A method, device and program product for calculating neoantigen load

Abstract

The application relates to the field of intelligent medical treatment, in particular to a new antigen load calculation method, equipment and program product. The method comprises the following steps: S1, obtaining sequencing data of a to-be-tested person, wherein the sequencing data is exome sequencing data or panel sequencing data; S2, extracting sequence and structural features of an antigen peptide, sequence and structural features of an MHC binding groove and sequence of a TCR variable region based on the sequencing data; S3, calculating MHC-antigen peptide affinity based on the sequence and structural features of the MHC binding groove and the sequence and structural features of the antigen peptide; S4, calculating the probability of TCR recognizing MHC-antigen peptide based on the sequence of the TCR variable region; and S5, calculating the new antigen load of the to-be-tested person based on the probability of TCR recognizing MHC-antigen peptide and the MHC-antigen peptide affinity. The method can calculate the new antigen load and has good clinical value.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of intelligent medical treatment, in particular to a new antigen load calculation method, device, program product and computer readable storage medium. BACKGROUND

[0002] Anti-PD-L1 (Programmed Death-Ligand 1) therapy is an immune checkpoint inhibitor therapy designed to enhance the body's own immune system's ability to recognize and attack tumor cells. This treatment method relieves the inhibition of the immune system by tumor cells by blocking the interaction between PD-L1 and PD-1, thereby activating T cell-mediated anti-tumor immune response. In esophageal cancer Anti-PD-L1 therapy, there is a lack of reliable biomarkers, and in the clinical application of immune checkpoint inhibitors CTLA4 and PD-L1 mAb, it is very important to identify patients by detecting biomarkers. The commonly used companion diagnostic markers currently include PD-L1 expression level, tumor mutation burden (TMB), microsatellite instability (MSI-H), and DNA mismatch repair deficiency (dMMR), etc. Although these prognostic markers play a guiding role in clinical application, more and more clinical cohort studies have shown that these markers cannot accurately distinguish which patients can benefit from immune checkpoint inhibitor therapy, therefore, the concept of tumor neoantigen burden (TNB) is proposed, which is the number of T cells actually targeting mutations, which can better judge the clinical efficacy of immune checkpoint inhibitors. Studies have shown that combining neoantigen quality with tumor CD8+ T cell penetration can accurately identify malignant glioma patients with the longest survival, while other studies have proposed a neoantigen screening algorithm that can effectively distinguish between long and short survival pancreatic cancer patients, but models that rely solely on the number of antigens cannot do this. SUMMARY

[0003] To solve the above problems, the present application provides a new antigen load calculation method, which specifically comprises:

[0004] S1, obtaining sequencing data of a to-be-tested person, the sequencing data being exome sequencing data or panel sequencing data;

[0005] S2, extracting sequence and structural features of antigen peptides, sequence and structural features of MHC binding grooves, and sequence of TCR variable regions based on the sequencing data;

[0006] S3, calculating MHC-antigen peptide affinity based on the sequence and structural features of the MHC binding groove and the sequence and structural features of the antigen peptide;

[0007] S4, calculating the probability of TCR recognizing MHC-antigen peptide based on the sequence of the TCR variable region;

[0008] S5, calculating the neoantigen load of the subject based on the probability of the TCR recognizing the MHC-antigen peptide and the MHC-antigen peptide affinity.

[0009] The extraction is performed by the trained first deep network model to obtain sequence and structural features of the antigen peptide, sequence and structural features of the MHC binding groove.

[0010] Optionally, the training process of the first deep network model is as follows:

[0011] Obtaining the sequence of the antigen peptide and the sequence of the MHC binding groove from a protein database;

[0012] Extracting structural features from the sequence to obtain the structural features;

[0013] Constructing a feature training set based on the sequence and the structural features;

[0014] Inputting the feature training set into the first deep network model for training to obtain the trained first deep network model;

[0015] Optionally, the structural features include one or more of the following: structural features, structural neighbor features, and volume accessible surface area.

[0016] Optionally, the feature extraction method uses one or more of the following: physical and chemical extraction, local structure entropy extraction, pairing potential energy extraction, and interaction tendency extraction.

[0017] Optionally, the first deep network model includes one or more of the following: GBRT, XGBoost, LightGBM, CatBoost, Random Forest, and AdaBoost.

[0018] S2 is replaced by: inputting the sequencing data into the trained second deep network model to obtain the MHC-antigen peptide affinity, and the training process of the second deep network model is as follows:

[0019] First step, obtaining MHC antigen peptide mass spectrum data set and IEDB data;

[0020] Second step, pre-training the neural network model by the IEDB data to obtain the weight of the training model;

[0021] Third step, migrating the weight of the training model to the second deep network model and inputting the MHC antigen peptide mass spectrum data set into the second deep network model for training to obtain the trained second deep network model;

[0022] Optionally, the first step further comprises MHC typing, obtaining antigen peptide data, MHC molecules, typing the MHC molecules to obtain MHC-I class data, constructing an MHC-I class antigen peptide mass spectrum dataset of MHC-I class data and antigen peptides based on mass spectrometry; inputting the MHC-I class antigen peptide mass spectrum dataset into the training model for training to obtain a second deep network model;

[0023] Optionally, the third step further comprises antigen deep features, obtaining antigen gene expression level data, extracting features of the antigen gene expression level data to obtain antigen deep features, and inputting the antigen deep features and the MHC-I class antigen peptide mass spectrum data into the second deep network model for training to obtain the trained second deep network model.

[0024] Optionally, the S2 is replaced by: inputting the sequencing data into the trained multi-source heterogeneous network model to obtain MHC-antigen peptide affinity, and the training process of the multi-source heterogeneous network model is:

[0025] Obtaining the structure of MHC molecules and the structure of antigen peptides to generate MHC network structure and antigen peptide network structure; the MHC molecules include MHC-I subtypes and MHC-II subtypes;

[0026] Calculating the similarity network of MHC-I subtypes and MHC-II subtypes, and the similarity network of antigen peptides;

[0027] Obtaining the association relationship between MHC molecules and antigen peptides;

[0028] Based on the similarity network of MHC-I subtypes and MHC-II subtypes, the similarity network of antigen peptides, and the association relationship between MHC molecules and antigen peptides, inputting into a third deep network model to generate a multi-source heterogeneous network;

[0029] Optionally, the multi-source heterogeneous network model uses one or more of the following: two-pass heterogeneous network, heterogeneous information network, graph neural network, and hypergraph;

[0030] Optionally, the S2 further comprises executing the second deep network model in parallel to obtain MHC-antigen peptide affinity, performing weight distribution calculation on the MHC-antigen peptide affinity obtained by the first deep network model and the second deep network model to obtain a second MHC-antigen peptide affinity, and calculating the neoantigen load of the subject based on the probability of TCR recognizing MHC-antigen peptide and the second MHC-antigen peptide affinity;

[0031] Optionally, the S2 further comprises obtaining MHC-antigen peptide affinities by parallel execution of the multi-source heterogeneous network model, performing weight distribution calculation on the MHC-antigen peptide affinities obtained by the first deep network model and the multi-source heterogeneous network model to obtain a third MHC-antigen peptide affinity, and calculating the neoantigen load of the subject based on the probability of the TCR recognizing the MHC-antigen peptide and the third MHC-antigen peptide affinity.

[0032] Optionally, the S2 further comprises obtaining MHC-antigen peptide affinities by parallel execution of the second deep network model and the multi-source heterogeneous network model, performing weight distribution calculation on the MHC-antigen peptide affinities obtained by the first deep network model, the multi-source heterogeneous network model, and the second deep network model to obtain a fourth MHC-antigen peptide affinity, and calculating the neoantigen load of the subject based on the probability of the TCR recognizing the MHC-antigen peptide and the fourth MHC-antigen peptide affinity.

[0033] The extraction of the sequence of the TCR variable region is obtained from the genomic data of the subject;

[0034] Optionally, the probability of the TCR recognizing the MHC-antigen peptide is calculated by an identification model, and the construction process of the identification model is as follows:

[0035] Obtaining antigen epitope-TCR interaction data and sequences and structures of TCR and peptide chains;

[0036] Extracting TCR features based on the sequence and structure data of the TCR and peptide chains;

[0037] Inputting the TCR features and antigen epitope-TCR interaction data into a fourth deep network model for training to obtain an identification model;

[0038] Optionally, the TCR features include one or more of the following: charge, hydrophobicity, and two-dimensional structure features of CDR3.

[0039] Optionally, the extraction of the TCR features is obtained by group sparse regularization regression analysis.

[0040] Optionally, the probability calculation of the TCR recognition further comprises TCR-antigen epitope binding prediction, and the probability of the TCR recognition is obtained by the identification model and TCR-antigen epitope binding prediction; the TCR-antigen epitope binding prediction is obtained by prediction by a trained TCR-antigen epitope binding prediction model.

[0041] Optionally, the construction process of the trained TCR-antigen epitope binding prediction model is as follows:

[0042] Obtaining sequences of CDR3 regions, genetic mutation data, and antigen peptides;

[0043] The sequence, gene mutation data, and antigenic peptides of the CDR3 region are converted into a feature matrix.

[0044] The feature matrix is ​​input into a neural network for training to obtain a TCR-antigen epitope binding prediction model;

[0045] Optionally, the feature matrix of the antigenic peptide includes both antigenic peptide sequence features and structural features;

[0046] Optionally, the TCR-antigen epitope binding prediction model further includes MHC-antigen peptide affinity calculation, which uses the characteristics of antigen peptide sequence and structure to predict affinity results.

[0047] The probability calculation for TCR recognition also includes the prediction of the immunogenicity of tumor antigens. The probability of TCR recognition is obtained through the recognition model and the prediction of the immunogenicity of tumor antigens. The prediction of the immunogenicity of tumor antigens is obtained by a trained tumor antigen immunogenicity prediction model.

[0048] Optionally, the process of constructing the immunogenicity prediction model for the tumor antigen is as follows:

[0049] Obtain RNA-seq data, tumor cell genome data, and immunogenicity data from different tumor-infiltrating lymphocytes;

[0050] TCR library profiling analysis was performed on the RNA-seq data to obtain tumor gene expression profile data, and the CDR3 sequence was obtained from the RNA-seq data;

[0051] Gene mutation and antigenic peptide data were obtained based on the tumor cell genomic data.

[0052] Feature matrix data was obtained by feature extraction from CDR3 sequence and antigenic peptide data;

[0053] The tumor gene expression profile data, gene mutation data, and feature matrix data are input into a neural network for training to obtain an immunogenicity prediction model.

[0054] The formula for calculating the neoantigen load is as follows:

[0055]

[0056] Where N represents the neoantigen load, and n represents the number of tumor clones obtained from the patient's exome sequencing data. b i It indicates affinity. r i c represents the probability of TCR recognition. j This indicates the expression level of gene j.

[0057] The present application aims to provide a tumor treatment prognosis prediction method based on neoantigen load, comprising:

[0058] Obtaining tumor section data and serum data of a subject to be tested;

[0059] Performing neoantigen screening based on the tumor section data and serum data to obtain neoantigen;

[0060] Calculating the load of the neoantigen based on the above neoantigen load calculation method to obtain neoantigen load;

[0061] Performing prognosis prediction based on the neoantigen load to obtain a prediction result.

[0062] The present application aims to provide a computer program product comprising a computer program or instructions thereon, which are executed by a processor to implement the above-mentioned neoantigen load calculation method or tumor treatment prognosis prediction method based on neoantigen load.

[0063] The present application aims to provide a computer device comprising a memory, a processor, and a computer program or instructions stored on the memory, which are executed by the processor to implement the above-mentioned neoantigen load calculation method or tumor treatment prognosis prediction method based on neoantigen load.

[0064] The present application aims to provide a computer-readable storage medium having a computer program or instructions stored thereon, which are executed by a processor to implement the above-mentioned neoantigen load calculation method or tumor treatment prognosis prediction method based on neoantigen load.

[0065] Advantages of the present application:

[0066] 1. A neoantigen load calculation method is proposed, which is calculated by the affinity of MHC antigen peptides and the probability of TCR recognition. The calculated neoantigen load is used as a biomarker to assist cancer diagnosis and improve the accuracy of cancer diagnosis, which is helpful for cancer treatment.

[0067] 2. The affinity of MHC antigen peptides is calculated by mass spectrometry data of MHC class I binding antigen peptides, sequence characteristics and structural characteristics of antigen peptides and MHC binding groove, and topological structure and ontology characteristics of MHC antigen peptide isomerization network. The affinity is calculated by integrating various data and characteristics, which enhances the accuracy and reliability of the affinity and improves the accuracy and reliability of the neoantigen load.

[0068] 3. The probability calculation of TCR recognition takes the antigen and TCR related features as input to predict the probability of TCR recognizing and binding to tumor antigens; the gene mutation-TCR corresponding matrix is constructed by integrating different tumor gene mutations, and the probability of tumor specific antigen being recognized by TCR is analyzed from the perspective of tumor clone evolution; the immunogenicity of tumor antigens is predicted by taking the tumor gene expression profile and antigen peptide feature matrix as the input of the deep neural network, combining the tumor infiltrating lymphocyte (TIL) data and the immunogenicity data obtained by experiments, and then the TCR recognition probability is predicted, the multi-source data is integrated for TCR recognition, and the accurate TCR recognition probability can be obtained, which is beneficial to the calculation accuracy of new antigen load. BRIEF DESCRIPTION OF DRAWINGS

[0069] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0070] Figure 1 The flowchart of the new antigen load calculation method provided by the embodiment of the present application is shown.

[0071] Figure 2 The schematic diagram of the new antigen load calculation system provided by the embodiment of the present application is shown.

[0072] Figure 3 The schematic diagram of the new antigen load calculation device provided by the embodiment of the present application is shown.

[0073] Figure 4 The multi-omics data driven tumor new antigen screening route provided by the embodiment of the present application is shown.

[0074] Figure 5 The transfer learning training high-precision MHC-antigen peptide affinity prediction model provided by the embodiment of the present application is shown, (a) a convolutional neural network is trained using the data in IEDB to obtain a pan-cancer MHC-antigen peptide affinity prediction model; (b) a one-step convolutional neural network is adopted by using the transfer learning technology, combining the transcriptome and proteomic data to obtain a prediction model for a specific cancer. DETAILED DESCRIPTION

[0075] In order to enable those skilled in the art to better understand the present application scheme, the technical solutions in the embodiments of the present application will be described clearly and completely in conjunction with the drawings in the embodiments of the present application.

[0076] In some of the flowcharts described in the description, claims, and drawings of the present application and the above description, a plurality of operations are included in a specific order, but it should be clearly understood that these operations can be performed in the order they appear in this paper or in parallel, the serial numbers of the operations such as S101, S102, etc. are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these flowcharts can include more or fewer operations, and these operations can be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second" and the like in this paper are used to distinguish different messages, devices, modules, etc., and do not represent the order of precedence, nor do "first" and "second" represent different types.

[0077] Figure 1 The new antigen load calculation method provided by the embodiment of the application includes the following steps:

[0078] S101: Obtain the sequencing data of the testee, wherein the sequencing data is exome sequencing data or panel sequencing data;

[0079] In one specific embodiment, the present application takes tumor immune cycle as the research context, analyzes the factors involved in tumor neoantigen screening from endogenous mechanism, influencing variables and corresponding data sources, refines key issues and conducts research. By carefully analyzing the biological mechanism of these tumor antigens presented to activated T cells, the present application conducts research on MHC-I and antigen peptide affinity, T, and TCR recognition of antigen epitopes, because these are the key to determining whether the antigen is presented and whether it can activate T cells after being presented to the cell surface.

[0080] S102: Extract the sequence and structure features of the antigen peptide, the sequence and structure features of the MHC binding groove, and the sequence of the TCR variable region based on the sequencing data;

[0081] In one embodiment, the extraction is obtained by a trained first deep network model to obtain the sequence and structure features of the antigen peptide and the sequence and structure features of the MHC binding groove.

[0082] In one embodiment, the training process of the first deep network model is as follows:

[0083] Obtain the sequence of the antigen peptide and the sequence of the MHC binding groove from the protein database;

[0084] Extract the structure features from the sequence to obtain the structure features;

[0085] Construct a feature training set based on the sequence and structure features;

[0086] Input the feature training set into the first deep network model for training to obtain the trained first deep network model.

[0087] In one embodiment, the structural features include one or more of: structural features, structural neighbor features, volume accessible surface area.

[0088] In one embodiment, the feature extraction method includes one or more of: physicochemical extraction, local structural entropy extraction, pair potential energy extraction, interaction propensity extraction.

[0089] In one embodiment, the first deep network model includes one or more of: GBRT, XGBoost, LightGBM, CatBoost, Random Forest, AdaBoost.

[0090] In one embodiment, the S102 is replaced by: inputting the sequencing data into a trained second deep network model to obtain MHC-antigen peptide affinity, and the training process of the second deep network model is:

[0091] The first step is to obtain MHC antigen peptide mass spectrometry data set, IEDB data;

[0092] The second step is to obtain the weight of the training model by pre-training the neural network model through the IEDB data;

[0093] The third step is to migrate the weight of the training model to the second deep network model and input the MHC antigen peptide mass spectrometry data set into the second deep network model for training to obtain the trained second deep network model.

[0094] Optionally, the first step further includes MHC typing, obtaining antigen peptide data, MHC molecules, typing the MHC molecules to obtain MHC-I class data, and constructing MHC-I class antigen peptide mass spectrometry data set based on mass spectrometry. The MHC-I class antigen peptide mass spectrometry data set is input into the training model for training to obtain the second deep network model.

[0095] In one embodiment, the third step further includes antigen deep features, obtaining antigen gene expression level data, extracting features of the antigen gene expression level data to obtain antigen deep features, and inputting the antigen deep features and MHC-I class antigen peptide mass spectrometry data into the second deep network model for training to obtain the trained second deep network model.

[0096] In one embodiment, the S102 is replaced by: inputting the sequencing data into a trained multi-source heterogeneous network model to obtain MHC-antigen peptide affinity, and the training process of the multi-source heterogeneous network model is:

[0097] obtaining the structure of MHC molecules and the structure of antigen peptides to generate MHC network structure and antigen peptide network structure; the MHC molecules include MHC-I subtypes and MHC-II subtypes;

[0098] calculating the similarity network of MHC-I subtypes and MHC-II subtypes, the similarity network of antigen peptides;

[0099] obtaining the association relationship between MHC molecules and antigen peptides;

[0100] based on the similarity network of MHC-I subtypes and MHC-II subtypes, the similarity network of antigen peptides, and the association relationship between MHC molecules and antigen peptides, inputting into a third deep network model to generate a multi-source heterogeneous network.

[0101] In one embodiment, the multi-source heterogeneous network model adopts one or more of the following: two-pass heterogeneous network, heterogeneous information network, graph neural network, hypergraph.

[0102] In one embodiment, the S102 further includes performing a second deep network model in parallel to obtain MHC-antigen peptide affinity, performing weight distribution calculation on the MHC-antigen peptide affinity obtained by the first deep network model and the second deep network model to obtain second MHC-antigen peptide affinity, and calculating the neoantigen load of the subject based on the probability of the TCR recognizing MHC-antigen peptide and the second MHC-antigen peptide affinity.

[0103] In one embodiment, the S2 further includes performing a multi-source heterogeneous network model in parallel to obtain MHC-antigen peptide affinity, performing weight distribution calculation on the MHC-antigen peptide affinity obtained by the first deep network model and the multi-source heterogeneous network model to obtain third MHC-antigen peptide affinity, and calculating the neoantigen load of the subject based on the probability of the TCR recognizing MHC-antigen peptide and the third MHC-antigen peptide affinity.

[0104] In one embodiment, the S2 further includes performing a second deep network model and a multi-source heterogeneous network model in parallel to obtain MHC-antigen peptide affinity, performing weight distribution calculation on the MHC-antigen peptide affinity obtained by the first deep network model, the multi-source heterogeneous network model, and the second deep network model to obtain fourth MHC-antigen peptide affinity, and calculating the neoantigen load of the subject based on the probability of the TCR recognizing MHC-antigen peptide and the fourth MHC-antigen peptide affinity.

[0105] In one embodiment, the sequence of the TCR variable region is extracted from the genomic data of the subject.

[0106] S103: calculating MHC-antigen peptide affinity based on the sequence and structural characteristics of the MHC binding groove and the sequence and structural characteristics of the antigen peptide;

[0107] In one embodiment, the MHC-I molecule and antigen peptide affinity is calculated: integrating mass spectrometry data of MHC class I binding antigen peptides, constructing mass spectrometry-based MHC-antigen peptide benchmark dataset and developing online database; analyzing genomic or whole exome sequencing data for MHC typing, extracting sequence and structural features of antigen peptide and MHC binding groove, proposing and developing deep learning-based MHC-antigen peptide affinity prediction model; constructing MHC-antigen peptide heterogeneous network integrating multi-source data, combining network topological structure features and ontology features, training deep feedforward network to predict MHC-antigen peptide affinity.

[0108] S104: calculating the probability of TCR recognizing MHC-antigen peptide based on the sequence of TCR variable region;

[0109] In one embodiment, the probability of TCR recognizing MHC-antigen peptide is calculated by an identification model, and the identification model is constructed as follows:

[0110] obtaining antigen epitope-TCR interaction data, sequence and structure of TCR and peptide chain;

[0111] extracting TCR features based on the sequence and structure data of TCR and peptide chain;

[0112] inputting the TCR features and antigen epitope-TCR interaction data into a fourth deep network model for training to obtain an identification model.

[0113] In one embodiment, the TCR features include one or more of the following: charge, hydrophobicity, two-dimensional structure features of CDR3.

[0114] In one embodiment, the TCR features are extracted by group sparse regularization regression analysis.

[0115] In one embodiment, the probability calculation of TCR recognition further includes TCR-antigen epitope binding prediction, and the probability of TCR recognition is obtained by the identification model and TCR-antigen epitope binding prediction; the TCR-antigen epitope binding prediction is obtained by predicting by a trained TCR-antigen epitope binding prediction model;

[0116] Optionally, the construction process of the trained TCR-antigen epitope binding prediction model is as follows:

[0117] obtaining sequence, gene mutation data and antigen peptide of CDR3 region;

[0118] converting the sequence, gene mutation data and antigen peptide of CDR3 region into a feature matrix;

[0119] The feature matrix is input into a neural network for training to obtain a TCR-antigen epitope binding prediction model.

[0120] In one embodiment, the feature matrix of the antigen peptide includes antigen peptide sequence features and structural features.

[0121] In one embodiment, the TCR-antigen epitope binding prediction model further includes MHC-antigen peptide affinity calculation, and the affinity prediction result is obtained by antigen peptide sequence features and structural features.

[0122] In one embodiment, the probability calculation of TCR recognition further includes immunogenicity prediction of tumor antigens, and the probability of TCR recognition is obtained by the recognition model and the immunogenicity prediction of tumor antigens, which is obtained by predicting the trained immunogenicity prediction model of tumor antigens.

[0123] In one embodiment, the construction process of the immunogenicity prediction model of tumor antigens is as follows:

[0124] Obtain RNA-seq data of different tumor infiltrating lymphocytes, tumor cell genomes, and immunogenicity data;

[0125] Perform TCR repertoire analysis on the RNA-seq data to obtain tumor gene expression profile data and obtain CDR3 sequences from the RNA-seq data;

[0126] Obtain gene mutations and antigen peptide data based on the tumor cell genome data;

[0127] Perform feature extraction on the CDR3 sequences and antigen peptide data to obtain feature matrix data;

[0128] Input the tumor gene expression profile data, gene mutations, and feature matrix data into a neural network for training to obtain an immunogenicity prediction model.

[0129] S105: Obtain the neoantigen load of the subject based on the probability of TCR recognizing MHC-antigen peptide and MHC-antigen peptide affinity calculation.

[0130] In one embodiment, the calculation formula of the neoantigen load is as follows:

[0131]

[0132] Wherein, N is the neoantigen load, n represents the number of tumor clones analyzed from the exome sequencing data of the patient, b i represents the affinity, r iProbability of TCR recognition, c j Expression level of gene j.

[0133] In one embodiment, the tumor-specific antigen recognized by the TCR is screened: the sequence and structural features of the antigen peptide and the TCR variable region are extracted from the genomic data, combined with the experimentally verified TCR-antigen peptide binding data set, and regression analysis and correlation analysis are performed to find the key factors of TCR recognition of tumor antigens; a deep learning model is constructed to predict the probability of TCR recognition and binding of tumor antigens by taking the antigen and TCR related features as input; the gene mutation-TCR corresponding matrix is constructed by integrating the gene mutations of different tumors, and the probability of tumor-specific antigen being recognized by TCR is analyzed from the perspective of tumor clonal evolution; the immunogenicity of tumor antigens is predicted by taking the tumor gene expression profile and antigen peptide feature matrix as the input of the deep neural network, combined with the tumor infiltrating lymphocyte (TIL) data and the experimentally obtained immunogenicity data.

[0134] Definition and calculation of neoantigen load index: tumor antigen activated T cell response is a multi-step complex process, in which antigen presentation and TCR recognition of pMHC complex are the main steps affecting T recognition and killing of target cells, and any step failure will lead to failure of immunotherapy. For antigen i of tumor clone j, assuming that the affinity of MHC-I molecule to the antigen peptide is bi, and the probability of TCR recognizing the antigen is ri, the neoantigen load of the patient is defined as:

[0135]

[0136] Where n represents the number of tumor clones analyzed from the patient's exome sequencing data.

[0137] It should be noted that the present application will use multi-omics data including genome or exon, transcriptome and mass spectrometry data when developing bioinformatics models. Once the prediction model is trained, only the patient's exome sequencing or large panel sequencing data is required during use, which is conducive to reducing the cost of clinical companion diagnosis and promoting market application and promotion.

[0138] In one embodiment, the present application aims to develop tumor neoantigen screening driven by big data and deep learning, and to deeply explore the affinity of MHC molecules and antigen peptides, TCR recognition of antigen peptides, and PD-L1 blocker clinical response biomarkers based on the integration of multi-omics data such as genome, transcriptome and protein mass spectrum, and to extract high-dimensional heterogeneous features from the biological processes of tumor antigen generation, presentation and immune response, and to develop and use various intelligent algorithms including deep learning to screen high-quality neoantigens with immunogenicity, such as Figure 4 as shown.

[0139] In one specific embodiment, mass spectrometry data of MHC class I binding antigen peptides are integrated, mass spectrometry based MHC-antigen peptide benchmark dataset is constructed and online database is developed; genomic or whole exome sequencing data is analyzed for MHC typing, sequence and structural features of antigen peptides and MHC binding groove are extracted, a deep learning based MHC-antigen peptide affinity prediction model is proposed and developed; MHC-antigen peptide heterogeneous network is constructed by fusing multi-source data, network topology features and ontology features are combined, and deep feedforward network is trained to predict MHC-antigen peptide affinity.

[0140] The prediction performance of MHC-I and antigen peptide affinity is improved from three aspects: 1) mass spectrometry data of MHC class I binding antigen peptides are integrated, mass spectrometry based MHC-antigen peptide benchmark dataset is constructed, deep neural network based model is pre-trained using IEDB data, then pre-trained model is transferred learning combined with mass spectrometry data to obtain higher performance prediction model, as shown in Figure 5 2) MHC-I and antigen peptide affinity prediction combined with sequence and structural features: sequence of antigen peptide and MHC-I binding groove is extracted from protein database, amino acid sequence is one-hot encoded, and sequence features including physical and chemical, local structure entropy, pairing potential, interaction tendency, etc. are extracted using multiple methods, while secondary structure features, structural neighbor features, solvent accessible surface area, etc. are extracted, these sequence and structure based features are combined to construct training set, gradient boosting regression tree model is learned, important factors of MHC-I molecule and antigen peptide affinity are determined, and online prediction service is developed. 3) MHC-I and MHC-II subtype similarity network and antigen peptide similarity network are calculated, MHC I molecule-antigen peptide association in IEDB is integrated, thereby constructing MHC-antigen peptide heterogeneous network fused with multi-source heterogeneous data, random walk algorithm on two-pass heterogeneous network is used to predict the probability of potential MHC molecule binding antigen peptide.

[0141] In one specific embodiment, TCR recognizes tumor specific antigen and immunogenicity prediction: sequence and structural features of antigen peptide and TCR variable region are extracted from genomic data, experimentally verified TCR-antigen peptide binding dataset is combined, regression analysis and correlation analysis are performed, key factors of TCR recognizing tumor antigen are found; deep learning model is constructed, antigen and TCR related features are used as input, probability of TCR recognizing and binding tumor antigen is predicted; gene mutation-TCR correspondence matrix is constructed by integrating gene mutations of different tumors, probability of tumor specific antigen being recognized by TCR is analyzed from the perspective of tumor clonal evolution; tumor infiltrating lymphocyte (TIL) data and experimentally obtained immunogenicity data are combined, tumor gene expression profile and antigen peptide feature matrix are used as input of deep neural network, immunogenicity of tumor antigen is predicted.

[0142] CD8+ T lymphocytes rely on the complementarity-determining region (CDR3) of their TCRs to recognize and kill tumor cells. The interaction between CDR3 and MHC-presented antigenic peptides is crucial for adaptive immune responses. TCRs are determined by V(D)J recombination, and the human TCR repertoire can accommodate up to 10¹⁵ different molecular types. Unless derived from the same clone, each T lymphocyte has a unique TCR, but different TCRs can bind to the same antigens, posing a significant challenge to the computational prediction of immunogenic tumor antigens. We believe that T cell TCR regions that recognize the same pMHC complex contain conserved sequence features (motifs). We plan to develop three methods to predict the probability of TCR recognizing tumor antigens and activating an immune response: 1) Identification of the main factors of TCR-antigen epitope binding: Collect and integrate experimentally validated antigen epitope-TCR interaction data, extract the sequence and structure of TCR and peptide chains, including the charge, hydrophobicity, two-dimensional structural features of CDR3 and the one-hot encoding of antigens, etc., use group sparse regularized regression analysis to analyze the impact of each feature on TCR recognition of antigens, construct a deep neural network model and train it using antigen-TCR interaction data to predict the probability of specific TCR subtypes recognizing specific antigens. 2) CNN-based TCR-antigen epitope binding prediction: Sequences of the CDR3 region are obtained from RNA-seq data, and autocross covariation is used to convert CDR3 sequences of different lengths into feature matrices of the same dimension. Gene mutations and antigenic peptides are obtained from genomic or exome sequencing data, and one-hot encoding is used to convert them into feature matrices. MHC-antigen peptide affinity prediction and TCR-antigen epitope binding prediction share the features of antigenic peptide sequences and structures. A multi-task deep learning model is constructed to simultaneously perform two tasks: MHC-antigen peptide affinity prediction and TCR-antigen epitope binding prediction. 3) TCR library profiling is performed on RNA-seq data of different tumor-infiltrating lymphocytes (TILs). Gene mutations are obtained by combining tumor cell genomic or exome sequencing data. Immunogenicity datasets obtained from experiments are collected. The tumor gene expression profile, antigenic peptide feature matrix, and CDR3 feature matrix are used as inputs to a deep neural network to predict the immunogenicity of tumor antigens.

[0143] In one specific embodiment, the approval of immune checkpoint inhibitors such as PD-L1 monoclonal antibody and CTLA-4 monoclonal antibody greatly promotes the research of tumor immunotherapy, but the population currently benefiting from immunotherapy is very limited, for example, the objective remission rate of PD-L1 inhibitor single drug is only about 30%. The companion diagnostic indicators TMB and PD-L1 protein level of immune checkpoint inhibitors cannot accurately screen patients who can benefit from PD-L1 monoclonal antibody treatment, because cytotoxic T lymphocytes cannot recognize tumor-specific antigen epitopes. Therefore, screening individual-specific biomarkers as companion diagnostics for immune checkpoint inhibitors by combining multi-omics data is the direction of precision medicine development. The present subject intends to collect and integrate various clinical data of tumor response to immune checkpoint inhibitor PD-L1, progression-free survival and overall survival, construct a benchmark PD-L1 inhibitor clinical data set, and develop two methods for biomarker screening based on this: 1) analyze the correlation of new antigens generated by tumor driver mutations and passenger mutations with the response and prognosis of PD-L1 inhibitors, and find driver gene mutations with statistical significance; 2) obtain gene mutations, MHC subtypes from exome sequencing data of tumor cells and paired normal cells, obtain gene expression and TCR library spectrum from RNA-seq data, integrate antigen peptide sequences, tumor gene expression profiles, MHC typing and TCR library and other related features, combine large-scale clinical immune response data, extract features from high-dimensional heterogeneous data and train a deep neural network model to achieve high-precision prediction of tumor-specific immune checkpoint response and prognosis, and provide more reference indicators for the companion diagnosis of immune checkpoint clinical treatment.

[0144] The present disclosure further discloses a computer program product or system, comprising a computer program which, when executed by a processor, implements the steps of the above-mentioned method for calculating neoantigen load.

[0145] Figure 2 The present disclosure further discloses a computer program product or system, comprising a computer program which, when executed by a processor, implements the steps of the above-mentioned method for calculating neoantigen load.

[0146] The acquisition unit acquires sequencing data of the subject to be tested, wherein the sequencing data is exome sequencing data or panel sequencing data.

[0147] The extraction unit extracts sequence and structural features of antigen peptides, sequence and structural features of MHC binding grooves, and sequences of TCR variable regions based on the sequencing data.

[0148] The affinity unit calculates MHC-antigen peptide affinity based on the sequence and structural features of the MHC binding groove and the sequence and structural features of the antigen peptide.

[0149] The recognition unit calculates the probability of TCR recognizing MHC-antigen peptide based on the sequence of the TCR variable region.

[0150] The computing unit: based on the probability of the TCR identifying the MHC-antigen peptide, the MHC-antigen peptide affinity, the neoantigen load of the to-be-tested person is calculated.

[0151] Figure 3 The computing device for the neoantigen load provided by the embodiment of the present application specifically comprises:

[0152] The memory and the processor; the memory is used for storing program instructions; the processor is used for calling the program instructions, when the program instructions are executed, any one of the above-mentioned neoantigen load calculation methods.

[0153] The embodiment of the present application further discloses a computer readable storage medium, the computer readable storage medium stores a computer program, and the computer program is executed by a processor to execute any one of the above-mentioned neoantigen load calculation methods.

[0154] The embodiment of the present application provides a tumor treatment prognosis prediction method based on a neoantigen load, which comprises:

[0155] Obtaining tumor section data and serum data of a to-be-tested person;

[0156] Performing neoantigen screening based on the tumor section data and the serum data to obtain a neoantigen;

[0157] Calculating the load of the neoantigen based on the above-mentioned neoantigen load calculation method to obtain a neoantigen load;

[0158] Performing prognosis prediction based on the neoantigen load to obtain a prediction result.

[0159] In one specific embodiment, esophageal cancer patient tumor and serum clinical sample sequencing and clinical data collection: among the patients who are first diagnosed as esophageal cancer in the hospital, the cases meeting the following conditions are selected as research objects: the age is between 18-70 years old; the clinical stage is ⅢA-Ⅳ stage, and surgery treatment cannot be performed; there is no history of malignant tumor; there is no history of chemotherapy; before chemotherapy, the patient's blood routine, heart, liver and kidney functions are normal. All research objects adopt a combined chemotherapy scheme mainly using cisplatin or carboplatin. Informed consent. All patients are treated with 4-6 cycles of chemotherapy, and a PD-L1 inhibitor is used during chemotherapy. According to the clinical chemotherapy data, the chemotherapy effect of the patients is evaluated according to the RECIST standard: complete remission (CR)-all target lesions disappear; partial remission (PR)-the sum of the longest diameters of the target lesions is reduced by at least 30%; stable disease (SD)-referring to the minimum value of the sum of the longest diameters at the beginning of treatment, which does not reach the PR standard nor the PD standard; progression (PD)-the sum of the longest diameters of the target lesions is increased by at least 20%, referring to the minimum value of the sum of the longest diameters at the beginning of treatment or when one or more new lesions appear.

[0160] All patients were evaluated for chemotherapy grade 3 and 4 toxicities according to the National Cancer Institute (NCI 3.0) evaluation criteria. Adverse reactions caused by chemotherapy include leukopenia, neutropenia, thrombocytopenia, anemia, nausea, vomiting, diarrhea, etc. Adverse reactions are routinely divided into three groups for follow-up analysis: (i) all 3 or 4 grade toxic reactions; (ii) all 3 or 4 grade hematologic toxic reactions; (iii) all 3 or 4 grade gastrointestinal toxic reactions.

[0161] Referring to the strategy of the United States NIH specimen library construction and tumor cohort study, an esophageal cancer specimen library is established according to the unified standards of the project, which has complete clinical treatment and follow-up data, experimental research data, and tumor tissue and body fluid samples at each treatment and follow-up stage. The third party sequencing agency is commissioned to perform whole exome sequencing on tumor and serum samples, obtain original fastq files and perform mutation analysis, extract tumor antigen sequences, and use the developed neoantigen screening model to calculate the neoantigen load index of each patient. At the same time, the expression level of PD-L1, TMB and MSI, etc. Prognostic indicators are collected as the basis for comparative analysis.

[0162] Clinical verification of the prognostic role of neoantigen load in anti-PD-L1 treatment of esophageal cancer patients: 100 patients were enrolled in a randomized controlled study, and the treatment group was treated with concurrent chemoradiotherapy combined with PD-L1 antibody Imfinzi (durvalumab, durvalumab) for treatment, and the control group was treated with concurrent chemoradiotherapy 5 combined with placebo. The treatment process is strictly in accordance with the NCCN clinical diagnosis and treatment scheme, and the patient's biochemical, immune, imaging and other clinical data are recorded. At the same time, related side effects, mainly evaluating hematologic toxicity, gastrointestinal reactions, cardiac damage, lung damage, skin damage and other side effects. According to the genomic sequencing data of the TCGA public database, the critical value of neoantigen load for stratification of esophageal cancer patients is calculated. Patients are divided into two groups, and the OS, objective response rate (ORR), overall response rate (OR), progression-free survival (PFS) and other prognostic indicators of the two groups of patients are statistically analyzed to determine the statistical significance of neoantigen load in each prognostic indicator. In particular, compared with PD-L1+, TMB and MSI, etc. Index, verify the prognostic role of neoantigen load.

[0163] In one specific instance, the approval and marketing of immune checkpoint inhibitors such as PD-L1 and CTLA-4 monoclonal antibodies has greatly promoted research into tumor immunotherapy. However, the number of people who currently benefit from immunotherapy is very limited; for example, the objective response rate of PD-L1 inhibitor monotherapy is only around 30%. Companion diagnostic markers for immune checkpoint inhibitors, such as TMB and PD-L1 protein levels, cannot accurately screen patients who will benefit from PD-L1 monoclonal antibody treatment because cytotoxic T lymphocytes cannot recognize tumor-specific antigenic epitopes. Therefore, combining multi-omics data to screen for individual-specific biomarkers as companion diagnostics for immune checkpoint inhibitors is a direction for the development of precision medicine. This invention aims to collect and integrate clinical data on the response of various tumors to the immune checkpoint inhibitor PD-L1, progression-free survival, and overall survival, constructing a benchmark PD-L1 inhibitor clinical dataset. Based on this dataset, two methods are developed for biomarker screening: 1) Analyzing the correlation between neoantigens generated by tumor driver mutations and passenger mutations and their response to PD-L1 inhibitors and prognosis, identifying statistically significant driver gene mutations; 2) Obtaining gene mutations and MHC subtypes from exome sequencing data of tumor cells and paired normal cells, obtaining gene expression and TCR library profiles from RNA-seq data, integrating relevant features such as antigen peptide sequences, tumor gene expression profiles, MHC typing, and TCR libraries, and combining large-scale clinical immune response data, extracting features from high-dimensional heterogeneous data and training a deep neural network model to achieve high-precision prediction of tumor-specific immune checkpoint responses and prognosis, providing more reference indicators for companion diagnostics of immune checkpoint clinical treatment.

[0164] In one specific embodiment, esophageal cancer patients underwent anti-PD-L1 clinical treatment. Enrolled patients were treated with the marketed PD-L1 antibody Imfinzi (durvalumab), and the treatment process strictly followed the NCCN clinical treatment protocol. Patients' biochemical, immunological, and imaging clinical data were recorded. Simultaneously, related toxic side effects were evaluated, primarily hematological toxicity, gastrointestinal reactions, cardiac injury, lung injury, and skin damage. Esophageal cancer patients were stratified based on calculated neoantigen load, and prognostic indicators such as overall survival (OS), objective response rate (ORR), overall response rate (OR), and progression-free survival (PFS) were statistically analyzed to determine the statistical significance of neoantigen load in each prognostic indicator. Specifically, the prognostic role of neoantigen load was verified by comparing it with indicators such as PD-L1+, TMB, and MSI.

[0165] The verification result of the verification embodiment shows that assigning inherent weights to the indications can improve the performance of the method compared with the default setting. It can be clearly understood by those skilled in the art that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the foregoing method embodiments, which will not be described here. In the several embodiments provided by the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. The division of the units is only a logical function division. There can be another division manner in actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interfaces, devices or units, and can be electrical, mechanical or other forms. The units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, i.e. can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme. In addition, each functional unit in the various embodiments of the present application can be integrated in one processing unit, or each unit can be a physically independent unit, or two or more units can be integrated in one unit. The integrated unit can be implemented in the form of hardware, or in the form of software functional units. Those skilled in the art can understand that all or part of the steps of the various methods in the above embodiments can be completed by programs instructing relevant hardware, and the programs can be stored in a computer readable storage medium, which can include read only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc.

[0166] Those skilled in the art can understand that all or part of the steps of the above-mentioned embodiment methods can be completed by programs instructing relevant hardware, and the programs can be stored in a computer readable storage medium, and the above-mentioned medium storage can be read only memory, magnetic disk or optical disk, etc.

[0167] The computer device provided by the present application has been described in detail above. For those skilled in the art, according to the idea of the embodiment of the present application, there will be changes in specific implementation and application range. In view of the above, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method for calculating neoantigen load, characterized in that, include: S1. Obtain the sequencing data of the subject, wherein the sequencing data is exome sequencing data or panel sequencing data; S2. Extracting the sequence and structural features of antigen peptides, the sequence and structural features of MHC binding grooves, and the sequence of TCR variable regions based on sequencing data; S3. Calculate the MHC-antigen peptide affinity based on the sequence and structural characteristics of the MHC binding groove and the sequence and structural characteristics of the antigen peptide. S4. Calculate the probability of TCR recognizing MHC-antigen peptides based on the sequence of the TCR variable region; S5. The neoantigen load of the test subject is calculated based on the probability of TCR recognizing MHC-antigen peptide and the affinity of MHC-antigen peptide. The formula for calculating the neoantigen load is as follows: Where N represents the neoantigen load, and n represents the number of tumor clones obtained from the patient's exome sequencing data. b i It indicates affinity. r i c represents the probability that the TCR recognizes the MHC-antigen peptide. j Indicates the expression level of gene j, clone j This represents the tumor clone lineage of gene j.

2. The method for calculating neoantigen load according to claim 1, characterized in that, The extraction process involves obtaining the sequence and structural features of the antigen peptide and the sequence and structural features of the MHC binding groove through a trained first deep network model.

3. The method for calculating neoantigen load according to claim 2, characterized in that, The training process for the first deep network model is as follows: Obtain antigen peptide sequences and MHC binding groove sequences from protein databases; Structural features were obtained by extracting structural features from the sequence; Construct a feature training set based on the aforementioned sequence and structural features; The feature training set is fed into the first deep network model for training to obtain the trained first deep network model.

4. The method for calculating neoantigen load according to claim 3, characterized in that, The structural features include one or more of the following: secondary structure features, structural neighbor features, and solvent-accessible surface area.

5. The method for calculating neoantigen load according to claim 3, characterized in that, The feature extraction method employs one or more of the following: physicochemical extraction, local structure entropy extraction, pairing potential energy extraction, and interaction tendency extraction.

6. The method for calculating neoantigen load according to claim 2, characterized in that, The first deep network model includes one or more of the following: GBRT, XGBoost, LightGBM, CatBoost, Random Forest, and AdaBoost.

7. The method for calculating neoantigen load according to claim 1, characterized in that, The TCR variable region sequence was extracted from the genomic data of the subject.

8. The method for calculating neoantigen load according to claim 1, characterized in that, The probability of TCR recognizing MHC-antigen peptides is calculated using a recognition model, the construction process of which is as follows: To obtain data on antigen epitope-TCR interaction, and the sequences and structures of TCR and peptide chains; TCR features were extracted based on the sequence and structure data of the TCR and peptide chain. The TCR features and antigen epitope-TCR interaction data are input into a fourth deep network model for training to obtain a recognition model.

9. The method for calculating neoantigen load according to claim 8, characterized in that, The TCR features include one or more of the following: charge, hydrophobicity, and two-dimensional structure features of CDR3.

10. The method for calculating neoantigen load according to claim 8, characterized in that, The TCR features were extracted using group sparse regularized regression analysis.

11. The method for calculating neoantigen load according to claim 8, characterized in that, The probability calculation of TCR recognition of MHC-antigen peptides also includes TCR-antigen epitope binding prediction. The probability of TCR recognition of MHC-antigen peptides is obtained through the recognition model and TCR-antigen epitope binding prediction. The TCR-antigen epitope binding prediction is obtained by a trained TCR-antigen epitope binding prediction model.

12. The method for calculating neoantigen load according to claim 11, characterized in that, The process of constructing the trained TCR-antigen epitope binding prediction model is as follows: Obtain the sequence, gene mutation data, and antigenic peptides of the CDR3 region; The sequence, gene mutation data, and antigenic peptides of the CDR3 region are converted into a feature matrix. The feature matrix is ​​fed into a neural network for training to obtain a TCR-antigen epitope binding prediction model.

13. The method for calculating neoantigen load according to claim 12, characterized in that, The feature matrix of an antigenic peptide includes both sequence and structural features.

14. The method for calculating neoantigen load according to claim 12, characterized in that, The TCR-antigen epitope binding prediction model also includes MHC-antigen peptide affinity calculation, which uses the characteristics of antigen peptide sequence and structure to predict affinity results.

15. The method for calculating neoantigen load according to claim 8, characterized in that, The probability calculation of TCR recognizing MHC-antigen peptides also includes the prediction of the immunogenicity of tumor antigens. The probability of TCR recognizing MHC-antigen peptides is obtained through the recognition model and the prediction of the immunogenicity of tumor antigens. The prediction of the immunogenicity of tumor antigens is obtained by a trained tumor antigen immunogenicity prediction model.

16. The method for calculating neoantigen load according to claim 15, characterized in that, The process of constructing the immunogenicity prediction model for the tumor antigen is as follows: Obtain RNA-seq data, tumor cell genome data, and immunogenicity data from different tumor-infiltrating lymphocytes; TCR library profiling analysis was performed on the RNA-seq data to obtain tumor gene expression profile data, and the CDR3 sequence was obtained from the RNA-seq data; Gene mutation and antigenic peptide data were obtained based on the tumor cell genomic data. Feature matrix data was obtained by feature extraction from CDR3 sequence and antigenic peptide data; The tumor gene expression profile data, gene mutation data, and feature matrix data are input into a neural network for training to obtain an immunogenicity prediction model.

17. A computer program product comprising a computer program or instructions, characterized in that, The computer program or instructions are executed by a processor to implement the method for calculating neoantigen load according to any one of claims 1-16.

18. A computer device comprising a memory, a processor, and a computer program or instructions stored in the memory, characterized in that, The computer program or instructions are executed by a processor to implement the method for calculating neoantigen load according to any one of claims 1-16.

19. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, The computer program or instructions are executed by a processor to implement the method for calculating neoantigen load according to any one of claims 1-16.

20. A computer program product comprising a computer program or instructions, characterized in that, The computer program or instructions are executed by a processor using a tumor treatment prognosis prediction method based on neoantigen load, including: Obtain tumor slide data and serum data from the subject; New antigens were obtained by screening for new antigens based on the tumor slice data and serum data. The neoantigen load is obtained by calculating the neoantigen load using the neoantigen load calculation method according to any one of claims 1-16; Prognostic predictions were obtained based on the neoantigen load.

21. A computer device comprising a memory, a processor, and a computer program or instructions stored in the memory, characterized in that, The computer program or instructions are executed by a processor using a tumor treatment prognosis prediction method based on neoantigen load, including: Obtain tumor slide data and serum data from the subject; New antigens were obtained by screening for new antigens based on the tumor slice data and serum data. The neoantigen load is obtained by calculating the neoantigen load using the neoantigen load calculation method according to any one of claims 1-16; Prognostic predictions were obtained based on the neoantigen load.

22. A computer-readable storage medium having a computer program or instructions stored thereon, characterized in that, The computer program or instructions are executed by a processor using a tumor treatment prognosis prediction method based on neoantigen load, including: Obtain tumor slide data and serum data from the subject; New antigens were obtained by screening for new antigens based on the tumor slice data and serum data. The neoantigen load is obtained by calculating the neoantigen load using the neoantigen load calculation method according to any one of claims 1-16; Prognostic predictions were obtained based on the neoantigen load.

Images