the spatially nearest neighboring nucleotide of each nucleotide in any of the RNA structures; obtaining conformational information between the nucleotide and the upstream nucleotide; Obtaining the positional relationship between the nucleotide and the adjacent nucleotide; Obtaining a pairing relationship between the nucleotide and the adjacent nucleotide; The structural information of any RNA structure is generated according to the coordinate information, the conformation information, the positional relationship and the pairing relationship.

20. The RNA structure search device according to claim 18, wherein: The device is also used for: Obtaining a potential space representation of any RNA structure according to the structural information of any RNA structure; The latent space representation is quantized to obtain a quantized vector, and a discretized sequence of any RNA structure is obtained from a pre-constructed discretized vocabulary according to the quantized vector.

21. The RNA structure search device according to claim 20, wherein: The device is also used for: Inputting the structural information of any RNA structure into a pre-trained first vector quantized variational autoencoder VQ-VAE model, and encoding the structural information by an encoder in the first VQ-VAE model to obtain the latent space representation; The latent space representation is quantized by the quantization layer in the first VQ-VAE model, and the quantized vector is matched in the discretized vocabulary to obtain a discretized sequence of any RNA structure.

22. The RNA structure search device according to claim 21, wherein: The discretization vocabulary includes multiple codebook vectors and state codes corresponding to the codebook vectors, wherein the discretization sequence of the RNA structure includes state code components corresponding to the codebook vectors in the discretization vocabulary that match the quantization vector.

23. The RNA structure search device according to claim 21, wherein: The training process of the VQ-VAE model includes: Obtaining a sample RNA structure and obtaining sample structure information of the sample RNA structure; Inputting the sample structure information into an initial VQ-VAE model, and encoding the sample structure information by an encoder in the VQ-VAE model to obtain a sample latent space representation; The sample latent space representation is quantized by a quantization layer in the VQ-VAE model, and the sample quantization vector is matched in the discretization vocabulary obtained by the current iteration to obtain a sample discretization representation of the sample RNA structure; The decoder in the VQ-VAE model decodes the discretized representation of the sample to obtain a restored RNA structure; According to the sample RNA structure and the reduced RNA structure, as well as the sample latent space representation and the sample discretization representation, the model parameters of the VQ-VAE model are adjusted, and training is continued until the termination condition is met to obtain a second VQ-VAE model, wherein the first VQ-VAE model only includes the encoder and quantization layer in the second VQ-VAE model.

24. The RNA structure search device according to claim 23, wherein: The device is also used for: determining a reconstruction loss based on the sample RNA structure and the reduced RNA structure; Determining a quantization loss according to the sample latent space representation and the sample discretization representation; According to the reconstruction loss and the quantization loss, the model parameters of the VQ-VAE model are adjusted.

25. An electronic device comprising: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 12.

26. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-12.

27. A computer program product comprising a computer program, which, when executed by a processor, implements the steps of the method according to any one of claims 1 to 12.

Images