B cell linear epitope prediction method and device and storage medium

Abstract

The invention discloses a B cell linear epitope prediction method and device and a storage medium, and belongs to the technical field of immunity and drug development. The method for predicting the linear epitope of the B cell comprises the following steps: respectively preprocessing sequences to be detected to obtain target input sequences; performing feature coding on each amino acid residue in the obtained target input sequence, and outputting a coding matrix of each sequence; inputting the obtained local feature matrix into a two-layer or five-layer BiGRU module in a trained B cell linear epitope prediction model, and outputting a global feature matrix of each sequence; and inputting the obtained global feature matrix of each sequence into three full connection layers in the B cell linear epitope prediction model, and outputting the probability that the to-be-detected sequence is the B cell epitope. In addition, the invention also provides a B cell linear epitope prediction device and a storage medium. According to the prediction method provided by the invention, the prediction accuracy of the linear epitope of the B cell is remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of immunity and drug development, and in particular to a method, device and storage medium for predicting B cell linear epitopes. Background Art

[0002] Adaptive immunity is composed of B cells and T cells, which recognize antigens of varying specificity. B cells primarily stimulate humoral immunity. B cells play a crucial role in immunity by secreting antibodies. B cell antigenic determinants (B cell epitopes) are protein fragments present on the surface of antigens that are specifically bound by B cell receptors and trigger an immune response. Epitopes can be divided into linear epitopes and conformational epitopes based on their structural characteristics. The main difference between them is whether the amino acids that make up the epitope are continuous on the antigen protein. Experimental methods for identifying B cell epitopes include peptide microarrays, X-ray crystallography, and enzyme-linked immunosorbent assays (ELISAs). However, these methods are often time-consuming, costly, and inefficient. Using computational methods to predict B cell epitopes can greatly accelerate the screening of B cell epitopes, improve the efficiency of immunotherapy, and accelerate the development of related drugs.

[0003] Numerous computational methods have been developed for B cell epitope screening. Linear epitope prediction methods are relatively old, having been around since the 1980s. The earliest models for linear B cell epitopes were based on propensity scales. Since the turn of the century, numerous machine learning-based models have emerged, such as ABCPred, BCPreds, SVMTriP, BepiPred-2.0, and epitope1D. In recent years, models using deep learning to model linear B cell epitopes have emerged. For example, EpiDope uses context-sensitive embeddings combined with ELMo+LSTM; NetBCE uses a CNN and BLSTM architecture for linear epitope prediction; and CALIBER utilizes the ESM-2 pre-trained model combined with a BLSTM model to simultaneously predict linear and conformational epitopes. However, current models often only utilize sequence amino acid composition and k-mer features during the encoding phase, failing to incorporate amino acid positional information and the physicochemical characteristics of each position into the modeling process. This results in poor prediction performance. Moreover, most current methods use machine learning methods, and there may be many hidden features that have not been fully explored. Summary of the Invention

[0004] The purpose of the present invention is to overcome the above technical deficiencies and provide a method, device and storage medium for predicting B cell linear epitopes, so as to solve the technical problem of how to improve the prediction accuracy of B cell linear epitopes in the prior art.

[0005] To achieve the above technical objectives, the technical solution of the present invention provides a method for predicting B cell linear epitopes, comprising the following steps:

[0006] S1. Obtain a sequence to be tested, and preprocess the sequence to be tested to obtain a target input sequence;

[0007] S2. Feature encoding each amino acid residue in the target input sequence obtained in step S1, and outputting a coding matrix for each sequence;

[0008] S3, inputting the encoding matrix obtained in step S2 into a two-layer convolutional module in the B cell linear epitope prediction model, and outputting a local feature matrix for each sequence;

[0009] S4. Input the local feature matrix obtained in step S3 into the 2-layer or 5-layer BiGRU module in the trained B cell linear epitope prediction model, and output the global feature matrix of each sequence;

[0010] S5. Input the obtained global feature matrix of each sequence into the three fully connected layers in the B cell linear epitope prediction model, and output the probability that the sequence to be tested is a B cell epitope.

[0011] In any embodiment, before step S5, the method further includes inputting the global feature matrix obtained in step S4 into a self-attention module in the B cell linear epitope prediction model, and outputting a feature matrix after feature rearrangement;

[0012] The feature matrix obtained by rearranging the features of each sequence is input into the three-layer fully connected layer in the B cell linear epitope prediction model, and the probability that the sequence to be tested is a B cell epitope is output.

[0013] In any embodiment, in step S1, the preprocessing includes: for sequences with more than 25 amino acids, cutting off the first 25 amino acids as the target input sequence; for sequences with less than 25 amino acids, appending "X" amino acids to the end of the sequence to make the length 25.

[0014] In any embodiment, in step S2, the feature coding includes physical and chemical feature coding, which is based on 28 features extracted from the AAIndex database that have been successfully applied in the field of protein binding as the physical and chemical feature coding of each amino acid residue in the target input sequence.

[0015] In any embodiment, in step S2, the 28 features that have been successfully applied to the field of protein binding are: CHOP780202, CIDH920103, CIDH920105, FAUJ880109, FAUJ880111, FINA910104, GEIM800104, GEIM800106, KANM800102, KLEP840101, KRIW710101, LIFS790101, ME EJ800101, OOBM770102, PALJ810107, QLAN880123, RACS770103, RADA880108, ROSM880102, SWER830101, Z IMJ680102, ZIMJ680104, AURR980120, MUNV940103, NADH010104, NADH010106, GUYH850105, MIYS990104.

[0016] In any embodiment, in step S3, each layer in the 2-layer convolution module includes a convolutional network, a batch normalization layer, a ReLU layer and a maximum pooling layer.

[0017] In any embodiment, in step S5, the self-attention model redistributes the weights of the feature vectors of each subsequence in the global feature matrix of the sequence, and outputs a feature matrix after the features are rearranged.

[0018] In any embodiment, in step S6, the three fully connected layers include two layers of linear layers and ReLU layers connected in series, and one layer of linear layer and Sigmoid layer; the feature matrix is ​​input into the two layers of linear layers and ReLU layers connected in series; and then passes through one layer of linear layer and Sigmoid layer to obtain the probability that the input sequence is a B cell epitope.

[0019] In addition, the present invention also provides a prediction device for B cell linear epitopes, comprising:

[0020] A data acquisition and preprocessing unit is used to acquire a sequence to be tested and preprocess the sequence to be tested to obtain a target input sequence;

[0021] a feature encoding unit, configured to perform feature encoding on each amino acid residue in the obtained target input sequence and output an encoding matrix for each sequence;

[0022] The first feature representation unit is used to input the obtained encoding matrix into a 2-layer or 5-layer convolution module in the B cell linear epitope prediction model and output a local feature matrix for each sequence;

[0023] The second feature representation unit is used to input the obtained local feature matrix into the two-layer bidirectional GRU module in the trained B cell linear epitope prediction model and output the global feature matrix of each sequence;

[0024] The prediction unit is used to input the obtained feature matrix of each sequence into the three fully connected layers in the B cell linear epitope prediction model, and output the probability that the sequence to be tested is a B cell epitope.

[0025] In addition, the present invention also provides a storage medium storing a computer program, which implements the above-mentioned method for predicting B cell linear epitopes when executed by a processor.

[0026] Compared with the prior art, the present invention has the following advantages: The method for predicting B-cell linear epitopes proposed in the present invention comprises the following steps: obtaining a test sequence and preprocessing the test sequence to obtain a target input sequence; S2, encoding each amino acid residue in the obtained target input sequence, outputting an encoding matrix for each sequence; inputting the obtained encoding matrix into a two-layer convolutional module in a B-cell linear epitope prediction model, outputting a local feature matrix for each sequence; inputting the obtained local feature matrix into a two-layer or five-layer bidirectional GRU module in a trained B-cell linear epitope prediction model, outputting a global feature matrix for each sequence; inputting the obtained global feature matrix into a self-attention module in the B-cell linear epitope prediction model, outputting a feature matrix after feature rearrangement; and inputting the obtained rearranged feature matrix for each sequence into a three-layer fully connected layer in the B-cell linear epitope prediction model, outputting the probability that the test sequence is a B-cell epitope. The method proposed in the present invention incorporates amino acid physicochemical feature encoding and a CNN+BiGRU deep learning model architecture. This method achieved performance superior to the original model in both data environments.

[0027] A further optimized approach, which incorporated amino acid physicochemical feature encoding and a deep learning model architecture combining CNN, BiGRU, and Attention, achieved state-of-the-art results across the majority of six independent test sets. Finally, the method successfully predicted epitope datasets for two severe acute respiratory syndrome coronavirus (SARS-CoV) datasets. Of the 10 SARS-CoV-1 epitopes, seven achieved prediction scores exceeding 99.9%. These multiple results demonstrate that this method is currently the best-performing linear B cell epitope prediction tool. The prediction method proposed in this invention significantly improves the accuracy of B cell linear epitope prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 Schematic diagram of the process of predicting B cell linear epitopes in Example 1 of the present invention.

[0029] Figure 2 This is a comparison chart of the ablation experiment results of Example 1 of the present invention.

[0030] Figure 3 1 is the AUROC and AUPR diagram of Example 1 of the present invention and the comparative example method. DETAILED DESCRIPTION

[0031] The "ranges" disclosed herein are defined in terms of lower and upper limits, where a given range is defined by selecting a lower limit and an upper limit, and the selected lower and upper limits define the boundaries of the particular range. Ranges defined in this manner can be inclusive or exclusive of the end values ​​and can be combined arbitrarily, i.e., any lower limit can be combined with any upper limit to form a range. For example, if ranges of 60 to 120 and 80 to 110 are listed for a particular parameter, it is understood that ranges of 60 to 110 and 80 to 120 are also contemplated. Furthermore, if minimum range values ​​of 1 and 2 are listed, and if maximum range values ​​of 3, 4, and 5 are listed, the following ranges are all contemplated: 1 to 3, 1 to 4, 1 to 5, 2 to 3, 2 to 4, and 2 to 5. In this application, unless otherwise indicated, the numerical range "a to b" is a shorthand representation of any combination of real numbers between a and b, where a and b are both real numbers. For example, a numerical range of "0-5" indicates that all real numbers between "0-5" are listed herein, and "0-5" is simply an abbreviation for these numerical combinations. Furthermore, when a parameter is expressed as an integer ≥ 2, this is equivalent to disclosing that the parameter is, for example, an integer of 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, etc.

[0032] Unless otherwise specified, the terms "include" and "comprising" used in this application may be open-ended or closed-ended. For example, "include" and "comprising" may mean that other components not listed may also be included or that only the listed components are included.

[0033] Unless otherwise specified, the term "or" is used in this application to be inclusive. For example, the phrase "A or B" means "A, B, or both A and B." More specifically, the condition "A or B" is satisfied if any of the following conditions are met: A is true (or exists) and B is false (or does not exist); A is false (or does not exist) and B is true (or exists); or both A and B are true (or exist).

[0034] This specific embodiment provides a method for predicting B cell linear epitopes, comprising the following steps:

[0035] S1. Obtain a sequence to be tested, and preprocess the sequence to be tested to obtain a target input sequence; the preprocessing includes: for a sequence with more than 25 amino acids, truncating the first 25 amino acids as the target input sequence; for a sequence with less than 25 amino acids, appending "X" amino acids to the end of the sequence to a length of 25;

[0036] S2. Perform feature coding on each amino acid residue in the target input sequence obtained in step S1, and output a coding matrix for each sequence; the feature coding includes physical and chemical feature coding, and the physical and chemical feature coding of each amino acid residue in the target input sequence is based on 28 features that have been successfully applied to the field of protein binding extracted from the AAIndex database; the 28 features that have been successfully applied to the field of protein binding are: CHOP780202, CIDH920103, CIDH920105, FAUJ880109, FAUJ880111, FINA910104, GEIM800104, G EIM800106, KANM800102, KLEP840101, KRIW710101, LIFS790101, MEEJ800101, OOBM770102, PALJ810107, QLAN880123, RACS770103, RADA88 0108, ROSM880102, SWER830101, ZIMJ680102, ZIMJ680104, AURR980120, MUNV940103, NADH010104, NADH010106, GUYH850105, MIYS990104;

[0037] S3, inputting the encoding matrix obtained in step S2 into a two-layer convolution module in the B cell linear epitope prediction model, and outputting a local feature matrix for each sequence; each layer in the two-layer convolution module includes a convolutional network, a batch normalization layer, a ReLU layer, and a maximum pooling layer;

[0038] S4, inputting the local feature matrix obtained in step S3 into a 2-layer or 5-layer bidirectional GRU module in the trained B cell linear epitope prediction model, and outputting a global feature matrix for each sequence; the 2-layer bidirectional GRU module includes a series of bidirectional GRU layers;

[0039] S5. Input the rearranged feature matrix of each sequence into the three-layer fully connected layer in the B cell linear epitope prediction model, and output the probability that the sequence to be tested is a B cell epitope; the three-layer fully connected layer includes two layers of linear layers and ReLU layers in series, and one layer of linear layer and Sigmoid layer; input the feature matrix into the two layers of linear layers and ReLU layers in series; and then pass through one layer of linear layer and Sigmoid layer to obtain the probability that the input sequence is a B cell epitope.

[0040] In some embodiments, before step S5, the method further includes: inputting the obtained global feature matrix into the self-attention module in the B-cell linear epitope prediction model, and outputting the feature matrix after feature rearrangement; the self-attention model redistributes the weights of the feature vectors of each subsequence in the global feature matrix of the sequence, and outputs the feature matrix after feature rearrangement; then inputting the obtained feature matrix after feature rearrangement of each sequence into the three-layer fully connected layer in the B-cell linear epitope prediction model, and outputting the probability that the sequence to be tested is a B cell epitope; further, the three-layer fully connected layer includes two layers of linear layers and ReLU layers connected in series, and one layer of linear layer and Sigmoid layer; inputting the rearranged feature matrix into the two layers of linear layers and ReLU layers connected in series; and then passing through one layer of linear layer and Sigmoid layer to obtain the probability that the input sequence is a B cell epitope.

[0041] In addition, this specific embodiment also proposes a prediction device for B cell linear epitopes, comprising:

[0042] A data acquisition and preprocessing unit is used to acquire a sequence to be tested and preprocess the sequence to be tested to obtain a target input sequence;

[0043] a feature encoding unit, configured to perform feature encoding on each amino acid residue in the obtained target input sequence and output an encoding matrix for each sequence;

[0044] The first feature representation unit is used to input the obtained encoding matrix into the two-layer convolution module in the B cell linear epitope prediction model and output the local feature matrix of each sequence;

[0045] The second feature representation unit is used to input the obtained local feature matrix into the 2-layer or 5-layer bidirectional GRU module in the trained B cell linear epitope prediction model, and output the global feature matrix of each sequence;

[0046] The prediction unit is used to input the rearranged feature matrix of each sequence obtained into the three fully connected layers in the B cell linear epitope prediction model, and output the probability that the sequence to be tested is a B cell epitope.

[0047] In some embodiments, the above-mentioned device further includes: a feature rearrangement unit, which is used to input the obtained global feature matrix into the self-attention module in the B cell linear epitope prediction model, and output the feature matrix after feature rearrangement.

[0048] In addition, this specific embodiment also proposes a storage medium on which a computer program is stored. When the program is executed by a processor, the above-mentioned method for predicting B cell linear epitopes is implemented.

[0049] It should be noted that B cell epitopes (determinants) are composed of hydrophilic amino acids on the surface of antigens, which are easily accessible to B cell receptors (BCR) and antibody molecules and recognized. They are generally composed of 3 to 5 spatially adjacent, continuous or discontinuous amino acid residues.

[0050] Epitopes are specialized chemical groups within an antigen molecule that determine its specificity, including B-cell and T-cell epitopes. B-cell epitopes are fragments that can be specifically recognized and bound by B-cell surface receptors or antibodies. Depending on their structural characteristics, B-cell epitopes can be classified as either linear or conformational.

[0051] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0052] In the present invention, references to “some embodiments”, “this embodiment”, examples, etc. describe a subset of all possible embodiments, but it can be understood that “some embodiments” may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict.

[0053] If similar descriptions of "first / second" appear in the application documents, the following explanation is added. In the following description, the terms "first\second\third" are merely used to distinguish similar objects and do not represent a specific order of the objects. It can be understood that "first\second\third" can be interchanged with the specific order or sequence where permitted, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein.

[0054] In this embodiment, the term "and / or" is merely a description of the association relationship between associated objects, indicating that three relationships may exist. For example, object A and / or object B may represent three situations: object A exists alone, object A and object B exist at the same time, and object B exists alone.

[0055] Below, the embodiment of the present application is described. The embodiment described below is exemplary and is only used to explain the present application, and is not to be construed as limiting the present application. Where specific techniques or conditions are not specified in the embodiments, the techniques or conditions described in the literature in this area or the product specifications are used. Reagents or instruments used that do not specify the manufacturer are conventional products that can be obtained commercially.

[0056] Example 1

[0057] Combine Figure 1 This embodiment proposes a B cell linear epitope prediction method based on convolution and gating networks, comprising the following steps:

[0058] Step 1: obtain a sequence to be tested, and preprocess the sequence to be tested to obtain a target input sequence;

[0059] For sequences with more than 25 amino acids, only the first 25 amino acids were truncated as the target input sequence;

[0060] For sequences with less than 25 amino acids, "X" amino acids were appended to the end of the sequence to make the length 25.

[0061] Step 2: Feature encode each amino acid residue in the target input sequence and output the encoding matrix of each sequence;

[0062] 28 features that have been successfully applied in the field of protein binding were extracted from the AAIndex database as the encoding of the physical and chemical characteristics of each amino acid residue in the target input sequence; the 28 features that have been successfully applied in the field of protein binding are: CHOP780202, CIDH920103, CIDH920105, FAUJ880109, FAUJ880111, FINA910104, GEIM800104, GEIM800106, KANM800102, KLEP840101, KRIW7 10101, LIFS790101, MEEJ800101, OOBM770102, PALJ810107, QLAN880123, RACS770103, RADA880108, ROSM880102, SWER830101, ZIMJ680102, ZIMJ680104, AURR980120, MUNV940103, NADH010104, NADH010106, GUYH850105, MIYS990104; as shown in Table 1.

[0063] Step 3: Input the encoding matrix into the two-layer convolution module in the trained B cell linear epitope prediction model, and output the local feature matrix of each sequence;

[0064] The encoding matrix of the sequence is input into the 2-layer convolutional layer to obtain the local feature matrix of the epitope;

[0065] Each of the 2 convolutional layers contains a convolutional network, a batch normalization layer, a ReLU layer, and a maximum pooling layer.

[0066] Step 4: Input the local feature matrix into the two-layer bidirectional GRU module in the trained B cell linear epitope prediction model, and output the global feature matrix of each sequence;

[0067] The local feature matrix of the sequence is input into the series of bidirectional GRU layers to obtain the global feature matrix.

[0068] Step 5: Input the global feature matrix into the self-attention module in the trained B cell linear epitope prediction model, and output the feature matrix after feature rearrangement;

[0069] The self-attention model is used to redistribute the weights of the feature vectors of each subsequence in the global feature matrix of the sequence.

[0070] In step 6, the rearranged feature matrix of each sequence is input into the three-layer fully connected layer in the B cell linear epitope prediction model to output the probability that the sequence to be tested is a B cell epitope; the rearranged feature matrix is ​​input into two layers of linear layers and ReLU layers in series; and then through one layer of linear layer and Sigmoid layer, the value is mapped to [0, 1], indicating the probability that the input sequence is a B cell epitope.

[0071] For example, if the sequence to be tested is "PLMESELVIGAVIIRGHLRMA", the model will eventually output that the probability of it being a B cell linear epitope is 0.7944.

[0072] Table 1 Coding of the physical and chemical characteristics of each amino acid

[0073]

[0074] In addition, this embodiment also provides a prediction device for B cell linear epitopes, comprising:

[0075] A data acquisition and preprocessing unit is used to acquire a sequence to be tested and preprocess the sequence to be tested to obtain a target input sequence;

[0076] a feature encoding unit, configured to perform feature encoding on each amino acid residue in the obtained target input sequence and output an encoding matrix for each sequence;

[0077] The first feature representation unit is used to input the obtained encoding matrix into the two-layer convolution module in the B cell linear epitope prediction model and output the local feature matrix of each sequence;

[0078] The second feature representation unit is used to input the obtained local feature matrix into the two-layer bidirectional GRU module in the trained B cell linear epitope prediction model and output the global feature matrix of each sequence;

[0079] A feature rearrangement unit is used to input the obtained global feature matrix into the self-attention module in the B cell linear epitope prediction model and output the feature matrix after feature rearrangement;

[0080] The prediction unit is used to input the rearranged feature matrix of each sequence obtained into the three fully connected layers in the B cell linear epitope prediction model, and output the probability that the sequence to be tested is a B cell epitope.

[0081] In addition, this embodiment also provides a storage medium storing a computer program, which implements the above-mentioned method for predicting B cell linear epitopes when executed by a processor.

[0082] Example 2

[0083] This embodiment proposes a B cell linear epitope prediction method based on convolution and gating networks. The difference from Example 1 is that a 5-layer bidirectional GRU module is used in step 4, and the other steps are the same as Example 1.

[0084] Example 3

[0085] This example proposes a B cell linear epitope prediction method based on convolution and gating networks. The difference from Example 2 is that step 5 of Example 2 is omitted, and the method includes the following steps:

[0086] Step 1: obtain a sequence to be tested, and preprocess the sequence to be tested to obtain a target input sequence;

[0087] For sequences with more than 25 amino acids, only the first 25 amino acids were truncated as the target input sequence;

[0088] For sequences with less than 25 amino acids, "X" amino acids were appended to the end of the sequence to make the length 25.

[0089] Step 2: Feature encode each amino acid residue in the target input sequence and output the encoding matrix of each sequence;

[0090] 28 features that have been successfully applied in the field of protein binding were extracted from the AAIndex database as the encoding of the physical and chemical characteristics of each amino acid residue in the target input sequence; the 28 features that have been successfully applied in the field of protein binding are: CHOP780202, CIDH920103, CIDH920105, FAUJ880109, FAUJ880111, FINA910104, GEIM800104, GEIM800106, KANM800102, KLEP840101, KRIW7 10101, LIFS790101, MEEJ800101, OOBM770102, PALJ810107, QLAN880123, RACS770103, RADA880108, ROSM880102, SWER830101, ZIMJ680102, ZIMJ680104, AURR980120, MUNV940103, NADH010104, NADH010106, GUYH850105, MIYS990104; as shown in Table 1.

[0091] Step 3: Input the encoding matrix into the two-layer convolution module in the trained B cell linear epitope prediction model, and output the local feature matrix of each sequence;

[0092] The encoding matrix of the sequence is input into the 2-layer convolutional layer to obtain the local feature matrix of the epitope;

[0093] Each of the 2 convolutional layers contains a convolutional network, a batch normalization layer, a ReLU layer, and a maximum pooling layer.

[0094] Step 4: Input the local feature matrix into the 5-layer bidirectional GRU module in the trained B cell linear epitope prediction model, and output the global feature matrix of each sequence;

[0095] The local feature matrix of the sequence is input into the series of bidirectional GRU layers to obtain the global feature matrix.

[0096] In step 5, the feature matrix of each sequence is input into the three-layer fully connected layer in the B cell linear epitope prediction model, and the probability that the sequence to be tested is a B cell epitope is output; the rearranged feature matrix is ​​input into two layers of linear layers and ReLU layers in series; and then through one layer of linear layer and Sigmoid layer, the value is mapped to [0, 1], indicating the probability that the input sequence is a B cell epitope.

[0097] Experimental verification

[0098] In order to verify the necessity of each module of the prediction method in this embodiment, we first conducted an ablation experiment. We selected the convolution unit and the gate unit as the objects of ablation. The experiment was conducted using the training and test sets of epitope1D. Figure 2 A. When the number of CNN layers is eliminated or reduced to 1, the results of our method will decrease to varying degrees. Figure 2 In B, we compared a wide range of BiGRU (i.e., bidirectional GRU) layer numbers. The results showed that when the BiGRU layer number was set to 5 (i.e., Example 2), all indicators achieved relatively optimal results. After comprehensive comparison, we selected a model that retained 2 BiGRU layers (our method 1, i.e., the prediction method of Example 1) and a model that removed the self-attention module (our method 2, i.e., the prediction method of Example 3) as the final result comparison tools.

[0099] We trained and tested our method using the datasets of epitope1D and NetBCE respectively, which ensured an absolutely fair comparison of the results. First, we trained and tested it using the data of epitope1D, and the results are shown in Figure 3 Left. Our method 1 and our method 2 achieved 0.9022 and 0.9335 on AUC, respectively, exceeding epitope1D by 4.74% and 8.38%. Furthermore, our method 1 and our method 2 achieved 0.0526 and 0.0640 on AUC10%, respectively, exceeding epitope1D by 56.55% and 90.40%. On AUPR, our method 1 and our method 2 achieved scores of 0.7023 and 0.7878, respectively exceeding epitope1D by 26.61% and 42.02%. We then retrained and tested the model on NetBCE data, and the results are shown in the table. Figure 3 Right. Our method 1 achieved scores of 0.8939, 0.0514, and 0.7224 in AUC, AUC10%, and AUPR, respectively, outperforming NetBCE by 7.14%, 46.98%, and 22.67%, respectively. Our method 2 achieved scores of 0.9058, 0.0729, and 0.8458 in AUC, AUC10%, and AUPR, respectively, outperforming NetBCE by 8.57%, 108.33%, and 43.63%, respectively. Our method achieved superior results on both dataset acquisition methods, demonstrating that it is indeed a superior model architecture.

[0100] We compare the results of our method on two independent test sets.

[0101] On the iBCE-EL test set, our method achieves a significant lead. Our method 1 (NetBCE) leads EpiDope, BepiPred-2.0, CLBTope, and EpitopeVec by 114.75%, 92.01%, 86.52%, and 63.12% in AUC, respectively. It also leads EpiDope, BepiPred-2.0, CLBTope, and EpitopeVec by 3167.26%, 738.89%, 647.93%, and 621.84% in AUC10%. It also leads EpiDope, BepiPred-2.0, CLBTope, and EpitopeVec by 133.88%, 100.02%, 87.38%, and 76.63% in AUPR.

[0102] In the LBTope dataset, all epitope sequences are of length 20. The LBTope dataset contains 7,824 positive samples and 7,853 negative samples. Our method 1 (NetBCE) achieved the best prediction results, with scores of 0.6158, 0.0117, and 0.6077 in AUC, AUC10%, and AUPR, respectively. Our method 1 (NetBCE) outperformed EpiDope, BepiPred-2.0, CLBTope, and EpitopeVec by 40.20%, 17.42%, 8.64%, and 12.30% in AUC, respectively. In terms of AUC10%, it outperformed EpiDope, BepiPred-2.0, CLBTope, and EpitopeVec by 318.95%, 139.40%, 22.19%, and 60.69%. In terms of AUPR, our method leads EpiDope, BepiPred-2.0, CLBTope, and EpitopeVec by 34.83%, 19.13%, 5.92%, and 11.58%, respectively. See Table 2 for details. Furthermore, Examples 2 and 3 of our method also achieve superior results compared to the comparison tools.

[0103] Table 2

[0104]

[0105] We collected 10 experimentally validated SARS-CoV-1 epitopes. These 10 sequences are all B-cell linear epitopes. Our method 1 (NetBCE) achieved a probability of over 99.9% for seven of these sequences. Our method 1 (NetBCE) also predicted the remaining three epitopes with a probability of around 80%. The results are shown in Table 3.

[0106] Table 3

[0107]

[0108] The present invention discloses a method for predicting B cell linear epitopes based on a convolutional and gating network. The method comprises: first, encoding the sequence to be tested using a feature matrix based on the biochemical properties of amino acids; then, these encoded matrices are fed into a two-layer convolutional unit. After exiting the convolution module, these matrices are fed into a gating module, which consists of a two-layer BiGRU or a five-layer BiGRU. Subsequently, the sequence is fed into a self-attention module. Finally, the sequence is fed into an output module, which maps the result of each sequence to a value between [0, 1] as the final output: the probability of whether it is a B cell epitope. The present invention can be used as a method for predicting B cell linear epitopes, with good prediction accuracy.

[0109] The specific embodiments of the present invention described above do not limit the scope of protection of the present invention. Any other corresponding changes and modifications made based on the technical concept of the present invention should be included in the scope of protection of the claims of the present invention.

Claims

1. A method for predicting B cell linear epitopes, characterized in that: The following steps are involved: S1, obtaining a sequence to be tested, and preprocessing the sequence to be tested to obtain a target input sequence; S2, performing feature encoding on each amino acid residue in the target input sequence obtained in step S1, and outputting the encoding matrix of each sequence; S3, inputting the encoding matrix obtained in step S2 into a 2-layer convolution module in the B cell linear epitope prediction model, and outputting a local feature matrix of each sequence; S4, inputting the local feature matrix obtained in step S3 into the 2-layer or 5-layer BiGRU module in the trained B cell linear epitope prediction model, and outputting the global feature matrix of each sequence; S5. Input the obtained global feature matrix of each sequence into the three fully connected layers in the B cell linear epitope prediction model, and output the probability that the sequence to be tested is a B cell epitope.

2. The method for predicting B cell linear epitopes according to claim 1, characterized in that: Before step S5, the method further includes inputting the global feature matrix obtained in step S4 into a self-attention module in the B cell linear epitope prediction model, and outputting a feature matrix after feature rearrangement; The feature matrix obtained by rearranging the features of each sequence is input into the three fully connected layers in the B cell linear epitope prediction model, and the probability that the sequence to be tested is a B cell epitope is output.

3. The method for predicting B cell linear epitopes according to claim 1, characterized in that: In step S1, the preprocessing includes: for sequences with more than 25 amino acids, the first 25 amino acids are cut off as the target input sequence; for sequences with less than 25 amino acids, "X" amino acids are added to the end of the sequence to a length of 25.

4. The method for predicting B cell linear epitopes according to claim 1, characterized in that: In step S2, the feature coding includes physical and chemical feature coding, and 28 features that have been successfully applied to the protein binding field extracted from the AAIndex database are used as the physical and chemical feature coding of each amino acid residue in the target input sequence.

5. The method for predicting B cell linear epitopes according to claim 4, characterized in that: In step S2, the 28 features that have been successfully applied to the field of protein binding are: CHOP780202, CIDH920103, CIDH920105, FAUJ880109, FAUJ880111, FINA910104, GEIM800104, GEIM800106, KANM800102, KLEP840101, KRIW710101, LIFS790101, MEEJ800 101. OOBM770102, PALJ810107, QLAN880123, RACS770103, RADA880108, ROSM880102, SWER830101, ZIM J680102, ZIMJ680104, AURR980120, MUNV940103, NADH010104, NADH010106, GUYH850105, MIYS990104.

6. The method for predicting B cell linear epitopes according to claim 1, characterized in that: In step S3, each layer in the 2-layer convolution module includes a convolutional network, a batch normalization layer, a ReLU layer, and a maximum pooling layer.

7. The method for predicting B cell linear epitopes according to claim 2, characterized in that: The self-attention model redistributes the weights of the feature vectors of each subsequence in the global feature matrix of the sequence and outputs a feature matrix after the features are rearranged.

8. The method for predicting B cell linear epitopes according to claim 1, characterized in that: In step S5, the three fully connected layers include two layers of linear layers and ReLU layers connected in series, and one layer of linear layer and Sigmoid layer; the feature matrix is ​​input into the two layers of linear layers and ReLU layers connected in series; and then passes through one layer of linear layer and Sigmoid layer to obtain the probability that the input sequence is a B cell epitope.

9. A prediction device for B cell linear epitopes, characterized in that: include: A data acquisition and preprocessing unit, used to acquire a sequence to be tested, and preprocess the sequence to be tested to obtain a target input sequence; A feature encoding unit, used for performing feature encoding on each amino acid residue in the obtained target input sequence and outputting a coding matrix for each sequence; The first feature representation unit is used to input the obtained encoding matrix into the 2-layer convolution module in the B cell linear epitope prediction model, and output the local feature matrix of each sequence; The second feature representation unit is used to input the obtained local feature matrix into a 2-layer or 5-layer bidirectional GRU module in the trained B cell linear epitope prediction model, and output a global feature matrix for each sequence; The prediction unit is used to input the obtained feature matrix of each sequence into the three fully connected layers in the B cell linear epitope prediction model, and output the probability that the sequence to be tested is a B cell epitope.

10. A storage medium, characterized in that: A computer program is stored thereon, and when the program is executed by a processor, the method for predicting B cell linear epitopes according to any one of claims 1 to 8 is implemented.

Images