An information processing method, device and equipment for respiratory syncytial virus typing

Abstract

The present invention discloses an information processing method, device and equipment for respiratory syncytial virus typing, which relates to the technical field of virus classification. The present invention innovatively proposes an information processing method for respiratory syncytial virus typing. On the basis of the original one-hot encoding and zero-padding strategies, a hierarchical one-hot encoding strategy is innovatively proposed, which reduces the interference of the traditional zero-padding method on one-hot encoding and enhances the robustness to insertion mutations. Based on the hierarchical one-hot encoding strategy and neural network image classification technology, it can help doctors and scientific researchers make accurate diagnoses and treatment guidance, improve work efficiency, and help improve the efficiency and diagnosis and treatment quality of the medical and public safety industries.

Description

Technical Field

[0001] The present invention relates to the technical field of virus classification, and in particular, to an information processing method, device, and equipment for typing respiratory syncytial virus. Background Art

[0002] Genotype typing of respiratory syncytial virus is of great significance. It helps to deeply understand the epidemiological characteristics of the virus. By accurately determining different genotypes, it is possible to clearly track the transmission routes and epidemic trends of the virus in different regions and different populations, providing a key basis for the formulation of public health prevention and control strategies. Clinically, genotype typing can assist doctors in more accurately judging the severity of the disease. Since respiratory syncytial virus of different genotypes may have differences in pathogenicity, clinical manifestations after infection, and the speed of disease progression, personalized treatment plans can be formulated accordingly to improve the treatment effect and the prognosis of patients.

[0003] In the Nextrain typing technology, an evolutionary tree is constructed based on the G gene sequence or the whole genome sequence. In the process of constructing the evolutionary tree, different respiratory syncytial virus strains are carefully classified and typed according to various characteristics such as sequence similarity and difference. This typing method has greatly improved our understanding of the diversity and evolutionary relationships of respiratory syncytial virus, helping to clearly track the transmission source, transmission route, and mutation trend of the virus in epidemiological studies. In clinical diagnosis, it can provide doctors with more accurate virus typing information to assist in judging the specific type of infection, thus formulating a more targeted treatment strategy. However, in the typing method based on the evolutionary tree, there are problems such as different classification results due to different reference evolutionary trees.

[0004] In the research and practice of virus classification, deep learning technology has demonstrated powerful potential and application value. The Deepvirusclassifier software focuses on the whole genome sequence and uses the one-hot encoding technology to effectively extract features from the whole genome sequence data of the virus. This encoding method has advantages compared with directly converting A, T, C, and G into 1, 2, 3, and 4, because the difference between A and G is not the same as the difference between T and G in terms of 2 and 3. Based on the deep learning network framework, the software accurately identifies the characteristic patterns unique to the whole genome sequence of the novel coronavirus, forming a sharp contrast with the sequence characteristics of other viruses, so as to efficiently and accurately determine whether an unknown virus sample belongs to the novel coronavirus. However, in the one-hot encoding of this software, there is no fixed reference gene, base substitutions are represented by ATCG, N represents that the base here is unknown, and the zero-padding strategy is adopted at the end to represent insertions and deletions between sequences. However, when genotyping the same species, this zero-padding strategy has some deficiencies, such as clustering errors, the zero-padding method will interfere with the one-hot encoding, and it has poor robustness to indel (insertion) mutations.

[0005] In view of this, the present invention is specifically proposed. Summary of the Invention

[0006] The purpose of the present invention is to provide a method, device and equipment for information processing of respiratory syncytial virus genotyping to solve the above technical problems.

[0007] The present invention is implemented as follows:

[0008] In the first aspect, the present invention provides a method for constructing a respiratory syncytial virus genotyping model, which includes the following steps:

[0009] S1: Construct a reference genome: Select the respiratory syncytial virus genome for constructing the reference genome from the respiratory syncytial virus genome sequence database, and divide the sequence into a training set, a validation set and a test set; perform sequence alignment on some genomes in the training set, find out the similar regions and variant regions among all genomes in the training set, and construct a pan-genome map containing all genome information based on the alignment results; merge the variant regions to obtain the reference genome;

[0010] S2: Perform hierarchical one-hot encoding: The length of the reference genome is X. The whole genome sequences of the training set and the validation set are aligned with the reference genome and one-hot encoded to obtain an image with the query vector * key vector * value vector of (X+1)*7*1. The key vector in the image refers to the following seven pieces of information: whether there is an A base, T base, C base, G base, missing or unknown base, insertion variation, and whether the base position is located in the G protein coding gene region after alignment with the reference genome;

[0011] The whole genome sequences of the training set and validation set are once again ont-hot encoded to obtain a query vector * key vector * value vector of 50*120 to 200*1 images. The key vector in the 50*120 to 200*1 image refers to the number of insertion variants. If there is at least one insertion variant, it is encoded in the first column of the 50*120 to 200*1 image.

[0012] S3: Fusing hierarchical features: Using the CNN network, extract features from the (X+1)*7*1 image and the 50*120~200*1 image, respectively, to obtain features of the target dimension, and then concatenate and fuse the obtained images to obtain a fused feature map.

[0013] S4: RSV genotyping based on neural network: input fusion feature map, use neural network to perform RSV genotyping and output;

[0014] S5: Input the full genome of the respiratory syncytial virus in the test set and proceed to steps S2-S4.

[0015] In a second aspect, the present invention provides an information processing method for respiratory syncytial virus typing, comprising the following steps:

[0016] A respiratory syncytial virus typing model is constructed according to the above-mentioned method for constructing a respiratory syncytial virus typing model, and then the full genome of the respiratory syncytial virus of the sample to be tested is input to perform steps S2-S4.

[0017] In a third aspect, the present invention further provides a device for typing respiratory syncytial virus, comprising: an input module, a control module, and an output module;

[0018] The input module was configured to: input the respiratory syncytial virus genome;

[0019] The control module includes: reference genome construction module, hierarchical one hot encoding module, fusion feature module and genotyping module;

[0020] The output module includes: outputting respiratory syncytial virus typing results;

[0021] Among them, the reference genome construction module is configured to: divide the sequences into a training set, a validation set, and a test set based on the input respiratory syncytial virus genome; perform sequence alignment on the genomes in part of the training set, find the similar regions and variant regions among all the genomes in the part of the training set, and construct a pan-genome map containing all the genome information based on the alignment results; merge the variant regions and output the reference genome, the training set, the validation set, and the test set;

[0022] The hierarchical one-hot encoding module is configured to: input the whole-genome sequences of the reference genome, the training set, and the validation set, align the whole-genome sequences of the training set and the validation set with the reference genome, and perform one-hot encoding to obtain a picture with query vector * key vector * value vector of (X + 1) * 7 * 1, where X is the length of the reference genome; the key vector in the picture refers to the following seven kinds of information: whether there is an A base, a T base, a C base, a G base, a deletion or an unknown base, an insertion mutation, and whether the base at the position is within the coding gene region of the G protein after alignment with the reference genome; output a picture with query vector * key vector * value vector of (X + 1) * 7 * 1; perform another one-hot encoding on the whole-genome sequences of the training set and the validation set, and output a picture with query vector * key vector * value vector of 50 * 120 - 200 * 1; the key vector in the picture of 50 * 120 - 200 * 1 refers to the number of insertion mutations; if there is at least one insertion mutation, encode it at the front of the picture of 50 * 120 - 200 * 1;

[0023] The fusion feature module is configured to: input the picture of (X + 1) * 7 * 1 and the picture of 50 * 120 - 200 * 1, perform feature extraction using a CNN network, obtain features of the target dimension respectively, splice and fuse the obtained images, and output a fusion feature map;

[0024] The genotyping module is configured to: input the fusion feature map, perform RSV genotyping using a neural network and output the respiratory syncytial virus genotyping result.

[0025] In a fourth aspect, the present invention also provides a respiratory syncytial virus genotyping device, which includes a processor and a memory. At least one instruction, at least one program, a code set or an instruction set is stored in the memory, and at least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor to implement the above-mentioned information processing method for respiratory syncytial virus genotyping.

[0026] In a fifth aspect, the present invention also provides a computer-readable storage medium, in which at least one instruction, at least one program, a code set or an instruction set is stored, and at least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor to implement the above-mentioned information processing method for respiratory syncytial virus genotyping.

[0027] The present invention has the following beneficial effects:

[0028] The present invention innovatively proposes an information processing method for respiratory syncytial virus genotyping. Based on the original one-hot encoding and zero-padding strategy, a hierarchical one-hot encoding strategy is innovatively proposed, which reduces the interference of the traditional zero-padding method on one-hot encoding and enhances the robustness to indel (insertion) mutations.

[0029] Based on the one-hot encoding, the present invention obtains two images, an (X + 1)*7*1 image and a 50*120 - 200*1 image. Based on the proposed CNN network, the features of the two images are extracted and registered and aligned, enhancing the ability to capture mutation information in RSV genomic images, making it easier for CNN to capture detailed information about the RSV genomic sequence. There is a stronger corresponding relationship between the features extracted by CNN. Therefore, it is easier to capture the supplementary information for insertion mutations in the second image during the fusion process, improving the final RSV genotype classification accuracy. By comparison, the classification accuracy of the RSV genotype classification method provided by the present invention is better than that of the classification method based on phylogenetic trees, and the clustering error rate is lower.

[0030] The development of the respiratory syncytial virus genotyping strategy will help doctors and researchers diagnose and evaluate more accurately. By optimizing the feature extraction of respiratory genomic sequences and applying deep learning techniques, more accurate diagnostic tools can be provided for the public safety field, helping doctors and researchers better understand respiratory syncytial virus and providing more effective treatment plans. Based on the hierarchical one-hot encoding strategy and neural network image classification technology, it can help doctors and researchers make accurate diagnoses and treatment guidance, improve work efficiency, and help improve the efficiency and diagnosis and treatment quality of the medical and public safety industries. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other related drawings can be obtained based on these drawings without creative efforts.

[0032] Figure 1 The first image for hierarchical ont hot encoding;

[0033] Figure 2 The second image for hierarchical ont hot encoding;

[0034] Figure 3It is a schematic diagram of a network architecture. Detailed implementation manners

[0035] Reference to the embodiments of the present invention will now be provided in detail, with one or more examples described below. Each example is provided for explanation and not for limiting the present invention. In fact, it will be obvious to those skilled in the art that various modifications and variations can be made to the present invention without departing from the scope or spirit of the present invention. For example, features described or illustrated as part of one embodiment can be used in another embodiment to produce a further embodiment.

[0036] In a first aspect, the present invention provides a method for constructing a respiratory syncytial virus genotyping model, which includes the following steps:

[0037] S1: Construct a reference genome: Select the respiratory syncytial virus genomes for constructing the reference genome from the respiratory syncytial virus genome sequence database, and divide the sequences into a training set, a validation set, and a test set; Perform sequence alignment on the genomes in part of the training set to find the similar regions and variant regions among all the genomes in part of the training set, and construct a pan-genome map containing all the genome information based on the alignment results; Merge the variant regions to obtain the reference genome;

[0038] S2: Perform hierarchical one-hot encoding: The length of the reference genome is X. Align the whole-genome sequences of the training set and the validation set with the reference genome and perform one-hot encoding to obtain a picture with query vector * key vector * value vector of (X + 1) * 7 * 1; The key vector in the picture refers to the following seven kinds of information: whether there is an A base, a T base, a C base, a G base, a deletion or an unknown base, an insertion mutation, and whether the base at the position is within the coding gene region of the G protein after alignment with the reference genome;

[0039] Perform one-hot encoding on the whole-genome sequences of the training set and the validation set again to obtain a picture with query vector * key vector * value vector of 50 * 120 to 200 * 1; The key vector in the picture of 50 * 120 to 200 * 1 refers to the number of insertion mutations; If there is at least one insertion mutation, encode it at the front of the picture of 50 * 120 to 200 * 1;

[0040] S3: Fuse the hierarchical features: Use a CNN network to extract features from the picture of (X + 1) * 7 * 1 and the picture of 50 * 120 to 200 * 1 respectively, obtain features of the target dimension, splice and fuse the obtained images, and obtain a fused feature map;

[0041] S4: Perform RSV genotyping based on a neural network: Input the fused feature map, and perform RSV genotyping with a neural network and output;

[0042] S5: Input the whole genome of respiratory syncytial virus in the test set, and perform steps S2 - S4.

[0043] The present invention first constructs a pan - genome based on the graphical pan - genome strategy as a reference genome. Based on the reference genome, the respiratory syncytial virus genome sequence is encoded, including A, T, C, G bases, as well as deletion and unknown bases, insertion bases, and the gene coding sequence (base weight) of the G protein. In order to capture insertion variant information more accurately and conveniently, a multi - graph strategy is adopted to represent possible insertions. Two graphs are obtained based on one - hot encoding. Features are extracted based on the proposed CNN network and registered and aligned to enhance the ability to capture mutation information in the RSV genome image, making it easier for the CNN to capture detailed information about the RSV genome sequence. There is a stronger correspondence relationship between the features extracted by the CNN. Therefore, it is easier to capture supplementary information about insertion mutations in the second picture during the fusion process, improving the final accuracy of RSV genotype classification.

[0044] After establishing the RSV typing model using the training set and the validation set, the test set is used to test the model, and then it can be used for typing the respiratory syncytial virus in the samples to be tested.

[0045] In one implementation, both deletions and unknowns are encoded using N, insertion variants are encoded using I, and within the G protein - coding gene range, it is encoded using S. In other implementations, replacements can also be made according to needs, not limited to the above - mentioned coding naming methods.

[0046] Add the base weight of the G protein gene sequence and perform accurate classification according to the G protein gene sequence variation information.

[0047] The ratio of the training set to the validation set is, for example, 6 - 8:2 - 4, such as 8:2, 6:4, or 7:3. The number of the test set can be adjusted according to needs.

[0048] In a preferred implementation of the application of the present invention, step S1 includes merging the variant regions according to the following rules: 1) If a position has only substitution variants, only retain the base with the highest mutation frequency; 2) If a position has deletion variants, retain the non - deleted bases; 3) If a position has insertion variants, retain the insertion variants.

[0049] In a preferred implementation of the application of the present invention, the one - hot encoding rule when obtaining a picture of (X + 1)*7*1 is as follows:

[0050] Optionally, 7 rows in the picture are respectively used as the coding rows for A base, T base, C base, G base, insertion variants, and within the coding gene region of the G protein;

[0051] After alignment with the reference genome, for each position, if there is an A base, the encoding row of the A base is encoded as 1 at that position; otherwise, it is encoded as 0. After alignment with the reference genome, for each position, if there is a T base, the encoding row of the T base is encoded as 1 at that position, and vice versa. After alignment with the reference genome, for each position, if there is a C base, the encoding row of the C base is encoded as 1 at that position, and vice versa. After alignment with the reference genome, for each position, if there is a G base, the encoding row of the G base is encoded as 1 at that position, and vice versa. After alignment with the reference genome, for each position, if there is a deletion or unknown base, the encoding row of the deletion or unknown base is encoded as 1 at that position, and vice versa. After alignment with the reference genome, for each position, if there is an insertion variant, the encoding row of the insertion variant is encoded as 1 at that position, and vice versa. For each position, whether it is located within the coding gene region of the G protein, if so, it is encoded as 1, and vice versa.

[0052] In a preferred embodiment of the application of the present invention, the rules for ont hot encoding when obtaining a 50*120 - 200*1 picture are as follows:

[0053] If there is no insertion variant in the sequence, all are filled with 0. If there is at least one insertion variant in the sequence, encoding is performed in the front row of the 50*120 - 200*1 picture. If the length of an insertion variant is 1bp, then one hot encoding of the inserted base is performed in rows 1 - 4. If the length of an insertion variant is 2bp, then one hot encoding of the inserted base is performed in rows 1 - 8, and so on, until the length of an insertion variant is 50bp, then one hot encoding of the inserted base is performed in rows 1 - 200.

[0054] In the 50*120 - 200*1 picture, 120 - 200 refers to the number of bases of the insertion variant being 30 - 50, and there are 4 possible bases in each insertion variant. Therefore, the key vector is a multiple of 4.

[0055] In a preferred embodiment of the application of the present invention, when obtaining a 50*120 - 200*1 picture, first perform sequence alignment on the whole genome sequences of the training set that did not participate in the construction of the reference genome and / or the RSV whole genome sequence in the validation set with the reference genome, judge the number of insertion variants in the training set and / or validation set that did not participate in the construction of the reference genome, and adjust the key vector according to the number of insertion variants.

[0056] This setting helps prevent the insertion mutation number from being too small when setting the key vector of the second picture, resulting in an excessive insertion mutation number in some parts of the genomic sequence that cannot be captured, causing missing insertion mutation information and thus unable to classify RSV more accurately. By aligning the whole-genome sequences of the training set that did not participate in the construction of the reference genome and / or the RSV whole-genome sequences in the validation set with the reference genome, the insertion mutation number in the training set and / or validation set that did not participate in the construction of the reference genome is judged, and the key vector is adjusted according to the insertion mutation number, so that the key vector can be adaptively adjusted according to the information in the database.

[0057] In a preferred embodiment of the application of the present invention, when fusing the hierarchical features in step S3, the ResNet50 network backbone layer is used to extract features, and each layer has a 3×3 convolutional layer, a Relu activation function and batch normalization processing to extract features.

[0058] Based on the one-hot encoding step, two pictures are obtained. The second picture is an effective supplement to the first picture. Therefore, fusing these two pictures can provide more comprehensive and rich features, thereby improving the final RSV genotype classification accuracy. The cross-modal features are fused by using the fusion method of feature weighted sum. The weight matrix of the fusion layer is trainable, and the initial weight value of the multi-layer feature fusion is set to 1. The feature fusion weight value will be backpropagated and updated together with the classification loss.

[0059] In a preferred embodiment of the application of the present invention, the neural network is selected from a BP neural network, an RNN neural network, a GRU neural network, an LSTM neural network or a Transformer architecture.

[0060] Preferably, it is a Transformer architecture. Compared with the BP neural network, the RNN neural network and the LSTM neural network, the Transformer architecture has more accurate classification.

[0061] The present invention applies the Transformer model to virus classification, breaking through the limitations of traditional methods. The self-attention mechanism of Transformer has the ability to establish global relationships, improving the model's perception ability of the whole-genome information of the genome. Through the Transformer position encoding, the model can take into account the relative positions of the features in the RSV whole-genome sequence, thereby better capturing the spatial information of the image and improving the accuracy of RSV genotype classification.

[0062] In a preferred embodiment of the present invention, step S4 inputs the fused feature map, uses the Transformer architecture to perform RSV genotyping and outputs it; the feature encoding layer Tranformer encoder is composed of Multi-head Self-Attention (MSA) and MLP layers;

[0063] In a preferred embodiment of the present invention, Layer Norm normalization is used before executing the MSA and MLP layers, and residual connection is used after the layers;

[0064] In a preferred embodiment of the present invention, if the maximum probability of typing after the whole genome sequence of the tested sample is compared with the reference genome is less than 0.95, the tested sample is judged to be a new RSV genotype;

[0065] In a preferred embodiment of the present invention, when RSV genotyping, RSV type A is divided into GA1, GA2, GA2.1, GA2.2, GA2.3, GA2.3.1, GA2.3.2, GA2.3.2a, GA2.3.2b, GA2.3.3, GA2.3.5, GA3, GA3.0.1 and GA3.0.2; RSV type B is divided into GB1, GB2, GB4, GB5, GB5.0.1, GB5.0.2, GB5.0.3, GB5.0.4a and GB5.0.5a.

[0066] In one embodiment, the respiratory syncytial virus genome sequence database is selected from the nextstrain website, such as https: / / nextstrain.org / rsv / a / genome / 6y. The whole genome sequences of RSVA and RSVB were downloaded, and sequences with a length of less than 12,000 were deleted using seqkit v2.2.0 software, and the sequences were retained for subsequent analysis.

[0067] In a second aspect, the present invention provides an information processing method for respiratory syncytial virus typing, comprising the following steps:

[0068] A respiratory syncytial virus typing model is constructed according to the above-mentioned method for constructing a respiratory syncytial virus typing model, and then the full genome of the respiratory syncytial virus of the sample to be tested is input to perform steps S2-S4.

[0069] In a third aspect, the present invention further provides a device for typing respiratory syncytial virus, comprising: an input module, a control module, and an output module;

[0070] The input module was configured to: input the respiratory syncytial virus genome;

[0071] The control module includes: a reference genome construction module, a hierarchical one-hot encoding module, a fusion feature module, and a genotyping module;

[0072] The output module includes: outputting the typing results of respiratory syncytial virus;

[0073] Among them, the reference genome construction module is configured to: divide the sequences into a training set, a validation set, and a test set based on the input respiratory syncytial virus genome; perform sequence alignment on part of the genomes in the training set, find the similar regions and variant regions among all the genomes in the training set, and construct a pan-genome map containing all the genome information based on the alignment results; merge the variant regions and output the reference genome, the training set, the validation set, and the test set;

[0074] The hierarchical one-hot encoding module is configured to: input the whole-genome sequences of the reference genome, the training set, and the validation set, align the whole-genome sequences of the training set and the validation set with the reference genome, and perform one-hot encoding to obtain a picture with query vector * key vector * value vector of (X + 1) * 7 * 1, where X is the length of the reference genome; the key vector in the picture refers to the following seven kinds of information: whether there is an A base, a T base, a C base, a G base, a deletion or an unknown base, an insertion mutation, and whether the base at the position is within the coding gene region of the G protein after alignment with the reference genome; output a picture with query vector * key vector * value vector of (X + 1) * 7 * 1; perform another one-hot encoding on the whole-genome sequences of the training set and the validation set, and output a picture with query vector * key vector * value vector of 50 * 120 - 200 * 1; the key vector in the picture of 50 * 120 - 200 * 1 refers to the number of insertion mutations; if there is at least one insertion mutation, encode it at the front of the picture of 5 * 120 - 200 * 1;

[0075] The fusion feature module is configured to: input the picture of (X + 1) * 7 * 1 and the picture of 50 * 120 - 200 * 1, perform feature extraction using a CNN network to obtain features of the target dimension respectively, splice and fuse the obtained images, and output a fusion feature map;

[0076] The genotyping module is configured to: input the fusion feature map, perform RSV genotyping using a neural network and output the typing results of respiratory syncytial virus.

[0077] Fourthly, the present invention also provides a device for establishing a respiratory syncytial virus typing model, including: an input module, a control module, and an output module;

[0078] The input module is configured to: input the respiratory syncytial virus genome in the database;

[0079] The control module includes: a reference genome construction module, a hierarchical one-hot encoding module, a fusion feature module, and a genotyping module;

[0080] The output module includes: an output respiratory syncytial virus genotyping model;

[0081] Among them, the reference genome construction module is configured to: divide the sequences into a training set, a validation set, and a test set based on the input respiratory syncytial virus genome; perform sequence alignment on the genomes in part of the training set to find the similar regions and variant regions among all the genomes in the part of the training set, and construct a pan-genome map containing all the genome information based on the alignment results; merge the variant regions and output the reference genome, the training set, the validation set, and the test set;

[0082] The hierarchical one-hot encoding module is configured to: input the whole genome sequences of the reference genome, the training set, and the validation set, align the whole genome sequences of the training set and the validation set with the reference genome, and perform one-hot encoding to obtain a picture with query vector * key vector * value vector of (X + 1) * 7 * 1, where X is the length of the reference genome; the key vector in the picture refers to the following seven kinds of information: whether there is an A base, a T base, a C base, a G base, a deletion or an unknown base, an insertion mutation, and whether the base at the position is within the coding gene region of the G protein after alignment with the reference genome; output a picture with query vector * key vector * value vector of (X + 1) * 7 * 1; perform another one-hot encoding on the whole genome sequences of the training set and the validation set, and output a picture with query vector * key vector * value vector of 50 * 120 to 200 * 1; the key vector in the picture of 50 * 120 to 200 * 1 refers to the number of insertion mutations; if there is at least one insertion mutation, encode it at the front row of the picture of 50 * 120 to 200 * 1;

[0083] The fusion feature module is configured to: input the picture of (X + 1) * 7 * 1 and the picture of 50 * 120 to 200 * 1, perform feature extraction using a CNN network, respectively obtain features of the target dimension, splice and fuse the obtained images, and output a fusion feature map;

[0084] The genotyping module is configured to: input the fusion feature map, perform RSV genotyping using a neural network, and output a respiratory syncytial virus genotyping model.

[0085] In a fifth aspect, the present invention also provides a respiratory syncytial virus genotyping device, which includes a processor and a memory. At least one instruction, at least one program, a code set, or an instruction set is stored in the memory, and at least one instruction, at least one program, a code set, or an instruction set is loaded and executed by the processor to implement the above-mentioned information processing method for respiratory syncytial virus genotyping.

[0086] In a sixth aspect, the present invention further provides a device for establishing a respiratory syncytial virus typing model. The device includes a processor and a memory. At least one instruction, at least one program, a code set or an instruction set is stored in the memory, and is loaded and executed by the processor to implement the above-mentioned method for constructing a respiratory syncytial virus typing model.

[0087] Specifically, the electronic device may include a memory, a processor, a bus, and a communication interface. The memory, the processor, and the communication interface are directly or indirectly electrically connected to each other to achieve data transmission or interaction. For example, these components may be electrically connected to each other through one or more buses or signal lines. The processor can process information and / or data related to target recognition to execute one or more functions described in this application.

[0088] The memory may be, but is not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read only memory (PROM), an erasable programmable read only memory (EPROM), an electrically erasable programmable read only memory (EEPROM), etc.

[0089] The processor may be an integrated circuit chip with signal processing capabilities. The processor may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it may also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0090] In a seventh aspect, the present invention further provides a computer-readable storage medium storing at least one instruction, at least one segment of program, a code set or an instruction set, which is loaded and executed by a processor to implement the above-mentioned information processing method for respiratory syncytial virus genotyping or the construction method for a respiratory syncytial virus genotyping model.

[0091] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. For those not specified in the embodiments, the conventional conditions or the conditions recommended by the manufacturer are followed. The reagents or instruments for which the manufacturers are not specified are all conventional products that can be obtained through commercial purchase.

[0092] The features and performance of the present invention will be further described in detail below with reference to the embodiments.

[0093] Embodiment 1

[0094] The embodiment provides a respiratory syncytial virus genotyping method based on hierarchical one-hot encoding and Transformer. First, a pan-genome is constructed based on the graph pan-genome strategy as a reference genome. Based on the pan-genome, the sequences are encoded, where deletions and unknowns are both encoded with N, a multi-graph strategy is used to represent possible insertions, and the base weights of the gene sequences of the G protein are also added. The features of the multi-graph are extracted based on CNN and feature fusion is performed. A multi-modal image classification technology based on Transformer.

[0095] 1. Constructing a reference genome based on the pan-genome concept

[0096] In order to perform subsequent hierarchical one-hot encoding, the present invention first constructs a reference genome based on the pan-genome concept to provide a reference basis for subsequent encoding.

[0097] First, the complete genome sequences of RSV A and RSV B are downloaded from the nextstrain website (https: / / nextstrain.org / rsv / a / genome / 6y), and the seqkit v2.2.0 software is used to filter out the sequences with a length less than 12,000, and 11,325 sequences are retained for subsequent analysis. The sequencing sequences are divided into a training set and a validation set in an 8:2 ratio, that is, 8,000 sequences are used for the training set, 2,000 sequences are used for the validation set, and 1,325 sequences are used for the test set.

[0098] Then, a pan-genome was constructed based on the whole-genome alignment strategy. Specifically, the mafft v7.526 software was used to perform multiple sequence alignment on 6,000 respiratory syncytial virus sequences in the training set without relying on a specific reference genome. Subsequently, the similar regions and different regions between all genomes were identified, including various types of variations such as SNPs and Indels. Then, a pan-genome map containing all genomic information was constructed based on these alignment results. Finally, these variant regions were merged according to the principle of maximum frequency and maximum degree to obtain a reference genome, that is, 1) if there is only a substitution variant at a certain position, only the base with the maximum frequency is retained; 2) if there is a deletion variant at a certain position, the non-deleted base is retained; 3) if there is an insertion variant at a certain position, the insertion variant is retained. Ultimately, the length of this reference genome is 26,739 bp.

[0099] Among the remaining 2,000 RSV whole-genome sequences in the training set and the 2,000 RSV whole-genome sequences in the validation set, compared with the reference genome obtained in this invention, 384 sequences have insertion variants.

[0100] 2. Hierarchical one-hot encoding strategy

[0101] Coding strategy for the first figure: The present invention innovatively proposes a hierarchical one-hot coding strategy. After determining the reference genome, the RSV full-genome sequences of the training set and the validation set are aligned with the reference genome and encoded to obtain a 26740*7*1 picture. Among them, 26740 is the width of the picture, the first 26739 refers to the length of the reference genome, and the last 1 is at the end of the picture, indicating that compared with the reference genome, an insertion mutation is still inserted at the end of the gene in the RSV full-genome sequence. It should be noted that the original RSV full-genome sequence is about 15000bp, and this difference is caused by adding a large number of insertion mutations based on the longest principle when constructing the reference genome. Among them, 7 refers to the height of the picture. The first row indicates whether there is an A base at a certain position after alignment with the reference genome. If there is an A base, it is encoded as 1, otherwise it is 0; the second row indicates whether there is a T base at a certain position after alignment with the reference genome. If there is a T base, the coding row of the T base is encoded as 1 at this position, otherwise it is 0; the third row indicates whether there is a C base at a certain position after alignment with the reference genome. If there is a C base, the coding row of the C base is encoded as 1 at this position, otherwise it is 0; the fourth row indicates whether there is a G base at a certain position after alignment with the reference genome. If there is a G base, the coding row of the G base is encoded as 1 at this position, otherwise it is 0; the fifth row indicates whether there is a deletion or unknown base at a certain position after alignment with the reference genome. If there is a deletion or unknown base, the coding row of the deletion or unknown base is encoded as 1 at this position, otherwise it is 0; the sixth row indicates whether there is an insertion at a certain position after alignment with the reference genome. If there is an insertion mutation, the coding row is encoded as 1 at this position, otherwise it is 0 and is encoded in another figure; the 7th row indicates whether the base is within the coding gene region of the G protein. If it is, it is encoded as 1, otherwise it is 0. The coding principle diagram of the first figure with hierarchical ont hot coding is referred to Figure 1 as shown.

[0102] Coding strategy for the second figure: In step 1, through alignment, it is found that among the 384 sequences with insertion mutations, the inserted length is less than 50bp and the number of insertions is less than 40. Therefore, the present invention creates a 50*200*1 picture to represent the insertion mutations. If there is no mutation, all are filled with 0; if there is 1 insertion mutation, it is encoded in the first column of this picture, and the remaining columns are filled with 0. If there are two mutations, they are encoded in the first two columns, and so on, up to 50 insertion mutations can be included. If there are more than 50 mutations, it may be a new genotype or the sequence of other viruses. If an insertion mutation has a length of 1bp, then the inserted base is one-hot encoded in rows 1 to 4; if an insertion mutation has a length of 2bp, then the inserted base is one-hot encoded in rows 1 to 8, and so on until 50bp. The coding principle diagram of the second figure with hierarchical ont hot coding is referred toFigure 2 as shown

[0103] 3. Hierarchical Feature Fusion

[0104] Based on the one-hot encoding step, the present invention obtains two pictures. The second picture is an effective supplement to the first picture. Therefore, fusing these two pictures can provide more comprehensive and rich features, thereby improving the final RSV genotype classification accuracy. The schematic diagram of the network architecture is referred to Figure 3 as shown, and the hierarchical feature fusion steps are as follows:

[0105] First, use a CNN network (in this invention, ResNet50 is used) to extract features from the two pictures. The input dimensions of the first picture are all 26740×7×1. The feature is extracted through the backbone layer of the ResNet50 network composed of residual blocks. Each layer has a 3×3 convolutional layer, a Relu activation function, and batch normalization processing to extract features. Since this picture is based on one-hot encoding, it is reduced to a feature of 896×7×1 through a 3-layer convolutional network.

[0106] The input dimensions of the second picture are all 50×200×1. The feature is also extracted through the backbone layer of the ResNet50 network composed of residual blocks. Each layer has a 3×3 convolutional layer, a Relu activation function, and batch normalization processing to extract features. For the convenience of fusion, the second picture also passes through 3 layers of convolution and becomes a feature of 128×7×1. The obtained images are spliced and fused to obtain a final feature map of 1024×7×1 for classification.

[0107] In this invention, a fusion method of feature weighted sum is used to fuse cross-modal features. The weight matrix of the fusion layer is trainable. The initial weight value of multi-layer feature fusion is set to 1, and the feature fusion weight value will be backpropagated and updated together with the classification loss.

[0108] 4. Transformer-Based Multimodal RSV Classification

[0109] In this invention, we explore the Transformer architecture for RSV. The network receives a feature map of size 1024×7×1 as input data. The feature encoding layer Tranformer encoder consists of Multi-head Self-Attention (MSA) and MLP layers proposed by Vaswani. Layer Norm is used before executing the MSA and MLP layers, and residual connections are used after the layers. The MLP contains two fully connected layers with GELU non-linearity.

[0110] In the classification layer, we set 23 categories. Among them, RSV A includes 14 categories, namely GA1, GA2, GA2.1, GA2.2, GA2.3, GA2.3.1, GA2.3.2, GA2.3.2a, GA2.3.2b, GA2.3.3, GA2.3.5, GA3, GA3.0.1, GA3.0.2 in sequence; RSV B includes 9 categories, namely GB1, GB2, GB4, GB5, GB5.0.1, GB5.0.2, GB5.0.3, GB5.0.4a and GB5.0.5a in sequence. According to the probability value of the genotyping, the samples with the maximum genotyping probability value less than 0.95 are defined as the new genotypes of RSV.

[0111] The Transformer architecture for RSV classification constructed in the present invention utilizes the self-attention mechanism of the Transformer to improve the perception ability of the classification model for the global information of the RSV sequence, enhance the modeling ability for complex relationships, and ultimately improve the accuracy of RSV classification.

[0112] Comparative Example 1

[0113] This comparative example performs RSV genotyping based on constructing an evolutionary tree:

[0114] (1) Construct an evolutionary tree based on 10,000 sequences; (2) Use the mafft software for multiple sequence alignment and the iqtree software to construct the evolutionary tree with bootstrap being 1000. Based on the evolutionary tree, use the pplacer software to classify 1325 RSV sequences in the test set and check which branch these RSV sequences belong to.

[0115] Experimental Example 1

[0116] According to the genotyping method provided in Example 1 of the present invention, among 1325 RSV sequences, 1300 sequences are predicted correctly and 15 sequences are predicted wrongly, with an accuracy rate of 98.87%;

[0117] According to the genotyping method provided in Comparative Example 1, 1279 sequences are predicted correctly and 36 sequences are predicted wrongly, with an accuracy rate of 96.53%. And in the evolutionary tree constructed from 10,000 sequences, there are 11 sequences with the problem of clustering errors.

[0118] Through comprehensive comparison, the present invention is superior to the classification based on the evolutionary tree.

[0119] Example 2

[0120] The present invention also provides a device for respiratory syncytial virus genotyping, including: an input module, a control module and an output module;

[0121] The input module is configured to: input the respiratory syncytial virus genome;

[0122] The control module includes: a reference genome construction module, a hierarchical one-hot encoding module, a fusion feature module, and a genotyping module;

[0123] The output module includes: outputting the typing results of respiratory syncytial virus;

[0124] Among them, the reference genome construction module is configured to: divide the sequences into a training set, a validation set, and a test set based on the input respiratory syncytial virus genome; perform sequence alignment on part of the genomes in the training set, find the similar regions and variant regions among all the genomes in the training set, and construct a pan-genome map containing all the genome information based on the alignment results; merge the variant regions and output the reference genome, the training set, the validation set, and the test set;

[0125] The hierarchical one-hot encoding module is configured to: input the whole-genome sequences of the reference genome, the training set, and the validation set, align the whole-genome sequences of the training set and the validation set with the reference genome, and perform one-hot encoding to obtain a picture with query vector * key vector * value vector of (X + 1) * 7 * 1, where X is the length of the reference genome; the key vector in the picture refers to the following seven kinds of information: whether there is an A base, a T base, a C base, a G base, a deletion or an unknown base, an insertion mutation, and whether the base at the position is within the coding gene region of the G protein after alignment with the reference genome; output a picture with query vector * key vector * value vector of (X + 1) * 7 * 1; perform another one-hot encoding on the whole-genome sequences of the training set and the validation set, and output a picture with query vector * key vector * value vector of 50 * 120 to 200 * 1; the key vector in the picture of 50 * 120 to 200 * 1 refers to the number of insertion mutations; if there is at least one insertion mutation, encode it at the front of the picture of 50 * 120 to 200 * 1;

[0126] The fusion feature module is configured to: input the picture of (X + 1) * 7 * 1 and the picture of 50 * 120 to 200 * 1, perform feature extraction using a CNN network, obtain features of the target dimension respectively, splice and fuse the obtained images, and output a fusion feature map;

[0127] The genotyping module is configured to: input the fusion feature map, perform RSV genotyping using a neural network and output the typing results of respiratory syncytial virus.

[0128] Example 3

[0129] The present invention also provides a device for establishing a respiratory syncytial virus typing model, including: an input module, a control module, and an output module;

[0130] The input module is configured to: input the respiratory syncytial virus genome in the database;

[0131] The control module includes: a reference genome construction module, a hierarchical one-hot encoding module, a fusion feature module, and a genotyping module;

[0132] The output module includes: outputting a respiratory syncytial virus genotyping model;

[0133] Among them, the reference genome construction module is configured to: divide the sequences into a training set, a validation set, and a test set based on the input respiratory syncytial virus genome; perform sequence alignment on the genomes in part of the training set, find the similar regions and variant regions among all the genomes in part of the training set, and construct a pan-genome map containing all the genome information based on the alignment results; merge the variant regions and output the reference genome, the training set, the validation set, and the test set;

[0134] The hierarchical one-hot encoding module is configured to: input the whole-genome sequences of the reference genome, the training set, and the validation set, align the whole-genome sequences of the training set and the validation set with the reference genome, and perform one-hot encoding to obtain a picture with query vector * key vector * value vector of (X + 1) * 7 * 1, where X is the length of the reference genome; the key vector in the picture refers to the following seven kinds of information: whether there is an A base, a T base, a C base, a G base, a deletion or an unknown base, an insertion mutation, and whether the base at the position is within the coding gene region of the G protein after alignment with the reference genome; output a picture with query vector * key vector * value vector of (X + 1) * 7 * 1; perform another one-hot encoding on the whole-genome sequences of the training set and the validation set, and output a picture with query vector * key vector * value vector of 50 * 120 - 200 * 1; the key vector in the picture of 50 * 120 - 200 * 1 refers to the number of insertion mutations; if there is at least one insertion mutation, encode it at the front of the picture of 50 * 120 - 200 * 1;

[0135] The fusion feature module is configured to: input the picture of (X + 1) * 7 * 1 and the picture of 50 * 120 - 200 * 1, perform feature extraction using a CNN network, obtain features of the target dimension respectively, splice and fuse the obtained images, and output a fusion feature map;

[0136] The genotyping module is configured to: input the fusion feature map, perform RSV genotyping using a neural network, and output a respiratory syncytial virus genotyping model.

[0137] Example 4

[0138] The present invention also provides a respiratory syncytial virus genotyping device, which includes a processor and a memory. At least one instruction, at least one program, a code set or an instruction set is stored in the memory, and the at least one instruction, at least one program, the code set or the instruction set is loaded and executed by the processor to implement the above-mentioned information processing method for respiratory syncytial virus genotyping.

[0139] Example 5

[0140] The present invention also provides a device for establishing a respiratory syncytial virus genotyping model, which includes a processor and a memory. At least one instruction, at least one program, a code set or an instruction set is stored in the memory, and the at least one instruction, at least one program, the code set or the instruction set is loaded and executed by the processor to implement the above-mentioned method for constructing a respiratory syncytial virus genotyping model.

[0141] Specifically, the electronic device may include a memory, a processor, a bus and a communication interface. The memory, the processor and the communication interface are directly or indirectly electrically connected to each other to achieve data transmission or interaction. For example, these components may be electrically connected to each other through one or more buses or signal lines. The processor can process information and / or data related to target recognition to execute one or more functions described in the present application.

[0142] The memory may be, but is not limited to, a random access memory (RAM), a read only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), etc.

[0143] The processor can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.

[0144] Embodiment 6

[0145] The present invention also provides a computer-readable storage medium, in which at least one instruction, at least one program segment, a code set or an instruction set is stored, and the at least one instruction, at least one program segment, the code set or the instruction set is loaded and executed by a processor to implement the above-mentioned information processing method for respiratory syncytial virus genotyping or the construction method for a respiratory syncytial virus genotyping model.

[0146] The present invention innovatively proposes an information processing method for respiratory syncytial virus genotyping. Based on the original one-hot encoding and zero-padding strategies, a hierarchical one-hot encoding strategy is innovatively proposed, which reduces the interference of the traditional zero-padding method on one-hot encoding and enhances the robustness to indel (insertion) mutations.

[0147] Based on one-hot encoding, the present invention obtains two images, an (X + 1)*7*1 image and a 50*120 - 200*1 image. Based on the proposed CNN network, the features of the two images are extracted and registered and aligned, enhancing the ability to capture mutation information in RSV genomic images, making it easier for CNN to capture detailed information about RSV genomic sequences. There is a stronger correspondence relationship between the features extracted by CNN. Therefore, it is easier to capture the supplementary information for insertion mutations in the second image during the fusion process, improving the final RSV genotype classification accuracy. By comparison, the classification accuracy of the RSV genotype classification method provided by the present invention is better than that of the classification method based on phylogenetic trees, and the clustering error rate is lower.

[0148] The development of the respiratory syncytial virus typing strategy will help doctors and researchers diagnose and evaluate more accurately. By optimizing the feature extraction of respiratory genome sequences and applying deep learning techniques, more accurate diagnostic tools can be provided for the public safety field, helping doctors and researchers better understand the respiratory syncytial virus and providing more effective treatment plans. Based on the hierarchical one-hot encoding strategy and neural network image classification technology, it can help doctors and researchers make accurate diagnoses and treatment guidance, improve work efficiency, and help improve the efficiency and diagnosis and treatment quality of the medical and public safety industries.

[0149] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, various changes and modifications can be made to the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for constructing a respiratory syncytial virus typing model, characterized in that: It includes the following steps: S1: Construction of a reference genome: Respiratory syncytial virus genomes for reference genome construction were selected from the respiratory syncytial virus genome sequence database, and the sequences were divided into training, validation, and test sets. Sequences of the genomes in the training set were aligned to identify similar and variant regions among all genomes in the training set. A pan-genome map containing all genome information was constructed based on the alignment results. The variant regions were merged to obtain the reference genome. S2: Perform hierarchical one-hot encoding: The length of the reference genome is X. The whole genome sequences of the training set and the validation set are aligned with the reference genome and one-hot encoded to obtain an image with the query vector * key vector * value vector of (X+1)*7*1. The key vector in the image refers to the following seven pieces of information: whether there is an A base, T base, C base, G base, missing or unknown base, insertion variation, and whether the base at the position is located in the G protein coding gene region after alignment with the reference genome; The whole genome sequences of the training set and the validation set are once again ont-hot encoded to obtain a query vector * key vector * value vector of 50*120~200*1 images; the key vector in the 50*120~200*1 image refers to the number of insertion variants; if there is at least one insertion variant, it is encoded in the front row of the 50*120~200*1 image; S3: Fusing hierarchical features: Using the CNN network, extract features from the (X+1)*7*1 image and the 50*120~200*1 image respectively to obtain features of the target dimensions. The obtained images are then spliced and fused to obtain a fused feature map. S4: RSV genotyping based on neural network: input fusion feature map, use neural network to perform RSV genotyping and output; S5: Input the full genome of the respiratory syncytial virus in the test set and proceed to steps S2-S4.

2. The method for constructing a respiratory syncytial virus typing model according to claim 1, wherein Step S1 includes merging the variant regions according to the following rules: 1) if a position is only a substitution variant, only the base with the highest variant frequency is retained; 2) if a position has a deletion variant, the non-deleted base is retained; 3) if a position has an insertion variant, the insertion variant is retained.

3. The method for constructing a respiratory syncytial virus typing model according to claim 1, wherein The rules for obtaining the ONT hot encoding of a (X+1)*7*1 image are as follows: Select any 7 lines in the image as the coding lines in the gene region encoding A base, T base, C base, G base, insertion mutation, and G protein respectively; After comparison with the reference genome, whether there is an A base at each position, if there is an A base, the coding line of the A base is coded as 1 at that position; otherwise it is coded as 0; whether there is a T base at each position after comparison with the reference genome, if there is a T base, the coding line of the T base is coded as 1 at that position, otherwise it is 0; whether there is a C base at each position after comparison with the reference genome, if there is a C base, the coding line of the C base is coded as 1 at that position, otherwise it is 0; whether there is a G base at each position after comparison with the reference genome, if there is a G base, the coding line of the G base is coded as 1 at that position, otherwise it is 0; whether there is a missing or unknown base at each position after comparison with the reference genome, if there is a missing or unknown base, the coding line of the missing or unknown base is coded as 1 at that position, otherwise it is 0; whether there is an insertion variation at each position after comparison with the reference genome, if there is an insertion variation, the coding line at that position is coded as 1, otherwise it is 0; whether each position is located in the G protein coding gene region, if so, it is coded as 1, otherwise it is 0.

4. The method for constructing a respiratory syncytial virus typing model according to claim 1, wherein The ONT hot encoding rules for images with sizes from 50*120 to 200*1 are as follows: If the sequence does not have any insertion mutation, it will be filled with all 0s; if the sequence has at least one insertion mutation, it will be encoded in the front row of the 50*120~200*1 picture; If an insertion variant is 1 bp in length, then one-hot encoding is performed on the inserted bases in lines 1 to 4; If an insertion variant is 2 bp in length, then one-hot encoding is performed on the inserted bases in lines 1 to 8, and so on, until an insertion variant is 50 bp in length, then one-hot encoding is performed on the inserted bases in lines 1 to 200.

5. The method for constructing a respiratory syncytial virus typing model according to claim 4, wherein When obtaining a 50*120~200*1 image, first perform a sequence alignment on the whole genome sequence of the training set that did not participate in the construction of the reference genome and / or the RSV whole genome sequence in the validation set with the reference genome, determine the number of inserted mutations in the training set and / or validation set that did not participate in the construction of the reference genome, and adjust the key vector according to the number of inserted mutations.

6. The method for constructing a respiratory syncytial virus typing model according to claim 5, wherein When step S3 fuses the hierarchical features, the ResNet50 network backbone layer is used to extract features. Each layer has a 3×3 convolution layer, ReLU activation function and batch normalization processing to extract features.

7. The method for constructing a respiratory syncytial virus typing model according to claim 1, wherein The neural network is selected from BP neural network, RNN neural network, LSTM neural network, GRU neural network or Transformer architecture.

8. The method for constructing a respiratory syncytial virus typing model according to claim 7, wherein In step S4, the fused feature map is input, and RSV genotyping is performed and output using the Transformer architecture; the feature encoding layer Tranformer encoder consists of a Multi-head Self-Attention and an MLP layer; the Multi-head Self-Attention is MSA.

9. The method for constructing a respiratory syncytial virus typing model according to claim 8, wherein Layer Norm normalization is used before executing the MSA and MLP layers, and residual connections are used after the layers.

10. The method for constructing a respiratory syncytial virus typing model according to claim 8, wherein If the maximum probability of typing after the whole genome sequence of the tested sample is compared with the reference genome is less than 0.95, the tested sample is judged to be a new RSV genotype.

11. The method for constructing a respiratory syncytial virus typing model according to claim 10, wherein: When RSV genotyping, RSV type A is divided into GA1, GA2, GA2.1, GA2.2, GA2.3, GA2.3.1, GA2.3.2, GA2.3.2a, GA2.3.2b, GA2.3.3, GA2.3.5, GA3, GA3.0.1 and GA3.0.2; RSV type B is divided into GB1, GB2, GB4, GB5, GB5.0.1, GB5.0.2, GB5.0.3, GB5.0.4a and GB5.0.5a.

12. An information processing method for respiratory syncytial virus typing, characterized in that: It includes the following steps: A respiratory syncytial virus typing model is constructed according to the method for constructing a respiratory syncytial virus typing model according to any one of claims 1 to 11, and then the full genome of the respiratory syncytial virus of the sample to be tested is input to perform steps S2 to S4.

13. A device for typing respiratory syncytial virus, characterized in that: include: Input module, control module and output module; The input module is configured to: input a respiratory syncytial virus genome; The control module includes: a reference genome construction module, a hierarchical one hot encoding module, a fusion feature module and a genotyping module; The output module includes: outputting respiratory syncytial virus typing results; The reference genome construction module is configured to: divide the sequence into a training set, a validation set, and a test set based on the input respiratory syncytial virus genome; perform sequence alignment on the genomes in part of the training set to find similar regions and variant regions between all genomes in the part of the training set, and construct a pan-genome map containing all genome information based on the alignment results; merge the variant regions to output the reference genome, training set, validation set, and test set; The hierarchical one hot encoding module is configured as follows: inputting the whole genome sequences of the reference genome, the training set, and the validation set, aligning the whole genome sequences of the training set and the validation set with the reference genome, and performing one hot encoding to obtain a picture with the query vector * key vector * value vector of (X+1)*7*1, where X is the length of the reference genome; the key vector in the picture refers to the following seven types of information: whether there is an A base, a T base, a C base, a G base, a missing or unknown base, an insertion variation, and whether the base at the position is located in the coding gene region of the G protein at each position after alignment with the reference genome; outputting a picture with the query vector * key vector * value vector of (X+1)*7*1; performing one hot encoding on the whole genome sequences of the training set and the validation set again. Hot encoding outputs a query vector * key vector * value vector of a 50*120~200*1 image. The key vector in the 50*120~200*1 image indicates the number of insertion mutations. If there is at least one insertion mutation, it is encoded in the first column of the 50*120~200*1 image. The fusion feature module is configured to: input the (X+1)*7*1 image and the 50*120~200*1 image, use the CNN network to perform feature extraction, obtain the features of the target dimension respectively, splice and fuse the obtained images, and output a fusion feature map; The genotyping module is configured to: input the fusion feature map, perform RSV genotyping using a neural network, and output a respiratory syncytial virus typing result.

14. A respiratory syncytial virus typing device, characterized in that: The device includes a processor and a memory, wherein the memory stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set is loaded and executed by the processor to implement the information processing method for respiratory syncytial virus typing as described in claim 12.

15. A computer-readable storage medium, characterized in that The storage medium stores at least one instruction, at least one program, a code set or an instruction set, and the at least one instruction, the at least one program, the code set or the instruction set are loaded and executed by the processor to implement the information processing method for respiratory syncytial virus typing as described in claim 12.

Images