Prediction Method and System for Affinity between Antibody Mutant and Antigen

Abstract

The present invention discloses a method and system for predicting the affinity between an antibody mutant and an antigen, comprising the following steps: (1) collecting experimental data on the affinity of an antibody and the interaction between the antibody and the antigen to construct a fine-tuning data set; (2) training the UniRep model using the fine-tuning data set to obtain an affinity prediction model for the antibody and the antigen; (3) mutating key sites in the antibody sequence to obtain an antibody mutant; and using the affinity prediction model to predict the affinity between the antibody mutant and the antigen. The method and system of the present invention can accurately evaluate the change in the antigen affinity after the antibody mutates at different sites, reduce the experimental screening cost, and improve the antibody design efficiency.

Description

Technical Field

[0001] The present invention relates to the technical fields of bioinformatics and protein engineering, and particularly to a method and system for predicting the affinity between an antibody mutant and an antigen with the aid of a protein language large model. Background Art

[0002] The affinity between an antibody and an antigen is a key factor determining the effect of an immune response. Studying the impact of antibody mutations on their affinity for a specific antigen is crucial for antibody drug development and vaccine design. However, due to the complexity of antibody-antigen interactions and the vastness of the antibody mutation space, the cost of experimentally screening antibody mutants is high and the efficiency is low. Therefore, it is particularly important to develop a computational model that can effectively predict the change in the affinity between an antibody and an antigen after mutation.

[0003] Currently, multiple models have been used to predict the affinity between an antibody and an antigen. For example, the DeepAb model extracts antibody sequence features through a convolutional neural network (CNN) and combines the structural information of the antibody and the antigen for affinity prediction; the TITAN model uses a graph neural network (GNN) and combines the three-dimensional structural data of the antibody and antigen interfaces to model their interactions. However, these models have some limitations in application. In particular, most of them rely on the three-dimensional structural data of the antibody or antigen. This results in great limitations in the application of these models in the absence of structural data. In addition, when dealing with complex multi-site mutations, the prediction accuracy of these models often decreases, and their computational cost is high, making it difficult to meet the needs of large-scale mutation screening. Related prior arts include: for example, Chinese patent document with publication number CN 117334247 A discloses a method for training an antigen-antibody affinity prediction model and an antibody screening method, which relates to artificial intelligence fields such as natural language processing and deep learning, including obtaining a target complex three-dimensional feature map extraction model and a first model parameter set of the target complex three-dimensional feature map extraction model; performing parameter assignment on the initial feature map extraction layer based on the first model parameter set to obtain a candidate affinity prediction model to be trained; obtaining a first training sample of the candidate affinity prediction model including a first complex sample and a set of mutant complex samples of the first complex sample; inputting the first complex sample and the set of mutant complex samples into the candidate affinity prediction model for model training until the training is completed to obtain a trained target affinity prediction model.

[0004] To solve these problems, protein language models have emerged. The development of such models began in 2019. Inspired by language models in natural language processing (such as Word2Vec and BERT), protein sequences are first regarded as "languages", amino acids as "words", and sequence structure and function information are extracted from a large number of protein sequences through deep learning. Representative protein language models include UniRep, ESM (Evolutionary Scale Modeling), and ProtTrans, etc. Through large-scale pre-training, they can capture complex patterns in protein sequences and apply them to prediction tasks such as folding, function, and activity.

[0005] A significant advantage of protein language models is that they can extract useful features from protein sequences without relying on three-dimensional structure data. This enables efficient and accurate affinity prediction even in the absence of structural data. In addition, such models show strong representation capabilities when dealing with complex sequence changes such as multi-site mutations and can quickly process large-scale mutation screening tasks. Through these characteristics, protein language models overcome the limitations of existing models and significantly improve computational efficiency and prediction accuracy. Summary of the Invention

[0006] The present invention provides a method and system for predicting the affinity between an antibody mutant and an antigen, which can accurately evaluate the change in the antigen affinity after the antibody mutates at different sites, reduce the experimental screening cost, and improve the antibody design efficiency.

[0007] The technical solution of the present invention is as follows:

[0008] A method for predicting the affinity between an antibody mutant and an antigen, comprising the following steps:

[0009] (1) Collect the affinity experimental data of the antibody and the interaction between the antibody and the antigen, and construct a fine-tuning data set;

[0010] (2) Use the fine-tuning data set to train the UniRep model to obtain an affinity prediction model for the antibody and the antigen;

[0011] (3) Mutate the key sites in the antibody sequence to obtain an antibody mutant; use the affinity prediction model to predict the affinity between the antibody mutant and the antigen.

[0012] The said fine-tuning data set includes the amino acid sequence of the antibody and the affinity measurement value of the interaction between the antibody and the antigen.

[0013] Each antibody sequence is represented by its amino acid sequence, and the amino acid sequence is represented by a single-letter symbol (such as A, C, D, E, etc.). Each amino acid in the antibody sequence is used as the input of the UniRep model.

[0014] Step (2) includes:

[0015] (2-1) Divide the fine-tuning data set into a training set and a test set;

[0016] The training set is used for supervised learning of the UniRep model. By minimizing the error between the predicted affinity value and the true affinity measurement value, the parameters of the UniRep model are gradually optimized; after training, an affinity prediction model is obtained;

[0017] The test set is used to evaluate the prediction ability of the model on unseen data.

[0018] In step (2-1), the fine-tuning data set is divided by the strategy of K-fold cross-validation. The fine-tuning data set is divided into K subsets of equal size. Each time, one subset is used as the test set, and the remaining K-1 subsets are used as the training set.

[0019] Step (2-2) includes:

[0020] (2-2i) Input the amino acid sequence of antibody X i into the UniRep model. The UniRep model extracts the sequence features of antibody X i and predicts the affinity of antibody X i interacting with the antigen; the formula is:

[0021] ;

[0022] where is the predicted affinity value of the i-th antibody X i ; UniRep(X) represents extracting sequence features from the amino acid sequence of antibody X i by the UniRep model; f is a regression function representing the mapping relationship from sequence features to affinity; ϵ is the prediction error term;

[0023] (2-2ii) Use the mean square error (MSE) as the loss function to calculate the error between the predicted affinity value and the measured affinity value:

[0024] ;

[0025] where Y i is the true affinity value measured experimentally, is the predicted value of the model, and N is the number of samples in the test set;

[0026] In the process of model training, calculate the gradient of the loss function with respect to the model parameters by the backpropagation algorithm, and update the model parameters using the gradient descent algorithm; the specific optimization formula is as follows:

[0027] ;

[0028] Among them, θ t is the current model parameter; η is the learning rate; is the gradient of the loss function with respect to the model parameter;

[0029] (2 - 2iv) By continuous iteration and optimization, an affinity prediction model is obtained.

[0030] Using MSE as the loss function can effectively measure the difference between the model prediction value and the true experimental data. By minimizing MSE, the model can adjust the parameters to improve the accuracy of affinity prediction.

[0031] In step (2 - 2i), the UniRep model extracts the sequence features of antibody X i including: converting the amino acid sequence of antibody X i into an amino acid feature vector through the embedding layer;

[0032] Processing the amino acid features at each time step through the LSTM network to generate context - dependent features of the amino acid sequence; expressed by the formula:

[0033] ;

[0034] where h t represents the hidden state of the LSTM network at time step t; x t represents the input feature of the amino acid sequence of the antibody at time step t; h t-1 represents the hidden state of the previous time step t - 1;

[0035] The hidden state of the last time step is the context - dependent feature of the amino acid sequence, which is the sequence feature of antibody X i ;

[0036] In step (3), mutating at the key sites of the antibody sequence includes: antibody mutants obtained by single - point mutation or multi - point mutation in the complementarity - determining regions of the antibody sequence.

[0037] In step (3), predicting the affinity between the antibody mutant and the antigen using the affinity prediction model includes: inputting the amino acid sequence of the antibody mutant into the affinity prediction model, and the affinity prediction model extracts the sequence features of the antibody mutant and predicts the affinity of the interaction between the antibody mutant and the antigen according to the sequence features. The formula is:

[0038] ;

[0039] where, is the predicted affinity value of the antibody mutant X by the affinity prediction model mut ; UniRep(X mut ) represents the sequence features of the antibody mutant extracted by the affinity prediction model; f is a regression function that directly maps the extracted features to the affinity value.

[0040] The antibody described is an anti-HER2 antibody; the antigen described is the HER2 antigen.

[0041] The present invention also provides a system for predicting the affinity between an antibody mutant and an antigen, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The memory also stores the affinity prediction model. When the processor executes the computer program, the following steps are implemented:

[0042] Input the amino acid sequence of the antibody mutant into the affinity prediction model, and output the predicted value of the affinity between the antibody mutant and the antigen.

[0043] Compared with the prior art, the beneficial effects of the present invention are:

[0044] Based on the protein language model UniRep model, the present invention proposes an innovative method and system for predicting the affinity between an antibody mutant and an antigen. Different from the traditional model that relies on three-dimensional structure information, the present invention only relies on the sequence data of the antibody and the antigen, can flexibly handle complex mutation situations, and adapts to the prediction task of a specific antigen (such as HER2) through fine-tuning after large-scale training.

[0045] The present invention not only improves the prediction accuracy of the affinity of the antibody mutant, but also can significantly accelerate the large-scale mutation screening task, reduce the experimental cost and time in antibody drug development, and has broad application prospects and technical advantages. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic diagram of the overall framework principle of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0047] The following details the embodiments of the present invention, where the same or similar reference numerals represent the same or similar elements or elements with similar functions throughout. The embodiments described below with reference to the drawings are exemplary and are only used to explain the present invention and cannot be used to limit the present invention.

[0048] Those skilled in the art can understand that, unless otherwise defined, all terms used herein (including technical terms and scientific terms) have the same meaning as the general understanding of those of ordinary skill in the art to which the present invention belongs. It should also be understood that terms defined in common dictionaries should be understood to have a meaning consistent with the context of the prior art, and will not be interpreted in an idealized or overly formal sense unless defined as herein.

[0049] The reference numerals of the steps mentioned in each embodiment are only for convenience of description and have no substantial chronological connection. Different steps in each specific embodiment can be combined in different chronological orders to achieve the object of the present invention.

[0050] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0051] Example:

[0052] As Figure 1 shown, a method for predicting the affinity between an antibody mutant and an antigen by means of a protein language large model specifically includes the following steps:

[0053] 101) Steps for constructing an antibody affinity prediction model

[0054] First, use the long short-term memory network (LSTM) in the UniRep model to extract deep features of the antibody sequence and construct an affinity prediction model between the antibody and the antigen. UniRep captures long-range dependence information by processing each amino acid site in the antibody sequence, thereby obtaining key features in the interaction between the antibody sequence and the antigen. The antibody sequence data is the input raw data, and the LSTM network is used to learn the characteristics of the antibody sequence, and gradually generate local and global feature representations of the sequence through multiple layers of recursion. These features can reflect the structure of the antibody and the potential influence of its functional regions, and thus lay a foundation for antibody affinity prediction.

[0055] Furthermore, the processing flow of the LSTM network is as follows:

[0056] Input the antibody sequence: Each antibody sequence is represented by its amino acid sequence, and the amino acid sequence is represented by a single-letter symbol (such as A, C, D, E, etc.). Each amino acid in the antibody sequence is used as an input. The UniRep model can process variable-length protein sequences, so there is no need to specifically crop or pad the sequence length, and then it is converted into a feature vector of a specific dimension through an embedding layer. Each amino acid sequence is represented as a continuous vector after embedding, and these vectors capture the semantic relationships between amino acids.

[0057] LSTM feature processing: The LSTM network processes the amino acid features of each time step in the sequence to generate a hidden state for each time step. The hidden state is the core of the LSTM's ability to capture contextual information. By combining information from previous time steps (hidden states), the current state is gradually updated. The hidden state not only reflects the information of the current amino acid but also incorporates the influence of previous amino acids, thereby capturing long-range dependencies within the antibody sequence.

[0058] Global and local feature output: The LSTM ultimately outputs a hidden state representation of the sequence. These hidden states contain both local feature representations (i.e., sequence information at the current time step) and global feature representations (i.e., contextual dependencies across the entire sequence). The UniRep model uses the final hidden state to represent global features and the average hidden state to reflect the overall characteristics of the sequence. These features are used in the subsequent affinity prediction model.

[0059] ;

[0060] Among them, h t represents the hidden state of LSTM at time step t, x t represents the input features of the antibody sequence at time step t, h t-1 Represents the hidden state at the previous moment.

[0061] Each hidden state h t The calculation will combine the input feature x of the current amino acid t and the hidden state h at the previous moment t-1 , which recursively processes the antibody sequence to generate deep features representing the functional regions of the antibody.

[0062] 102) Dataset preparation and fine-tuning

[0063] First, a dataset containing experimental data on the affinity of anti-HER2 antibodies for the HER2 antigen was selected. This dataset includes antibody sequences and their affinity measurements for the HER2 antigen. To ensure the model can effectively generalize across a wide range of conditions, the dataset was divided into training and test sets. The training set was used for supervised learning of the UniRep model, enabling it to learn key features from the antibody sequences, while the test set was used to evaluate the model's predictive ability on unseen data.

[0064] Specifically, the training set is used to minimize the error between the predicted affinity values and the true experimental values. By optimizing the model parameters, it effectively captures the structural characteristics of the antibody sequence and its binding relationship with the HER2 antigen. During this process, the UniRep model utilizes its embedding layer and LSTM network to represent each amino acid in the antibody sequence, thereby generating a deep feature representation of the sequence.

[0065] During the model fine-tuning process, the UniRep model first encodes each amino acid in the antibody sequence into a high-dimensional feature vector through an embedding layer, and then uses an LSTM network to process these feature vectors to generate context-dependent features of the sequence. Finally, these features are mapped to affinity prediction values through a fully connected layer. The prediction formula is:

[0066] ;

[0067] where is the predicted affinity value of the i-th antibody sequence X i ; UniRep(X) represents the extraction of sequence features by the UniRep model; f is a regression function representing the mapping relationship from features to affinity; ϵ is the prediction error term.

[0068] At this stage, the core task of the model is to minimize the gap between the predicted value and the real experimental data by optimizing the model parameters.

[0069] Loss function and optimization: To optimize the parameters of the UniRep model, the mean squared error (MSE) is used as the loss function. MSE can quantify the difference between the model's predicted value and the real value, helping the model gradually adjust the parameters to improve the prediction accuracy. The specific formula of the loss function is:

[0070] ;

[0071] where, Y i is the real affinity value measured experimentally, is the predicted value of the model, and N is the number of samples. MSE helps the model continuously adjust its internal weights by measuring the error between the predicted value and the real value, thereby improving the prediction performance. During the model training process, the gradient of the loss function with respect to the model parameters is calculated through the backpropagation algorithm, and the model parameters are updated using the gradient descent algorithm. The specific optimization formula is as follows:

[0072] ;

[0073] where θ t are the current model parameters; η is the learning rate, which is used to control the step size of parameter update; is the gradient of the loss function with respect to the model parameters. Through continuous iteration and optimization, the model can learn the deep features related to affinity in the antibody sequence and gradually improve the performance of the model in the affinity prediction task.

[0074] Dataset Partitioning and Cross-Validation. To ensure the generalization ability of the model, the K-fold cross-validation strategy is adopted during dataset partitioning. K-fold cross-validation divides the dataset into K subsets of equal size. Each time, one subset is used as the test set, and the remaining K - 1 subsets are used as the training set. Through cross-validation, the model can be trained and tested under different data partitions, further verifying its robustness and generalization ability on different data.

[0075] 103) Mutant Generation and Affinity Prediction

[0076] After the model is fine-tuned and trained, the next step is to generate antibody mutants and predict their affinities. Mutation designs are carried out at key sites (such as complementarity-determining regions, CDRs) of the antibody sequence. The CDR region is crucial for the binding of the antibody to the antigen, and mutations may have a significant impact on the binding ability of the antibody. Therefore, mutation designs usually choose to be carried out at key sites in the CDR region, and these sites may simulate different antibody variants through single-point mutations or multi-point mutations.

[0077] After the mutation design is completed, the mutant sequences are input into the fine-tuned UniRep model. The model predicts the affinity changes between them and the HER2 antigen according to the characteristics of the mutant sequences.

[0078] ;

[0079] Among them, is the affinity prediction value of the fine-tuned UniRep model for the mutant antibody sequence X mut , UniRep(X mut ) represents the feature representation of the mutant sequence extracted by the UniRep model, and f is a regression function that directly maps the extracted features to affinity values.

[0080] In this step, the model no longer performs further error calculations or model optimizations. The fine-tuned UniRep model directly outputs the affinity prediction values of the mutants. This prediction process is very efficient and can handle large-scale mutant data, helping researchers quickly screen out excellent antibody mutants.

[0081] Mutant Screening and Application. By predicting the affinity changes of different mutants with the HER2 antigen, antibody mutants with high affinity can be quickly screened out. The screened mutants can be used for further experimental verification or applied in the early stage of antibody drug development. Through the screening and optimization of mutants, researchers can more efficiently design antibody drugs with better therapeutic potential.

[0082] In summary, step 103) generates the affinity prediction results of antibody mutants based on the fine-tuned UniRep model, and obtains high-affinity antibody mutants through rapid screening, providing important technical support for the development and optimization of antibody drugs.

[0083] Although the above methods are illustrated and described as a series of actions for simplicity of explanation, it should be understood and appreciated that these methods are not limited by the order of the actions, because according to one or more embodiments, some actions may occur in a different order and / or concurrently with other actions not illustrated and described herein but understood by those skilled in the art.

[0084] The present invention also provides a system for predicting the affinity between an antibody mutant and an antigen, including a memory, a processor, and a computer program stored in the memory and executable on the processor. The memory also stores the affinity prediction model. When the processor executes the computer program, the following steps are implemented:

[0085] Input the amino acid sequence of the antibody mutant into the affinity prediction model, and output the predicted value of the affinity between the antibody mutant and the antigen.

[0086] Those skilled in the art will further appreciate that the various illustrative logical blocks, modules, circuits, and algorithm steps described in connection with the embodiments disclosed herein can be implemented as electronic hardware, computer software, or a combination of the two. To clearly illustrate this interchangeability of hardware and software, the various illustrative components, blocks, modules, circuits, and steps are described above in terms of their functionality. Whether such functionality is implemented as hardware or software depends on the particular application and the design constraints imposed on the overall system. Skilled artisans may implement the described functionality in different ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.

[0087] The various illustrative logical blocks, modules, and circuits described in connection with the embodiments disclosed herein can be implemented using a general-purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic device, discrete gate or transistor logic, discrete hardware components, or any combination thereof designed to perform the functions described herein. The general-purpose processor may be a microprocessor, but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. The processor may also be implemented as a combination of computing devices, such as a combination of a DSP and a microprocessor, multiple microprocessors, one or more microprocessors cooperating with a DSP core, or any other such configuration.

[0088] The steps of the methods or algorithms described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. The software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, a hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art. An exemplary storage medium is coupled to the processor such that the processor can read from, and write to, the storage medium. In an alternative, the storage medium may be integrated into the processor. The processor and the storage medium may reside in an ASIC. The ASIC may reside in a user terminal. In an alternative, the processor and the storage medium may reside as discrete components in a user terminal.

[0089] In one or more exemplary embodiments, the described functionality may be implemented in hardware, software, firmware, or any combination thereof. If implemented in software as a computer program product, the functions may be stored on or transmitted via a computer-readable medium as one or more instructions or code. The computer-readable medium includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. A storage medium may be any available medium that can be accessed by a computer. By way of example and not limitation, such computer-readable medium can comprise RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer. Any connection is properly termed a computer-readable medium. For example, if the software is transmitted from a Web site, server, or other remote source using a coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave, then the coaxial cable, fiber optic cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwave are included in the definition of medium. As used herein, disk and disc include compact disc (CD), laser disc, optical disc, digital versatile disc (DVD), floppy disk, and Blu-ray disc where disks usually reproduce data magnetically, while discs reproduce data optically with lasers. Combinations of the above should also be included within the scope of computer-readable medium.

[0090] The above-described embodiments have elaborated on the technical solutions and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the present invention. Any modifications, supplements, equivalent replacements, etc. made within the scope of the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting the affinity between an antibody mutant and an antigen, characterized in that Comprising the following steps: (1) Collect experimental data on the affinity of the antibody-antigen interaction, and construct a fine-tuning data set; the fine-tuning data set includes the amino acid sequence of the antibody and the measured affinity value of the antibody-antigen interaction; the antibody is an anti-HER2 antibody; the antigen is a HER2 antigen; (2) Use the fine-tuning data set to train the UniRep model to obtain an affinity prediction model for the antibody-antigen, including: (2-1) Divide the fine-tuning data set into a training set and a test set, including: adopting a K-fold cross-validation strategy to divide the fine-tuning data set, dividing the fine-tuning data set into K subsets of equal size, and using one subset as the test set each time, and the remaining K-1 subsets as the training set; (2-2) The training set is used for supervised learning of the UniRep model. By minimizing the error between the predicted affinity value and the true measured affinity value, the parameters of the UniRep model are gradually optimized; after training, an affinity prediction model is obtained, including: (2 - 2i) inputs the amino acid sequence of antibody X i into the UniRep model, and the UniRep model extracts the sequence features of antibody X i and predicts the affinity of antibody X i for interacting with the antigen; the formula is: ； wherein, is the predicted affinity value of the i-th antibody X i ; UniRep(X) represents the extraction of sequence features from the amino acid sequence of antibody X i ; f is a regression function representing the mapping relationship from sequence features to affinity; ϵ is the prediction error term; (2-2ii) Use the mean square error as the loss function to calculate the error between the predicted affinity value and the measured affinity value: ； Among them, Y i is the true affinity value measured experimentally, and N is the number of samples in the test set; (2-2iii) During the model training process, calculate the gradient of the loss function with respect to the model parameters through the backpropagation algorithm, and use the gradient descent algorithm to update the model parameters; the specific optimization formula is as follows: ； Among them, θ t is the current model parameter; η is the learning rate; is the gradient of the loss function with respect to the model parameter; (2-2iv) Through continuous iteration and optimization, an affinity prediction model is obtained; (2-3) The test set is used to evaluate the prediction ability of the model on unseen data; (3) Mutate the key sites in the antibody sequence to obtain antibody mutants; use the affinity prediction model to predict the affinity between the antibody mutants and the antigen; Mutating the key sites in the antibody sequence includes: antibody mutants obtained by single-point mutation or multi-point mutation in the complementarity-determining region of the antibody sequence.

2. The method for predicting the affinity of the antibody mutant according to claim 1, wherein In step (2-2i), the UniRep model extracts the sequence features of antibody X i , including: converting the amino acid sequence of antibody X i into an amino acid feature vector through an embedding layer; Gradually process the amino acid features at each time step through the LSTM network to generate context-dependent features of the amino acid sequence; the formula is expressed as: ； Among them, h t represents the hidden state of the LSTM network at time step t; x t represents the input feature of the amino acid sequence of the antibody at time step t; h t-1 represents the hidden state of the previous time step t - 1; The hidden state at the last time step is the context-dependent feature of the amino acid sequence, which is the antibody X i sequence feature.

3. The method for predicting the affinity between the antibody mutant and the antigen according to claim 1, wherein In step (3), using the affinity prediction model to predict the affinity between the antibody mutant and the antigen includes: inputting the amino acid sequence of the antibody mutant into the affinity prediction model, and the affinity prediction model extracts the sequence features of the antibody mutant and predicts the affinity of the antibody mutant-antigen interaction according to the sequence features, and the formula is: ； Among them, is the predicted affinity value of the antibody mutant X by the affinity prediction model; mut UniRep(X mut ) represents the sequence features of the antibody mutant extracted by the affinity prediction model; f is a regression function that directly maps the extracted features to affinity values.

4. A prediction system for the affinity between an antibody mutant and an antigen, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, The memory also stores an affinity prediction model obtained by the prediction method according to any one of claims 1-3. When the processor executes the computer program, the following steps are implemented: Input the amino acid sequence of the antibody mutant into the affinity prediction model, and output the predicted value of the affinity between the antibody mutant and the antigen.

Images