Identification of cell surface antigens that induce T cell responses and their use

Abstract

The present invention relates to a computer-implemented method for identifying at least one antigen in a subject that is predicted to induce a T cell response that attacks unhealthy cells in the subject. The invention further relates to the use of said antigens in the design of personalized vaccines, and the use of such personalized vaccines and personalized vaccine compositions in kits and methods of treating disease.

Description

[Technical Field]

[0001] The present invention relates to a system and computer-implemented method for predicting the efficacy of T cell responses induced by HLA-presented surface antigens in different individuals, particularly in infectious diseases, cancer, and autoimmune diseases. This prediction can be used to develop a range of personalized medicines, including vaccines, T cell therapies, and diagnostic tests. Furthermore, the present disclosure details a method for creating personalized vaccines from peptide antigens containing a range of highly immunogenic epitopes capable of inducing strong CD8 and CD4 T cell responses. Furthermore, this technology provides useful insights for predicting individual responses to vaccines and T cell therapies, enhancing the accuracy and efficacy of personalized medicine interventions. [Background technology]

[0002] An antigen is defined as a peptide that elicits an immune response. Immune responses to diseases such as cancer and viral infections are known to be regulated by a group of highly polymorphic related proteins encoded by human leukocyte antigens (HLA), also known as major histocompatibility complexes (MHC). After infection, viral proteins are processed into epitopes. Epitopes are peptides that bind to HLA molecules. Some of these epitopes are transported to the cell surface and activate T cell receptors (TCRs). This mechanism is schematically illustrated in Figure 1, which shows a TCR in a T cell being activated by epitope 150 that is tightly bound to an individual's HLA allele. Antigens can be defined by their specificity (i.e., amino acid sequence) and potency (i.e., the strength with which they stimulate the TCR to elicit a T cell response). Upon stimulation, responding T cells proliferate and, when their TCRs recognize infected cells displaying the same epitope, kill these cells. After clearance of infected cells, a subset of epitope-specific T cells forms a memory population that can rapidly respond to new diseases containing the same epitope.

[0003] Human HLA genotypes are highly diverse. Each individual inherits one allele of HLA-A, one allele of HLA-B, and one allele of HLA-C from each parent, and therefore each individual has six different HLA class I molecules. These HLA class I molecules typically present epitopes with lengths of 8 to 14 amino acids. These epitopes are expressed by CD8 + presented to cytotoxic T cells. Individuals also have CD4 + HLA class II molecules also present epitopes to helper T cells. The epitopes presented by HLA class II molecules are longer than those presented by HLA class I molecules, generally 11–20 amino acids in length. HLA class II molecules are encoded by three distinct loci: HLA-DR, HLA-DQ, and HLA-DP. These molecules contain two identical peptides (e.g., DQA and DQB) in the α and β chains. HLA-DR is the most polymorphic, with over 700 known alleles for HLA-DRB and only three variants for HLA-DRA. In contrast, both HLA-DQ and HLA-DP chains are polymorphic. However, only a few alleles predominate in HLA-DP, the most notable being the heterodimer DPA1*0103 / DPB1*0401 (DP401).

[0004] The potency of T cell responses is one of the most complex aspects of human immunology, relevant to numerous diseases, including but not limited to COVID-19 (a disease caused by SARS-CoV-2 infection). For example, over 1,400 epitopes have been identified from SARS-CoV-2 that elicited T cell responses in at least 1 out of 1,187 individuals (see Reference 5 (Bukhari et al.)). Experimental methods can only test a small subset of putative epitopes due to sample limitations, and antigen-specific T cell responses are not reproducible across individuals due to HLA diversity and disease variability. Therefore, conventional techniques are limited to measuring the potency of a small number of epitopes predicted to bind to one or two HLA molecules. For example, EP3370065 describes a method for identifying polypeptide fragments capable of binding to at least two HLA molecules in a specific human subject as immunogenic in that subject.

[0005] Systems for predicting epitope-HLA binding have also been developed. For example, Reference 5 (Bukhari et al.) describes several machine learning (ML) models for predicting epitope-HLA binding. One of these models, known as the NetMHCpan-4.1 model, is described in detail in "NetMHCpan-4.1 and NetMHCIIpan-4.0: Improved Predictions of MHC Antigen Presentation by Concurrent Motif Deconvolution and Integration of MS MHC Eluted Ligand Data" by Reynisson et al., published in Nucleic Acids Research (2020). This model was trained on 850,000 quantitative binding affinities (BAs) and mass spectrometry eluted ligands (ELs) to accurately predict the strength of epitope-HLA interactions. The NetMHCpan-4.1 model calculates a score for any epitope-HLA allele pair, called the EL score, which represents the likelihood that the epitope will be presented on the cell surface by that HLA allele. However, as pointed out in reference 7 (Saini et al.), such predicted epitopes rarely induce T cell responses in HLA allele-matched individuals.

[0006] Applicants have recognized the need for new methods for determining T cell responses to individuals. Summary of the Invention

[0007] Summary of the Invention In a first aspect of the present invention, there is provided a computer-implemented method for identifying in a subject at least one antigen that is expected to induce a T cell response that attacks unhealthy cells in the subject, the method comprising the steps of: obtaining subject data identifying a plurality of major histocompatibility complex (MHC) molecules of the subject; obtaining sequence data for a plurality of epitopes within a protein expressed in the unhealthy cell, each epitope being an amino acid sequence within the protein; obtaining, for each epitope within the plurality of epitopes, a potency score indicative of the likelihood of presenting the epitope on the surface of the identified plurality of MHC molecules; generating a ranked list of epitopes based on the determined potency scores; identifying at least one antigen by selecting at least one epitope that is highly ranked in the ranked list; and Outputting at least one of a ranked list and sequence data for the at least one identified antigen.

[0008] In one embodiment, further comprising: selecting a plurality of epitopes that are highly ranked in the ranked list; identifying subitopes for each selected epitope using a directed graph network; and Identifying each subtope common to the selected epitopes.

[0009] In one embodiment, the first aspect of the invention further comprises identifying at least one antigen by selecting common subtopes that are highly ranked in the ranked list.

[0010] In one embodiment, the method further comprises identifying at least one antigen by selecting the longest common subtopes.

[0011] In one embodiment, the method further comprises identifying the shortest common subtopes as the target sequence.

[0012] In one embodiment, obtaining the potency score comprises selecting an epitope from the plurality of epitopes, and calculating an epitope weight score for each epitope in the plurality of epitopes; obtaining, for each identified MHC molecule, a probability score indicating the probability that each MHC molecule carries the selected epitope; and summing at least some of the probability scores to calculate an epitope weight score for the selected epitope.

[0013] In one embodiment, in the method of aspect 4, the epitope weight score is calculated from:

number

[0014] In one embodiment, the method further comprises calculating at least one additional score for each epitope that indicates whether subtopes of each epitope are capable of eliciting the same T cell response.

[0015] In one embodiment, the method further comprises calculating at least one additional score for each epitope that indicates whether subtopes of each epitope are capable of eliciting the same T cell response.

[0016] In a second aspect of the present invention, there is provided a method for designing a personalised vaccine for inducing a T cell response that attacks unhealthy cells in a subject, the method comprising the steps of: obtaining subject data identifying a plurality of major histocompatibility complex (MHC) molecules for the subject, the subject data identifying at least one of a set of HLA class I molecules and a set of HLA class II molecules; obtaining sequence data for a plurality of epitopes within a protein expressed in the unhealthy cell, each epitope being an amino acid sequence within the protein; obtaining, for each epitope within the plurality of epitopes, a first potency score indicative of the likelihood of each epitope being presented on the surface of the identified set of HLA class I molecules; generating a first ranked list of epitopes based on the determined first potency scores; selecting a plurality of epitopes that are highly ranked in the first ranked list; identifying subtopes of each selected epitope using a directed graph network; identifying each subtope common to two or more of the selected epitopes; identifying at least one first antigen by selecting at least one subtopes that is itself a highly ranked epitope in the first ranked list; and Outputting the identified at least one first antigen for use in a personalized vaccine.

[0017] In one embodiment, the method further comprises the steps of: obtaining, for each epitope within the plurality of epitopes, a second potency score indicative of the likelihood that each epitope will be presented on the surface of the identified set of HLA class II molecules; generating a second ranked list of epitopes based on the determined second potency scores; selecting a plurality of epitopes that are highly ranked in the second ranked list; identifying subtopes of each selected epitope using a directed graph network; identifying each subtope common to two or more of the selected epitopes; identifying at least one second antigen by selecting at least one subtopes that is itself a highly ranked epitope in the second ranked list; and Outputting the identified at least one second antigen for use in a personalized vaccine.

[0018] In one embodiment, the further step includes: determining whether the identified first antigen is a highly ranked epitope in the second ranked list; and If the first antigen is a highly ranked epitope in the second ranked list, designing a personalized vaccine based on the first antigen.

[0019] In one embodiment, the antigens are highly ranked in the ranked list, are capable of being presented by MHC class I and / or MHC class II molecules on the surface of cells of the subject, induce a CD4 and / or CD8 T cell response, and contain multiple epitopes that share at least one common sequence capable of eliciting the same T cell response, and optionally overlapping epitopes.

[0020] In a third aspect of the present invention, there is provided a method for determining the efficacy of an immune response as a means of controlling safety and efficacy in a recipient of a personalised vaccine composition comprising at least one antigen identified by a method according to an aspect of the present invention, the method comprising the steps of: generating all potential epitopes from the sequence of the selected antigen; creating a ranked list based on the potency scores of these epitopes; confirming that epitopes derived from proteins expressed in unhealthy cells of the recipient induce a strong immune response as indicated by a high potency score, thereby ensuring efficacy; confirming that the epitopes derived from the cell membrane-penetrating peptide and / or excipient are immunologically inactive as indicated by a low potency score, thereby ensuring safety; A step to confirm that epitopes with high potency scores are not components of proteins expressed in healthy cells, thereby further contributing to the safety of personalized vaccines.

[0021] In a fourth aspect of the present invention, there is provided a personalized vaccine composition comprising a peptide antigen comprising a plurality of top-ranked epitopes selected according to the method of the aspects of the present invention, derived from a protein expressed in an unhealthy cell of a recipient, and optionally a cell membrane-penetrating peptide and / or an excipient.

[0022] In one embodiment, the method further comprises at least two peptide antigens derived from proteins expressed in unhealthy cells of the recipient, the peptide antigens comprising a plurality of the top-ranking epitopes selected according to the methods according to aspects of the invention.

[0023] In one embodiment, the cell membrane-permeable peptide is positioned between two antigens.

[0024] In a fifth aspect of the present invention, there is provided a method of treating or preventing a disease in a subject comprising administering at least one personalized vaccine according to an aspect of the present invention, wherein the at least one personalized vaccine composition is administered alone or in combination with an additional therapeutic agent, said administration being simultaneous or sequential, and optionally wherein at least two personalized vaccine compositions are administered.

[0025] In one embodiment, the disease is cancer or an autoimmune disease, or a viral infection.

[0026] In a sixth aspect of the invention, there is provided a method of inducing an antigen-specific immune response in a subject, the method comprising administering to the subject a personalised vaccine composition according to an aspect of the invention.

[0027] In a seventh aspect of the invention, there is provided a kit comprising a number of items necessary to prepare a personalised vaccine composition according to an aspect of the invention for an individual, said kit comprising: two synthetic peptides comprising at least one antigen selected according to an embodiment of the present invention and further comprising at least a portion of a cell membrane-penetrating peptide; and A means for covalently linking two synthetic peptides during the preparation of a personalized vaccine, thereby reconstituting the function of a cell membrane-penetrating peptide.

[0028] In an eighth aspect of the present invention, there is provided a method for preparing a personalized peptide vaccine or therapeutic composition, the method comprising the steps of preparing a first amino acid sequence and preparing a second amino acid sequence, both of which comprise an antigen comprising a set of top-ranked epitopes derived from a protein expressed in an individual's unhealthy cells and at least a portion of a cell membrane-penetrating peptide, and covalently linking the first amino acid sequence and the second amino acid sequence to form the personalized vaccine so as to reconstitute the function of the cell membrane-penetrating peptide located between the first antigen and the second antigen.

[0029] Detailed Summary The present invention provides an apparatus and method as set out in the accompanying claims. Other features of the invention will become apparent from the dependent claims and the following description.

[0030] We describe a computer-implemented method for identifying at least one antigen in a subject that is predicted to induce a T cell response that attacks unhealthy (also referred to as diseased or infected) cells of the subject. Because each unhealthy cell expresses a set of disease-associated proteins, T cells that specifically destroy the unhealthy cells and spare healthy cells can restore the patient's health. The method includes the following steps: obtaining subject data identifying a plurality of major histocompatibility complex (MHC) molecules for the subject; obtaining sequence data for a plurality of epitopes within the protein, each epitope being an amino acid sequence within the protein; generating, for each epitope within the plurality of epitopes, a potency score indicating the likelihood that each identified MHC molecule will present the epitope on a cell surface; generating a ranked list of epitopes, which is a subset of the plurality of epitopes ranked based on the determined potency scores; selecting at least one top-ranked epitope to identify at least one antigen; and outputting at least one of the ranked list and sequence data for the identified at least one antigen.

[0031] The disease may be a viral infectious disease such as SARS-CoV-2 or cancer. The unhealthy cells may be virus-infected cells or cancer cells. An antigen may be defined as a peptide that triggers an immune response. An antigen may be defined by its specificity (i.e., amino acid sequence) and its potency (i.e., the strength with which it stimulates the T cell receptor (TCR) to trigger a T cell response).

[0032] The method may further include the steps of selecting a plurality of epitopes that are highly ranked in the ranked list, identifying subitopes of each selected epitope using a directed graph network, and identifying each subitopes that are common to the selected plurality of epitopes. Each subitopes that is common to the selected plurality of epitopes may be referred to as a core amino acid sequence.

[0033] At least one antigen can be identified by selecting a common subitopes (also called subepitopes) that rank highly in the ranked list (i.e., by selecting a highly ranked core). It will be understood that the most common core is usually the shortest sequence recognized by the TCR. A longer core contains multiple smaller cores and is therefore more likely to induce a strong T cell response. Thus, the method can include identifying at least one antigen by selecting the longest epitope that includes a set of highly ranked epitopes with a common subitopes. Typically, cell surface antigens are longer than highly ranked epitopes with subitopes. The potency score of such an antigen can be calculated by summing the potency scores of overlapping epitopes with the same score.

[0034] The method may include identifying the shortest common subtopes as target sequences for TCRs, for example, for T cell therapy or other treatments discussed below.

[0035] Obtaining the efficacy score may include calculating an epitope weight score for each epitope in the plurality of epitopes. This includes the steps of selecting an epitope from the plurality of epitopes; obtaining a probability score for each identified MHC molecule indicating the probability that each MHC molecule will carry the selected epitope; and summing at least some of the probability scores to calculate the epitope weight score for the selected epitope. In other words, each epitope is paired with each self-MHC molecule to calculate the epitope weight score for the epitope.

[0036] Each probability score may be obtained from a database containing probability scores for multiple MHC molecule-epitope pairs. The probability score used to calculate the epitope weight score may be the eluted ligand score (ELS). The eluted ligand score may be calculated using the scoring mechanism (algorithm) taught in "NetMHCpan-4.1 and NetMHCIIpan-4.0: Improved Predictions of MHC Antigen Presentation by Concurrent Motif Deconvolution and Integration of MS MHC Eluted Ligand Data" by Reynisson et al., published in Nucleic Acids Research in 2020. These algorithms are trained on experimental data, including binding affinities between epitopes (ligands) and MHC molecules and ligand elution data. Because eluted ligands pass through the natural antigen processing and presentation pathway, the ligand elution data contains unique information that cannot be obtained when only epitope-HLA binding is considered. Furthermore, high-throughput ligand elution assays make it possible to identify thousands of natural ligands in a single experiment, and large training datasets are available. The NetMHC method described in the paper provides an eluted ligand score to estimate the likelihood that an epitope will be eluted from a given MHC molecule, and it performs better in predicting epitopes on cell surfaces than methods based on binding affinity data. As described in detail below, a database can be constructed using large amounts of experimental data to predict the density of epitopes presented on the cell surface by MHC molecules. The database can store the eluted ligand score for each paired MHC molecule and epitope.

[0037] As noted above, the epitope weight score EWS can be calculated by summing at least some of the probability scores. Thus, EWS can be a weighted sum of at least some of the probability scores. The epitope weight score can be expressed as:

number

[0038] This method may further include calculating at least one additional score for each epitope, taking into account overlapping subepitopes (also called subitopes) that can trigger the same T cell receptor. A subitopes is a fragment of the epitope under consideration. The at least one additional score is selected from a left (or first) subitopes score and a right (or second) subitopes score, which are calculated based on the epitope weight scores of the subitopes in the left and right subgraphs, respectively. The left subgraph includes all subitopes that are one amino acid removed from the right side of the upper epitope and have a length exceeding a sequence threshold. Similarly, the right subgraph includes all subitopes that are one amino acid removed from the left side of the upper epitope and have a length exceeding a sequence threshold. It is understood that there is overlap between the right and left subgraphs. Therefore, the right subitopes score (SWS2) may exclude subepitopes already included in the left subitopes score (SWS1).

[0039] The left subitopic score (SWS1) and the right subitopic score (SWS2) may be calculated as follows:

number

[0040] The method may further include calculating an overall weight score OWS for each epitope in the plurality of epitopes. For example, the overall weight score OWS may be a combination of the epitope weight score and at least one additional score. The overall weight score OWS may be calculated from:

number

number

[0041] The subject may be an animal or a human. If the subject is a human, the subject data may identify at least a portion of the subject's human leukocyte antigen (HLA) genotype. The subject data may identify at least one of a set of HLA class I molecules and a set of HLA class II molecules. The set of HLA class I molecules may include six molecules (encoded by three different loci: two HLA-A, two HLA-B, and two HLA-C). The set of HLA class II molecules may include 12 molecules (encoded by three different loci: HLA-DR, HLA-DQ, and HLA-DP). Thus, this method uses multiple HLA molecules (or alleles) rather than a single HLA allele (or a small subset of HLA alleles). Considering all self-HLA reflects the fact that the immune response is based on T cells responding to epitopes presented by multiple HLAs.

[0042] The data identifying a subject's HLA genotype can be multiple numbers (e.g., a minimum of four digits and a maximum of eight digits for HLA class I and II molecules). These multiple numbers represent the amino acid sequence of the epitope-binding pocket of the HLA molecule. The HLA genotype data can be obtained, for example, by a clinician, a nurse, or the subject entering the data into a system. Modern molecular pathology can sequence the entire HLA genome from a single swab of the oral mucosa or a small blood sample. The HLA genotype can be accurately determined using standard techniques. The HLA genotype data can be the complete sequence of the HLA gene, if available.

[0043] If the subject data identifies a set of HLA class I molecules, each of the multiple epitopes can be an amino acid sequence of 8 to 14 amino acids. When calculating the left and right subtopes scores as described above, i ranges from a to b, so a is 8 and b is 14. If the subject data identifies a set of HLA class II molecules, each of the multiple epitopes can be an amino acid sequence of at least 9 amino acids. The maximum length is 20 amino acids, but epitopes can exceed 20 amino acids. In this example, a is 9 and b is 20. By way of example only, for the SARS-CoV-2 protein, there can be 70,000 and 100,000 smaller sequences of subtopes and epitopes that pair with HLA class I and HLA class II molecules, respectively.

[0044] The method may further include obtaining, for each epitope within the plurality of epitopes, a first potency score indicating the likelihood that each molecule identified in the set of HLA class I molecules will present the epitope on a cell surface. The method may also include obtaining, for each epitope within the plurality of epitopes, a second potency score indicating the likelihood that each molecule identified in the set of HLA class II molecules will present the epitope on a cell surface. The first and second potency scores may be used to generate a first ranked list and a second ranked list: the first ranked list covers epitopes that are highly ranked for HLA class I molecules, and the second ranked list covers epitopes that are highly ranked for HLA class II molecules. There may be some overlap between the first and second ranked lists, and highly ranked epitopes that appear in both lists may elicit both CD4+ and CD8+ T cell responses.

[0045] The protein can be any suitable protein expressed in a cell. For example, in the case of T cell-based drug development, the protein can be a protein specifically expressed in diseased cells to avoid T cell-mediated killing of healthy cells. The protein can be expressed in infected cells, e.g., one or more SARS-CoV-2 proteins, and the disease can be COVID-19. Alternatively, the protein can be expressed in tumor cells, e.g., AKAP-4, and the disease can be cancer, e.g., metastatic breast cancer.

[0046] Potency of T cell responses The potency of T cell responses is one of the most complex characteristics of human immunology, relevant not only to cancer and COVID-19, but also to many other diseases. Experimental methods can only measure a small subset of immunogenic epitopes in a subject due to sample limitations. To date, no method has been developed to rank protein epitopes expressed in tumors and infected cells in a subject based on their potency in stimulating individual T cells. Similarly, no method exists to determine the potency of T cell responses to an individual's tumor or a virus like SARS-CoV-2 and identify the individual's T cell antigens that represent the top-ranked epitopes most likely to elicit a T cell response. The phenotypic investigation of T cell responses described here required the development of novel methods. The breadth and magnitude of T cell responses to tumor and viral proteins are determined by the highest density (top-ranked) of epitopes capable of activating T cells.

[0047] The extreme diversity of human HLA genotypes is essential to ensure that at least some individuals within a population can mount an appropriate immune response to an emerging infectious disease (e.g., SARS-CoV-2) through unique, epitope-specific T cell responses, maximizing the probability of survival. Polymorphic HLA genotypes are likely the reason behind the strong CD8+ and CD4+ T cell responses in asymptomatic individuals infected with SARS-CoV-2 or other viruses. Subjects with weak T cell responses are thought to be more likely to develop symptomatic disease (e.g., COVID-19).

[0048] Thus, the method includes determining a total score using at least a portion of the efficacy weighted scores of the plurality of epitopes, and comparing the total score to an overall threshold to predict the subject's response to the disease. Thus, in another aspect, we describe a method for predicting a subject's response to a disease that causes unhealthy cells in the subject. The method includes the following steps: obtaining subject data identifying a plurality of major histocompatibility complex (MHC) molecules of the subject; obtaining sequence data of a plurality of epitopes within a protein expressed in the unhealthy cells, where each epitope is an amino acid sequence within the protein; obtaining a efficacy score for each epitope within the plurality of epitopes, indicating the likelihood that each identified MHC molecule will present the epitope on its surface; generating a ranked list of epitopes based on the determined efficacy scores; determining a total score using at least a portion of the efficacy weighted scores of the plurality of epitopes, and outputting at least one of the ranked list and the total score.

[0049] Determining the total score may involve summing the potency scores above a potency threshold. That is, the sum may be considered a weighted sum, with scores below the threshold weighted at 0 and scores at or above the threshold weighted at 1. Setting the potency threshold may include ranking each epitope based on its potency score and selecting the potency score value for a particular rank as the potency threshold. Alternatively, the total score may be the average potency score of a certain number (e.g., 100) of the highest-ranking epitopes (when ranked by their scores).

[0050] As another alternative, the total score may be determined by a machine learning model. The total score may be the probability that the epitope is an immunogenic epitope for that individual. The machine learning model may determine the total score based on a feature vector including some or all of the epitope weight score, left subtope score, right subtope score, and overall weight score. Other features may be input into the feature vector, such as a "log fold change" that quantifies the efficacy of the epitope compared to no epitope or a baseline and indicates the proliferation of antigen-specific T cells after infection of the body. Other features may include, for example, the major cores of each of the subject's HLA-A, HLA-B, and HLA-C class I alleles. The core is the pattern or sequence of amino acids within the epitope that is recognized by the TCR.

[0051] Patient stratification If the total score is less than the efficacy threshold, the subject can be classified as having a weak T cell response.Weak T cell response indicates poor prognosis of disease.That is, the subject has a high risk of disease, and therefore needs treatment and / or vaccination.If the total score is equal to or greater than the efficacy threshold, the subject can be classified as having a good T cell response.Good T cell response indicates good prognosis of disease, and therefore may be able to avoid treatment and / or vaccination.

[0052] For some diseases, such as COVID-19 and cancer, the challenge is how to vaccinate individuals at high risk of developing the disease. One strategy can be to induce T cell responses in individuals with HLA genotypes that only support weak T cell responses to kill infected cells and tumor cells. These individuals may experience symptomatic disease. COVID-19 patients with symptomatic disease are also more likely to spread the virus to more people than asymptomatic individuals. In another embodiment, therefore, there is a method for stratifying the group of subjects to be vaccinated by predicting the response of each subject to the above-mentioned tumor or disease, and classifying subjects whose total score is below the efficacy threshold as those who are prioritized for vaccination.

[0053] This method can also be used to stratify patient groups to determine treatment.In another embodiment, there is a method for stratifying a group of subjects to be treated by predicting the response of each subject to tumor or disease as described above, and classifying subjects whose total score is less than the efficacy threshold as treatment priority.Similarly, in another embodiment, there is a method for treating a subject, the method comprises predicting the subject's response to disease as described above.If the efficacy score is less than the overall threshold, a first type of treatment can be recommended, and if the efficacy is equal to or greater than the overall threshold, a second type of treatment that is less invasive than the first type of treatment can be recommended.The invasive treatment can be a vaccine combined with T cell therapy or other treatments (for example, a combination of drugs).The less invasive treatment can be a vaccine alone or no treatment, because the immune system is more likely to eliminate the disease.

[0054] Vaccination In another embodiment, there is a method for designing a personalized vaccine to induce a T cell response against a disease-associated protein, the method comprising selecting one or more of at least one output antigen generated by the method for incorporation into the vaccine. As described above, the subject data may identify at least one of a set of HLA class I molecules and a set of HLA class II molecules. The personalized vaccine includes two or more T cell antigens to induce CD8+ and CD4+ T cell responses against proteins expressed on unhealthy cells, such as tumor cells or infected cells. Such personalized vaccines may be designed by covalently linking two T cell antigens that are co-expressed in diseased cells derived from the same or different proteins. Alternatively, personalized vaccines may be designed by identifying T cell antigens with a set of epitopes that can be simultaneously presented by both the subject's HLA class I and class II molecules. Such antigens presented by the subject's own HLA molecules contain a "core" that stimulates the same TCR.

[0055] In another aspect, there is a method for designing a personalized vaccine for inducing a T cell response that attacks unhealthy cells in a subject, the method comprising the steps of: obtaining subject data identifying a plurality of major histocompatibility complex (MHC) molecules for the subject, the subject data identifying at least one of a set of HLA class I molecules and a set of HLA class II molecules; obtaining sequence data for a plurality of epitopes within a protein expressed in the unhealthy cells, each epitope being an amino acid sequence within the protein; and obtaining, for each epitope within the plurality of epitopes, a first efficacy score indicative of the likelihood that each epitope of the identified set of HLA class I molecules will be presented on the cell surface. generating a first ranked list of epitopes based on the determined first efficacy score; selecting a plurality of epitopes that are highly ranked in the first ranked list, identifying subtopes of each selected epitope using a directed graph network, and identifying each subtope that is common to two or more of the selected epitopes; identifying at least one first antigen by selecting at least one subtope that is itself a highly ranked epitope in the first ranked list; and outputting the identified at least one first antigen as a personalized vaccine for the prevention or treatment of a disease caused by unhealthy cells in the subject.

[0056] The method may then include the following steps: for each epitope within the plurality of epitopes, determining a second potency score indicating the likelihood that each epitope will be presented on the surface of the set of identified HLA class II molecules; generating a second ranked list of epitopes based on the determined second potency scores; selecting a plurality of epitopes that are highly ranked in the second ranked list; identifying subtopes for each selected epitope using a directed graph network or other method; identifying each subtope that is common to a plurality of the selected epitopes; identifying at least one second antigen by selecting at least one subtope that is itself a highly ranked epitope in the second ranked list; and outputting the identified at least one second antigen as a personalized vaccine. Alternatively, the first highly ranked set of epitopes may overlap with the second highly ranked set of epitopes. In this example, the antigen included in the personalized vaccine may comprise a set of epitopes presented by both class I and class II HLA molecules on the cell surface, and such a single antigen can stimulate both CD8 and CD4 T cell responses in a subject.

[0057] Another aspect of the invention is a computer-implemented method for developing an industrial vaccine (or a universal vaccine) suitable for protecting a subpopulation of patients. The industrial vaccine may be designed by selecting a plurality of personal vaccines, selecting at least one antigen that is more frequently used in the personalized vaccines, and including the at least one selected antigen in the universal vaccine.

[0058] While each personalized vaccine is designed to induce a strong T cell response against tumor-specific antigens in related individuals (e.g., subjects with matching HLA genotypes), a universal vaccine may not be suitable for every individual. Therefore, the efficacy of a universal vaccine may be evaluated in new subjects before administering the vaccine. The computer-implemented method of the present invention can identify subgroups of patients likely to respond to a vaccine. Potential responders must have an HLA genotype that supports a strong T cell response to at least one epitope in a commercial vaccine. Therefore, the above method can be used to predict a subject's response to a commercial vaccine using the subject's HLA genotype. For example, the prediction method may include the following steps: obtaining subject data identifying multiple major histocompatibility complex (MHC) molecules for the subject; obtaining sequence data for each epitope in the vaccine, where each epitope is an amino acid sequence; obtaining a potency score for each epitope in the vaccine that indicates the likelihood that each epitope of the identified MHC molecule will be presented on the cell surface; and predicting the subject's response to the vaccine based on the potency score. Predictions based on efficacy scores can be made, for example, by comparison to machine learning models and / or thresholds, as described above.

[0059] T cell therapy In another aspect, there is a method for identifying targets for T cell therapy derived from proteins expressed in unhealthy cells, the method comprising the steps of: obtaining subject data identifying a plurality of major histocompatibility complex (MHC) molecules of the subject; obtaining sequence data for a plurality of epitopes within the protein, each epitope being an amino acid sequence within the protein; obtaining, for each epitope within the plurality of epitopes, a potency score indicating the likelihood that each epitope will be presented on the surface of the identified plurality of MHC molecules; selecting a plurality of epitopes that are highly ranked in a ranked list, identifying subtopes for each selected epitope using a directed graph network, and identifying each subtope that is common to two or more of the selected epitopes; determining the length and sequence of each identified subtopes, and outputting the subtopes with the shortest length (also referred to as the core) as the target for the TCR.

[0060] As described above, the subject data may identify at least one of a set of HLA class I molecules and a set of HLA class II molecules. One or more cores are selected as targets for personalized T cell therapy. Such personalized T cell therapy may be designed for unhealthy cells that simultaneously express two or more cores of T cell antigens derived from the same or different proteins. Alternatively, personalized T cell therapy may be designed to target a core of T cell antigens simultaneously presented by both HLA class I and class II molecules. Such antigens presented by self-HLA molecules comprise a "core" that stimulates the TCR of therapeutic T cells.

[0061] Another aspect of the invention is a computer-implemented method for developing industrial T cell therapies (or universal T cell therapies) suited to patient subgroups. The industrial T cell therapy may be designed by selecting multiple targets for a personalized T cell therapy, selecting at least one target that is more frequently used in individualized T cell therapy, and including the selected at least one target in the universal T cell therapy.

[0062] Each personalized T cell therapy is designed to kill cells expressing tumor-specific antigens in the relevant individual, but a generic T cell therapy may not be suitable for every individual. Therefore, the efficacy of T cell therapy in a subject may be evaluated before administering the T cell therapy. The computer-implemented method of the present invention can identify a subgroup of patients who are likely to respond to T cell therapy. Potential responders must have an HLA genotype that supports at least one of the top-ranked epitopes targeted by the T cell therapy. Therefore, the above method can be used to predict a subject's response to industrial T cell therapy using the subject's HLA genotype. For example, a method for predicting a subject's response to T cell therapy may include the following steps: obtaining subject data identifying a plurality of major histocompatibility complex (MHC) molecules of the subject; obtaining sequence data for each epitope targeted by the T cell therapy, where each epitope is an amino acid sequence; obtaining, for each epitope targeted by the T cell therapy, a efficacy score indicating the likelihood that the identified MHC molecule will present the epitope on its cell surface; and predicting the subject's response to the T cell therapy based on the efficacy score. Predictions based on the efficacy score may be made, for example, by comparison with a machine learning model and / or a threshold, as described above.

[0063] Personalized vaccine and therapeutic compositions and methods for testing safety and efficacy in recipients In one aspect of the present invention, there is provided a personalised vaccine or therapeutic composition, and optionally a cell membrane-permeable peptide, prepared by a method according to the above aspect of the present invention.

[0064] In another aspect, there is provided a personalized vaccine or therapeutic composition prepared by a method according to any other aspect of the invention. In one embodiment, the personalized vaccine or therapeutic composition comprises at least one peptide antigen derived from a protein expressed in an unhealthy cell of the subject, the peptide antigen being capable of being presented on the cell surface by an HLA class I and / or HLA class II molecule of the subject, eliciting a CD4 and / or CD8 T cell response, and comprising a plurality of top-ranked epitopes that share at least one common sequence and, optionally, a cell membrane-penetrating peptide.

[0065] In a further embodiment, the cell membrane-permeable peptide is an immunologically inert cell membrane-permeable peptide.The immunologically inert cell membrane-permeable peptide comprises multiple putative epitopes, and none of these putative epitopes are ranked high in the first list or the second list according to the computer-implemented method of the present invention, so that they are unlikely to induce an immune response in a subject.An example of an immunologically inert cell membrane-permeable peptide is a peptide consisting of at least 8-mer polyarginine.

[0066] In one embodiment, the personalized vaccine or therapeutic composition comprises at least two peptide antigens, both of which contain multiple top-ranked epitopes, at least one antigen that induces a CD4 T cell response and at least one antigen that induces a CD8 T cell response. In a further embodiment, the personalized vaccine or therapeutic composition comprises at least two antigens, at least one antigen that induces both a CD4 T cell response and a CD8 T cell response. In one embodiment, the cell membrane-penetrating peptide is positioned between the at least two antigens.

[0067] Another aspect of the present invention is a method for determining the efficacy of an immune response in a recipient of a personalized vaccine composition, which serves as a safety and efficacy control measure. The method includes the steps of generating all potential epitopes derived from the sequence of a selected antigen; creating a ranked list of these epitopes based on their potency scores; ensuring that epitopes derived from proteins expressed in unhealthy cells of the recipient induce a strong immune response, as indicated by their high potency scores, thereby ensuring efficacy; confirming that epitopes derived from cell membrane-penetrating peptides and / or excipients are immunologically inactive, as indicated by their low potency scores, thereby ensuring safety; and confirming that epitopes with high potency scores are not components of proteins expressed in healthy cells, thereby further contributing to the safety of the personalized vaccine.

[0068] Treatment method In another aspect, there can be a method for treating subject with the above-mentioned vaccine or T cell therapy.This method can include the step of predicting the response of subject to vaccine or T cell therapy, and if the subject is predicted to respond, select the subject for treatment.In another aspect, there is a method for treating subject with vaccine, which can be the above-mentioned personalized vaccine or group vaccine.This method can include the step of predicting the response of subject to vaccine, and if the subject is predicted to respond well to vaccine, select the subject for treatment.

[0069] In a further aspect of the invention, there is provided a method of treating or preventing a disease in a subject, the method comprising administering to the subject a personalised vaccine or therapeutic composition according to any other aspect of the invention.

[0070] In one embodiment, the personalized vaccine or therapeutic composition is administered in combination with an additional therapeutic agent, either simultaneously or sequentially. In a further embodiment, at least two personalized vaccine or therapeutic compositions comprising different peptide antigens comprising a set of top-ranked epitopes are administered to a subject simultaneously or sequentially. In yet another embodiment, the personalized vaccine or therapeutic composition is administered with an adjuvant, either simultaneously or sequentially.

[0071] In one embodiment, the disease is cancer, an autoimmune disease, or a viral infection.

[0072] In another aspect of the invention, there is provided a method of inducing antigen-specific immunity in a subject comprising administering to the subject a personalized vaccine or therapeutic composition according to any of the other aspects of the invention.

[0073] In another aspect of the invention there is provided a personalised vaccine or therapeutic composition according to any of the other aspects of the invention for use in the prevention or treatment of disease.

[0074] Vaccine preparation and manufacturing In another aspect of the present invention, a method for preparing a personalized vaccine or therapeutic composition is provided, the method comprising preparing a first amino acid sequence and preparing a second amino acid sequence, wherein the first amino acid sequence comprises a first peptide antigen that induces a CD4 and / or CD8 T cell response and at least a portion of a cell membrane-penetrating peptide, and the second amino acid sequence comprises a second peptide antigen that induces a CD4 and / or CD8 T cell response and at least a portion of the cell membrane-penetrating peptide, and covalently linking the first and second amino acid sequences to form a personalized vaccine comprising the cell membrane-penetrating peptide located between the first antigen and the second antigen.

[0075] In another aspect of the invention, there is provided a kit for preparing a personalised vaccine or therapeutic composition according to any other aspect of the invention, optionally further comprising a pharmaceutically acceptable excipient and, optionally, instructions for use.

[0076] In another aspect of the present invention, a kit is provided that includes a number of items necessary to prepare a personalized vaccine composition for an individual, the kit including two synthetic peptides, each synthetic peptide comprising at least one selected antigen, and further including at least a portion of a cell membrane-permeable peptide, and the formation of a covalent bond between the two synthetic peptides during preparation of the personalized vaccine to reconstitute the function of the cell membrane-permeable peptide.

[0077] For a better understanding of the present invention and to show how embodiments thereof may be carried into effect, reference will now be made, by way of example, to the accompanying drawings in which: [Brief explanation of the drawings]

[0078] [Figure 1] FIG. 1 is a schematic diagram showing the state of the art demonstrating that HLA alleles transport epitopes to the cell surface, which then trigger T cell receptors (TCRs). [Figure 2a] FIG. 1 is a flow chart showing the key steps for investigating the potency of T cell responses. [Figure 2b] FIG. 1 is a schematic diagram of an algorithm for a computer-implemented method for identifying cell surface antigens that are presented by an HLA molecule of interest and that contain a set of top-ranked overlapping epitopes and a T cell receptor (TCR) elicited by a "core" of overlapping epitopes. [Figure 3] A multi-level directed graph network is shown that helps identify and manage epitope and sub-epitope relationships. [Figure 4] 2b is a flowchart of the scores that can be calculated by the scoring module of the system of FIG. 2a. [Figure 5a] 1 is a flowchart of an example process for selecting cores for, e.g., a vaccine. [Figure 5b] 1 is a flow chart showing a method for using selected cores in a vaccine. [Figure 5c] 10 is a flowchart of an example process for selecting a core, e.g., for treatment. [Figure 6] FIG. 1 is a block diagram of a system for implementing the above method. [Figure 7a-7b] The dataset used for class I HLA is shown. [Figure 7c] The available data sets for class II HLA are shown. [Figure 8a] FIG. 1 shows a schematic diagram of a system incorporating a machine learning model used to validate the described method. [Figure 8b] 8b is a flowchart illustrating a method for training and using the machine learning model of FIG. 8a. [Figure 9a-9b] The ROC curve and PR curve of the current method and the two comparison methods are plotted, respectively. [Figure 9c] 10 is a bar graph comparing the mean rank accuracy of the top 10, top 20, top 50, and top 100 epitopes by the current method and the EL-Max model for each individual averaged across all individuals. [Figure 9d-9f] This compares the accuracy, precision, and recall metrics of the top 20 ranked epitopes between individuals. [Figure 10a] The proposed method (VERDI) is compared with the EL Max model, and the average accuracy ranks of the top 1, top 2, top 3, top 5, and top 10 epitopes across all individuals are plotted. [Figure 10b] The average PU index across all individuals is plotted for the top 1, top 2, top 3, top 5, and top 10 epitopes, comparing the proposed method (VERDI) with the EL Max model. [Figure 10c] The proposed method (VERDI) is compared with the EL Max model, and the number of individuals and accuracy rank in the top three are plotted. [Figure 10d]The proposed method (VERDI) is compared with the EL Max model, and the PU index values ​​and the number of individuals in the top three ranks are plotted. [Figure 11a-11b] Plot of potency against specificity for the top 50 ranked epitopes as 9-mer cores for two de-identified individuals from the adaptive dataset. [Figures 11c-11d] The average potency of the top 50 ranked epitopes is plotted against their abundance (%) in the population. [Figure 12a] Composition of the VERDI vaccine: two cell surface antigens containing a set of overlapping epitopes of interest selected using the methods of the present invention, and optionally an immunologically inert cell membrane-penetrating peptide; a mechanism for inducing CD8 cytotoxic T cells and CD4 helper T cells. [Figure 12b] Comparison of cellular uptake between the VERDI vaccine composition designed using a model antigen and a control vaccine. The figure shows the cellular uptake of the VERDI vaccine and the control vaccine in HeLa cells after 30 and 120 minutes of incubation. The cellular uptake was visualized by fluorescence microscopy. The results clearly show that the VERDI vaccine exhibits significantly increased cellular uptake compared to the control vaccine. The improved intracellular delivery highlights the potential efficacy of the VERDI vaccine in immunization. [Figure 13a] FIG. 1 is a diagram of the structure of the E7 (HPV16) specific VERDI vaccine individualized to the HLA genotype of one of the inventors. [Figure 13b] Graph of quantification of cellular uptake of personalized HPV-specific VERDI vaccine (blue) and control (red) after 30 minutes of incubation with human cells. Intracellular uptake was assessed by flow cytometry, and fluorescence intensity was measured as an indicator of vaccine uptake by cells. Results show significantly higher uptake of VERDI vaccine (blue) compared to control (red), improving its intracellular delivery and indicating its potential efficacy in HPV-specific immunization. [Figure 14a]FIG. 1 is a diagram of the structure of the AKAP-specific VERDI vaccine individualized to the HLA genotype of one of the inventors. [Figure 14b] Graph of quantification of cellular uptake of the control (red) compared to the personalized AKAP-4-specific VERDI vaccine (blue). Intracellular uptake was assessed by flow cytometry, and fluorescence intensity was measured as an indicator of cellular vaccine uptake. Results show significantly higher uptake of the VERDI vaccine (blue) compared to the control (red), improving its intracellular delivery and indicating its potential efficacy in AKAP-specific immunization. [Figure 15a] Diagram of dendritic cell uptake of the control vaccine compared to the VERDI vaccine designed with a model antigen. This figure shows the cellular uptake of the control vaccine compared to the VERDI vaccine in dendritic cells after 120 minutes of incubation. The cellular uptake was visualized by fluorescence microscopy. The results clearly show that the VERDI vaccine has significantly increased cellular uptake compared to the control vaccine. The VERDI vaccine has improved delivery to dendritic cells as well as HeLa cells, suggesting the potential efficacy of the VERDI vaccine in immunization. [Figure 15b] Figure 1 shows a diagnostic test for T cell responses in cancer patients. Transcriptome analysis from a patient's tumor biopsy identified several proteins (CEP55, BIRC5, ATAD2, WT1) that are specifically expressed in tumors. The figure shows the potency (y-axis) of the top-ranked epitopes identified by the method of the present invention (the epitope's position on the protein sequence is shown on the x-axis). Top-ranked epitopes presented on the cell surface by the patient's HLA class I molecules (blue) are involved in CD8 responses. Top-ranked epitopes presented on the cell surface by the patient's HLA class II molecules (orange) are involved in CD4 responses. [Figure 16a]22b is a diagram of cell surface antigen selection in cancer patients using the present method and the T cell response diagnostic test of the present invention. The selected potent antigens are shown in green and are derived from the CEP55, BIRC5, ATAD2, and WT1 proteins (FIG. 22b). These antigens comprise a set of top-ranked epitopes involved in CD8 and CD4 T cell responses in patients. Antigen selection is based on the number and potency of epitopes predicted by the computer-implemented method of the present invention. [Figure 16b] Figure 1 shows cell surface antigen selection for a Murcia patient with metastatic signet ring cell adenocarcinoma using the methods of the present invention and a T cell response diagnostic test. Selected potent antigens useful for personalized VERDI vaccine design are derived from the KKLC1 and SSX4 tumor-specific proteins and are shown in green. This patient's antigens include a set of top-ranked epitopes involved in CD8 and CD4 T cell responses (protein location and epitope potency), shown in blue and orange. Antigen selection is based on the number and potency of epitopes predicted by the computer-implemented methods of the present invention. [Figure 17]This figure shows T cell responses before and after personalized VERDI vaccination in a Murcia patient. A Murcia patient received a single dose of 10 different VERDI vaccines. The QuantiFERON test quantifies the cumulative CD4 and CD8 responses elicited by the 10 different vaccines. Of note, the KKLC1-specific VERDI vaccines (C1 and C2, serving as positive controls) consistently induced strong T cell responses after a single dose. In stark contrast, the SSX4-specific VERDI vaccines (C4 and C5, serving as negative controls) did not induce any detectable responses. This evidence supports the following compelling hypothesis: the KKLC1 antigen identified by our computer-implemented method induced an immune response against the KKLC1-specific protein expressed in the patient's tumor cells, whereas the lack of a detectable response to the SSX4-specific vaccine suggests that the patient's tumor did not express this specific target. Transcriptome analysis of tumor biopsies taken after vaccination confirmed the absence of tumor cells expressing any of the proteins targeted by the personalized VERDI vaccine. [Figure 18] Graph of blood biomarker response before and after personalized VERDI vaccination in Murcia patients. High alkaline phosphatase levels are common in patients with bone metastases. High carbohydrate antigen 19-9 (CA19-9) levels also predict clinicopathological status, recurrence, and prognosis in gastric cancer. [Figure 19a] Figure 1 shows the patient journey during personalized VERDI vaccination. [Figure 19b] FIG. 1 is a diagram of predictive diagnostics and vaccine design software. [Figure 20a] FIG. 1 is a diagram of the tumor transcriptome analysis software feeding data into the diagnostic software. [Figure 20b] FIG. 1 is a diagram of a secure web-based application for downloading predictive diagnostic and vaccine design results for cancer patients and uploading clinical data. DETAILED DESCRIPTION OF THE INVENTION

[0079] Detailed Description of Drawings As described below, we rank all putative epitopes based on their predicted density on the cell surface (e.g., tumor cells or infected cells, e.g., including SARS-CoV-2 infection) and then investigate the potency of T cell responses to proteins expressed on the cells. Using a computer-implemented method, we determine the potency of all viral epitopes presented by multiple HLA alleles encoding six HLA class I alleles and / or 12 class II molecules. As described below, this method leverages large amounts of experimental data to assess epitope density on an individual's cell surface and rank epitopes that elicit a T cell response. The analytical method described below is referred to as VERDI (Viral Epitopes Ranked by Digital Intelligence). As demonstrated, this method is rapid and accurate, providing a complete profiling of an individual's T cell response to one or more proteins. Therefore, this method is applicable to the development of a range of novel pharmaceuticals, including: These include vaccines that induce strong T-cell responses in 12 HLA allele-matched recipients, precision vaccines that induce T-cell responses in selected subpopulations, companion diagnostic tests for vaccines to predict the efficacy of vaccine-induced T-cell responses in recipients, T-cell therapies that target high-density epitopes with the same "core," and tests to diagnose an individual's HLA genotype predisposition to diseases / illnesses (including cancer, infectious diseases, and autoimmune diseases, such as COVID-19).

[0080] Predictive diagnosis of antigen-specific T cell responses using VERDI FIG. 2a is a flowchart outlining a method used for predictive diagnosis of antigen-specific T cell responses. As shown, there are two independent inputs: HLA genotype data of the test subject obtained in step S210, and input epitope data obtained in step S212. As shown, the data may be obtained simultaneously, but it is understood that the data may be obtained sequentially in any order. The HLA genotype data may be obtained using any suitable technique, including common HLA genotype diagnostic tests used in transplantation. The HLA genotype data may include four-digit HLA class I genotype data representing the amino acid sequence of the peptide-binding pocket of the HLA molecule and / or four-digit HLA class II genotype data representing the amino acid sequence of the peptide-binding pocket of the HLA molecule. The complete set of HLA genotype data is shown in Table 1 below, collected along with the identity of the individual from whom it was collected and, if applicable, their region: [Table 1]

[0081] Because matching of HLA alleles between transplant recipients and donors is essential for clinical outcomes, four-digit HLA class I data are routinely obtained from blood or oral swabs in clinical settings, diagnostic laboratories, and stem cell donor registries. Currently available methods are reviewed in "Bioinformatics Strategies, Challenges, and Opportunities for Next Generation Sequencing-Based HLA Genotyping" by Klasberg et al., published in Transfusion Medicine and Hemotherapy (2019). Data can be obtained from individuals for purposes such as classifying T cell responses for personalized vaccine development, predicting the efficacy of vaccine-induced T cell responses to match existing vaccines to individual populations, prioritizing vaccination, matching recipients for TCR-T cell therapy, and determining predisposition to specific diseases.

[0082] The epitope is determined by the disease being studied. For example, when studying T cell responses to SARS-CoV-2, the epitope could be one of 70,000 epitopes derived from viral proteins expressed in infected cells. Each epitope is input in step S212 along with its sub-epitopes, which are defined in detail in Figure 3. For example, when studying HLA class I data, the input epitope could be an 8-14 amino acid sequence, and for a 14 amino acid epitope, its sub-epitope could be an 8-13 amino acid sequence.

[0083] Separate inputs are entered as pairs of epitope and HLA data, and at least one score is generated for each pair in step S214. For example, when considering HLA class I molecules, six scores may be calculated as follows: Each score may be derived from an EL score generated by NetMCHPan4.1 or other appropriate method. As described in the Background Art section and shown in Figure 1, HLA transports epitopes derived from proteins to the cell surface and induces T cell responses against cells expressing the protein. In Background Art methods, each HLA is typically considered alone. Figure 2b illustrates the biological mechanism represented by the computer-implemented method of Figure 2a.

[0084] Figure 2b, like Figure 1, illustrates the presence of antigen-presenting cells and T cells. However, instead of considering a single HLA molecule, Figure 2b illustrates the consideration of multiple HLA molecules. For ease of reference, one each of HLA-A, HLA-B, and HLA-C class I alleles is shown, each presenting epitopes 150a, 150b, and 150c. These overlapping epitopes share a common core and stimulate the same TCR. The TCR core is shown in the inset, detailing the binding of the TCR core (red) to an individual's six HLA class I molecules. An individual's T cells respond sequentially by repeatedly stimulating the TCR with the core. Because core-TCR binding is far more important for stimulating the TCR than HLA-TCR binding, the TCR can also be stimulated by the core presented by either HLA molecule (class I or class II). As described in detail below, current methods and systems identify epitopes of different lengths transported to the cell surface by the individual's HLA. Overlapping epitopes with the same core are identified, and from these overlapping epitopes, T cell antigens for inclusion in vaccines and targeting in T cell therapy are identified. While only HLA class I molecules are shown in Figure 2b, it will be understood that this method is also applicable to identifying CD8 and CD4 T cell antigens and their cores.

[0085] Returning to FIG. 2a, steps S212 and S214 are repeated for multiple input epitopes. The at least one score generated for each pair of epitope and individual in step S214 is used to rank the multiple input epitopes in step S216 to generate a ranked list. The ranked list may include only the top epitopes, e.g., the top 50 or top 100 epitopes. The ranked list defines an individual's T cell antigen repertoire and can be used to predict the strength of an individual's T cell response to a particular disease, such as SARS-CoV-2 or cancer. In step S218, a ranking is output that includes the specificity of each epitope (i.e., the amino acid sequence within the epitope) and the potency of each epitope (e.g., based on at least one score generated).

[0086] Figure 3 shows a multi-level directed graph network that can be used to identify and manage epitope and subepitope relationships. Each node (e.g., 110, 112, 114) identifies a peptide, and each peptide is classified into different levels based on its amino acid length. Any node that is not at the lowest level is considered a parent node, and each parent node has a left child node and a right child node, each of which is one amino acid shorter than the parent node. The left child omits the prefix (i.e., the first amino acid) of the parent peptide. The right child omits the suffix (i.e., the last amino acid) of the parent peptide. Parent-child relationships are defined by edges where the source is the parent and the sink is the child.

[0087] By way of example only, parent node 110 may be input epitope EKMKKDFRAMKDLAQQINLS (SEQ ID NO: 1111), a CD4 epitope. The left child 112 of this parent node is EKMKKDFRAMKDLAQQINL (SEQ ID NO: 1112), and the right child 114 is KMKKDFRAMKDLAQQINLS (SEQ ID NO: 113). Each child is also a node and has a left child and a right child. Because epitope EKMKKDFRAMKDLAQQINLS (SEQ ID NO: 1114) is a 20-mer peptide, and the minimum length for CD4 in the method implemented in FIG. 2a is currently 11, the graph of all subepitopes of this input epitope will be 10 levels deep, including the parent node. The top level has the longest amino acid length (20), and the bottom level has the shortest amino acid length (11). The bottom levels of the graphs for CD8 and CD4 epitopes are different. The minimum length of each CD4 subepitope is 11, and the minimum length of a CD8 subepitope is currently 8.

[0088] Calculating the score Figure 4 is a flowchart illustrating the scores that can be calculated using the method of Figure 2a. Each input epitope is paired with an HLA allele (class I or class II separately). For each pair, a trafficking score is calculated (step S400). The purpose of the trafficking score is to identify the likelihood that the epitope will be presented on the cell surface by the HLA allele (i.e., the likelihood that it will be transported by the HLA) or the amount of the epitope that the HLA allele is most likely to transport to the cell surface. Both types of HLA alleles transport and present peptides from degraded proteins on the cell surface, but proteins are degraded in different ways within the cell. In other words, the trafficking score can be defined as the contribution of the HLA allele to presenting the epitope on the cell surface.

[0089] The transport score may be the Eluted Ligand Score (EL-score) predicted by the scoring mechanism taught by Reynisson et al. in "NetMHCpan-4.1 and NetMHCIIpan-4.0: Improved Predictions of MHC Antigen Presentation by Concurrent Motif Deconvolution and Integration of MS MHC Eluted Ligand Data," published in Nucleic Acids Research in 2020. This system utilizes the largest set of experimental data (including data derived from eluted ligands (EL)) that identifies the ability of cell surface epitopes to transmit signals from HLA to T cell receptors (TCRs). The EL score may be considered an indicator of transport efficiency, proportional to cell surface epitope density. The EL score may also be considered the probability or likelihood of epitope transport. The value of the EL score is between 0 and 1. Optionally, the identity of each epitope and HLA allele may be stored along with the associated EL score; such a dataset may be referred to as a ligand dataset.

[0090] If all HLA class I alleles are used, six transport scores (or EL scores—these terms can be used interchangeably) are calculated for each epitope in step S400. Similarly, if all HLA class II alleles are used, 12 transport scores are calculated for each epitope in step S400. Thus, in step S402, the number of scores input into the feature vector can be reduced by summing at least some of the transport scores. The summed transport scores can be referred to as epitope weight scores (EWS). The summation can be performed for each locus in the HLA genotype data. That is, for HLA class I data, there can be a first epitope weight score for the HLA-A locus, a second epitope weight score for the HLA-B locus, and a third epitope weight score for the HLA-C locus. Similarly, for HLA class II data, there may be a first epitope weight score for the HLA-DP conformation, a second epitope weight score for the HLA-DQ conformation, and a third epitope weight score for the HLA-DR conformation.

[0091] When calculating each EWS, it is possible to select only contributing pairs to reduce computational complexity. Such selection may also exclude HLA alleles that do not bind to the epitope. Contributing HLA-epitope pairs are defined as those whose weighted EL score exceeds an EL threshold (ELT), and therefore only weighted EL scores that exceed the EL threshold are summed. Simply by way of example, setting the EL threshold to 0.2 will typically remove low-scoring HLA-epitope pairs that are unlikely to be transferred to the cell surface. Each EWS can be expressed as:

number

[0092] It will be understood that other total scores can be generated. For example, a first EWS sums the EL scores of all HLA class I molecules and can therefore range from 0 to 6. Such an EWS can be considered an indication of the likelihood that an epitope will be processed, transported, and presented to T cells by each of the six HLA class I molecules. Similarly, a second EWS sums all EL scores for HLA class II molecules and can range from 0 to 12. Such a second EWS can be considered an indication of the likelihood that an epitope will be processed, transported, and presented to T cells by each of the 12 HLA class II molecules. Any or all EWS values ​​can be used to rank epitopes as desired, with higher-ranked epitopes being more likely to elicit a T cell response.

[0093] By way of example only, the EWS scores of some epitopes derived from proteins expressed on tumor cells that are presented by HLA class I and II molecules in patients with metastatic breast cancer are shown in the table below. The epitopes are ranked by their EWS score. [Table 2]

[0094] High-ranking epitopes presented by HLA class I molecules are more likely to induce cytotoxic CD8 T cell responses, whereas high-ranking epitopes presented by HLA class II molecules are more likely to induce CD4 T cell responses.

[0095] If the peptide length is longer than the minimum number (e.g., at least 9 when considering HLA class I alleles, or at least 12 when considering HLA class II alleles), additional scores can be calculated. These scores are called the left subitope score and the right subitope score, and are calculated in steps S406 and S408, respectively. The left subitope score, SubitopeWeightScore_1 (SWS1), is the sum of the epitope weight scores of the epitopes in the left subgraph (LEWS). The right subitope score, SubitopeWeightScore_2 (SWS2), is the sum of the epitope weight scores of the epitopes in the right subgraph (REWS), excluding epitopes that are also part of the left subgraph (LEWS).

[0096] Returning to Figure 3, it can be seen that the left subgraph contains all subepitopes that are one amino acid removed from the right side of the top peptide. Similarly, the right subgraph contains all subepitopes that are one amino acid removed from the left side of the top peptide. It can be seen that there is overlap between the right and left subgraphs. Thus, the right subepitope score (SWS2) excludes subepitopes already included in the left subepitope score (SWS1). The subepitopes included in each score are shown in Figure 3.

[0097] 4, for each individual, SubitopeWeightScore_1 (SWS1) is calculated in step S404, and SubitopeWeightScore_2 (SWS2) is calculated in step S406. These calculations may be defined as follows:

number

[0098] In step S408, there is also the option to calculate a score named "OverallWeightScore" (OWS). For the shortest peptide, the overall weight score can be calculated as follows:

number

[0099] For each peptide of length greater than a minimum length (e.g., 9 or greater when considering HLA class I alleles, or 12 or greater when considering HLA class II alleles), the OverallWeightScore (OWS) is the sum of each EpitopeWeightScore, SubitopeWeightScore_1, and SubitopeWeightScore_2, i.e.:

number

[0100] Optionally, only EWS scores above an EWS threshold may be included when generating this overall score. If a threshold is used, OWS may be calculated as follows:

number

[0101] Among the overlapping, potent epitopes in this cancer patient, the core is the 12 amino acid long KDFRAMKDLAQQ (SEQ ID NO: 40), which contains the TCR-triggering pattern described above in connection with Figure 2b.

[0102] Optionally, other scores may be calculated in step S410. For example, a maximum EL score, a minimum EL score, and / or an average EL score may be determined. The scores calculated in any of steps S402 through S408 are output in step S412 to characterize the potency of each epitope to elicit a T cell response, as well as the potency of a set of top-ranked epitopes from the viral proteome to elicit a T cell response in the individual. As described above in connection with FIG. 2a, some or all of these scores are output as features in a feature vector, which are then processed by a machine learning module to determine the probability that the epitope is an immunogenic epitope for that individual.

[0103] Core Selection Figure 5a illustrates one method for selecting core epitopes using the scores calculated above. In the first step, S500, a ranked list of epitopes is obtained, for example, using the method of Figure 2a. For illustrative purposes, the following table shows an example of a ranked list for an individual associated with the SP17 protein, a known antigen in metastatic breast cancer patients. All 16 examples shown are CD4 epitopes, and therefore, the EWS scores for HLA class II alleles are shown, along with the SWS1 scores for both class I and class II alleles and the SWS2 scores for both class I and class II alleles. Several epitopes of this small SP17 protein have high EWS scores for HLA class II alleles and are therefore likely to elicit strong CD4+ T cell responses in metastatic breast cancer patients. By including the SWS1 and SWS2 scores for class I alleles, we confirm whether the sequence of these epitopes contains favorable CD8 epitopes (sub-epitopes). [Table 4]

[0104] In the next step S502, at least two top-ranked epitopes are selected, for example, each of epitopes 1-16 listed above. In step S504, subepitopes of each selected epitope are identified, for example, using the directed graph shown in FIG. 3. In step S506, the subepitopes are compared to determine whether there is at least one common subepitope, for example, by performing an intersection operation on the identified subepitopes. If there is no common subepitope, the method returns to step S502, where a higher-ranked epitope may be selected.

[0105] For example, using the above table, the first listed epitope is selected along with epitopes ID Nos. 3, 5, 7, 8, 11, 14, 15, and 16. Each of these epitopes has a 13-amino acid long common subepitope (PAFAAAYFESLLE (SEQ ID NO: 57)). The next step, S508, determines whether the common subepitope is ranked in the list. For example, because the common subepitope (PAFAAAYFESLLE (SEQ ID NO: 57)) is not individually listed among the top-ranked epitopes presented by self-HLA molecules, it is less likely to induce a T cell response than higher-ranked epitopes. If not, the method may return to step S502, select a higher-ranked epitope, and begin the process again.

[0106] It will be appreciated that if many epitopes are selected in step S502, a negative result in step S506 is likely to be avoided. For example, if the top 30 epitopes are selected in step S502, it is statistically unlikely that there will be no common sub-epitopes. Similarly, a negative result in step S508 is likely to be avoided if the top-ranked list contains a sufficient number of epitopes (e.g., if there are 200 epitopes in the ranked list). That is, the top 30 epitopes can be selected in step S502, and common sub-epitopes can be found among the top 200 epitopes in step S508.

[0107] In step S510, at least one core is identified from the common subepitopes. Any suitable mechanism can be used for this determination. For example, the ranks of all common subepitopes can be considered, and the highest-ranked common subepitope can be identified as the core. For example, NIPAFAAAYFESLLE (SEQ ID NO: 58) is a common subepitope of epitopes ID NOs. 3, 7, and 15, and is listed individually as number 11. Therefore, this common subepitope has the highest rank and may be output as the core. Alternatively, the potency score of each common subepitope can be compared to a threshold, and a common subepitope with a potency score exceeding the threshold can be identified as the core. Alternatively, the EWS score and SWS2 score can be compared to a threshold, and if one or both scores exceed the threshold, the common subepitope can be output as the core. Note that the core may be longer than the smallest core common to many of the top-ranked epitopes and contain multiple overlapping sequences. When using the method of FIG. 5a, the cores selected will typically be longer than the shortest sequence common to most examples.

[0108] 5a illustrates a method for selecting a core based on a ranked list of epitopes, with the table used in the example showing scores associated with HLA class II alleles. Once an appropriate core has been selected for a particular type of HLA data, the method can be repeated as needed for other types of HLA data to identify at least one acceptable core for the other HLA data in step S512.

[0109] The cores identified in step S510 may also be identified in step S512. That is, there may be cores that have the potential to induce both strong CD8+ T cell responses and strong CD4+ T cell responses. An example is shown in the following data. [Table 5]

[0110] Each epitope with the shortest common subepitope, PAFAAAYFESLLE (SEQ ID NO: 57), also has a high SWS1 score ranging from 1.7 to 2.1. More specifically, the core NIPAFAAAYFESLLE (SEQ ID NO: 58) also ranks highly in the table above. This suggests that this antigen is likely to induce a strong CD8+ T cell response. This core may be output in step S516 as suitable for eliciting both CD4+ and CD8+ T cell responses. It will be appreciated that two different cores, one for HLA class I molecules and one for HLA class II molecules, are likely to be output, as shown in step S518.

[0111] Vaccine development An important application of the top-ranked epitopes is to select T cell antigens for inclusion in a vaccine to induce a T cell response in an individual. Personalized vaccines typically contain multiple T cell antigens derived from one or more proteins expressed on infected cells (e.g., T cell antigens derived from the SARS-CoV-2 proteome). The inclusion of multiple T cell antigens may increase the probability of killing diseased cells (tumor cells or virus-infected cells). Preferably, the vaccine contains at least two T cell antigens to induce both CD8+ and CD4+ T cell responses against proteins expressed on diseased cells. Figure 5b illustrates several steps that can be taken when developing a vaccine.

[0112] As shown in step S530, for example, using the method of FIG. 5a, an output core is obtained. NIPAFAAAYFESLLE (SEQ ID NO: 58) is a rare T cell antigen that appears to be capable of inducing very strong CD4+ and CD8+ T cell responses in this individual by triggering the same TCR, although the possibility that the epitope may trigger a different TCR cannot be excluded. Therefore, this core is used to identify suitable antigens as an optimal personalized vaccine for this individual. Therefore, in step S532, the sequence of this core may be output as a potential vaccine for this individual.

[0113] As an example using two different antigens, a vaccine targeting the EPCAM protein for a metastatic breast cancer patient may be formed by combining two output cores: YVDEKAPEF (SEQ ID NO: 59) (Class I EWS = 2.9) and DVAYYFEKDVKGESL (SEQ ID NO: 60) (Class II EWS = 2.0) to form a potent vaccine with a total score (i.e., the sum of the first and second EWS) of 4.9. Such a potent T cell vaccine is likely to kill the patient's cells expressing the EPCAM protein. Such a vaccine is personalized to the patient. The combined sequence may be output in step S532. It should be understood that two independent output cores may overlap or be so close to each other that they fit into a single long peptide, and the sequence of this long peptide may be output as a potential vaccine.

[0114] As another example, the following ranked epitopes could be used to develop a hepatitis B vaccine for the following three individuals: [Table 6] TIFF2025540913000018.tif238159TIFF2025540913000019.tif240159TIFF2025540913000020.tif241159TIFF2025540913000021.tif165159

[0115] For subject 10881, examples of overlapping epitopes shown in the table above, "Top Epitopes by EWS in Three Hepatitis B Patients," include the following: AAYPAVSTF (SEQ ID NO: 261) (EWS = 3.74), KAAYPAVSTF (SEQ ID NO: 273) (EWS = 2.09), RKAAYPAVSTF (SEQ ID NO: 274) (EWS = 2.03), and AYPAVSTF (SEQ ID NO: 278) (EWS = 1.82), each of which shares the common subepitope AYPAVSTF (SEQ ID NO: 278). Of the four overlapping epitopes, RKAAYPAVSTF (SEQ ID NO: 274) was selected as a T cell antigen to elicit a CD8+ T cell response because it contains the sequences of all four epitopes. When the RKAAYPAVSTF (SEQ ID NO: 274) antigen is introduced into cells, it is processed into smaller peptides to be presented on the cell surface of 10881. The potency of the T cell response induced by the RKAAYPAVSTF (SEQ ID NO: 274) antigen can be estimated by the sum of the EWS of the four epitopes. The "core" of the four epitopes is AYPAVSTF (SEQ ID NO: 278), which triggers the same TCR and induces a strong T cell response.

[0116] In subject 3821, only two epitopes, AAYPAVSTFEK (SEQ ID NO: 215) (EWS=0.80) and AAYPAVSTF (SEQ ID NO: 173) (EWS=1.32), share the common sub-epitope AYPAVSTF (SEQ ID NO: 278). From these overlapping epitopes, the T cell antigen selected for vaccine development may be AAYPAVSTFEK (SEQ ID NO: 215). Not only is the T cell antigen of 3821 different from that of 10881, but the sum of their EWS suggests that it will induce a much weaker T cell response in recipients. Subject 3819 has two top epitopes with a common subepitope (AYPAVSTF): AAYPAVSTFEK (SEQ ID NO: 107) (EWS = 0.79) and AAYPAVSTF (SEQ ID NO: 91) (EWS = 0.88). The selected T cell antigen is AAYPAVSTFEK (SEQ ID NO: 107), the same as subject 3821. However, the sum of their EWS suggests a lower density of "core" epitopes on the cell surface compared to subject 10881. Subject 3819 has a very low density of HBV-derived epitopes, and the EWS suggests that HBV induces a weak T cell response. The top epitope is HLYSHPIIL (SEQ ID NO: 61) (EWS=1.65), with no overlapping epitopes among the top 100 epitopes. The T cell antigens for subject 3821 are likely to be AAYPAVSTFEK (SEQ ID NO: 107) and HLYSHPIIL (SEQ ID NO: 61), which appear to have similarly weak potency in stimulating T cells. These two T cell antigens would be candidates for the development of an HBV vaccine for 3821.

[0117] In the above example, for each patient, AAYPAVSTFEK (SEQ ID NO: 107) is an antigen that can be included in a personalized vaccine. It will be appreciated that this step can be repeated for multiple individuals, and simple statistical techniques can be used to identify T cell antigens presented by diseased cells in most individuals with a specified disease (e.g., COVID, Hepatitis B, or cancer). These repeatedly reported antigens can be included in a universal vaccine appropriate for the population. For example, as shown in Figure 5b, the output (cores or longer sequences containing at least one core) can be analyzed to determine whether they are suitable for multiple individuals in step S534. If the output is acceptable, it can be proposed as a component of a universal vaccine, as shown in step S536. If the output is unacceptable, no universal vaccine is proposed.

[0118] Predicting patient response to vaccines The efficacy of a vaccine in a subject can be determined by the methods of the present invention, for example, as shown in step S538 and reported in a diagnostic test. By way of example only, efficacy can be determined as the average EWS of the vaccine's top-ranking epitopes. Responder selection involves obtaining a 4- to 8-digit HLA genotype of the patient and confirming that the protein targeted by the vaccine is expressed in the patient's unhealthy cells. For example, the target protein could be the spike of a circulating viral variant or AKAP4, which is expressed in breast cancer cells but not in healthy cells.

[0119] Vaccines that induce strong T cell responses that rapidly kill unhealthy cells (e.g., SARS-CoV-2-infected cells or tumor cells) are useful for the prevention and treatment of diseases (e.g., COVID-19 or cancer). Therefore, the VERDI method can also be used to select patients who exhibit strong T cell responses to a vaccine, for example, by comparing the efficacy with a threshold, as shown in step S540. In other words, the VERDI method is used to determine whether any of the putative epitopes in the vaccine are presented at high density on an individual's unhealthy cells. If there are highly ranked epitopes shared with the vaccine among the epitopes presented on the individual's unhealthy cells, the individual is more likely to respond well to the vaccine. In other words, the more highly ranked epitopes present in both the circulating virus and the vaccine, the higher the success rate. Therefore, if the efficacy exceeds a threshold, as shown in step S542, vaccination may be recommended for the individual. However, if the efficacy is below the threshold in step S544, vaccination may not be recommended.

[0120] In the prior art, epitope selection methods based on HLA binding have been used in vaccine development. Epitopes in personalized vaccines created for breast cancer patients were selected to bind to at least three HLAs of the subject, as described in U.S. Patent Application Publication No. 2018 / 264094. As an example of predicting the efficacy of vaccine-induced T cell responses, the following table shows the efficacy of eight long peptide vaccines used to immunize metastatic breast cancer patients. The efficacy of T cell responses against tumor antigens was predicted using the above system (using HLA class I EWS, HLA class II EWS, and total EWS). After the first immunization, the efficacy of T cell responses against vaccine peptides was measured by ELISPOT assay after 5 days of culture of peripheral mononuclear cells isolated from patient blood samples (spots / 10 6 PBMCs). [Table 7]

[0121] The results show that the HLA class I and class II EWS were 1.8 and 1.9, respectively, and only one peptide (SVYADQVNIDYLMNRPQNLR (SEQ ID NO: 362)) had acceptable potency. As expected, this peptide induced a strong T cell response capable of attacking tumor cells (366 spots / million cells). Most other peptides induced weak T cell responses (less than 100 spots / million cells). Only two peptides could induce relevant T cell responses (100–300 spots / million cells), likely due to their acceptable HLA class II EWS (2.3, 1.4). However, these peptides have low HLA class I EWS and may not induce CD8 cytotoxic T cell responses capable of killing tumor cells. Although the sum of HLA class I and class II EWS may not be linear as shown in the example above, a CD4 T cell response is required to induce a strong CD8 T cell response capable of killing tumor cells. The results indicated that the predicted efficacy of peptides included in personalized vaccines for breast cancer patients was lower than that of the top-ranked epitopes identified above. The results also indicate that selection of poor T cell antigen selection in vaccine development results in weak T cell responses.

[0122] Development of T cell therapy FIG. 5c illustrates a method for selecting suitable cores as targets for T cell therapy. Because there is some overlap with the method of FIG. 5a, a first step S600 obtains a ranked list of epitopes, e.g., using the method of FIG. 2a. In step S602, at least two highly ranked epitopes are selected, and in step S604, subepitopes of the selected epitopes are identified, e.g., using the directed graph shown in FIG. 3. In step S606, the subepitopes are compared to determine whether at least one common subepitope exists, e.g., by performing an intersection of the identified subepitopes. If no common subepitope exists, the method returns to step S602 to select a higher-ranked epitope.

[0123] In step S610, at least one core is identified from the common sub-epitopes. When developing a T cell therapy, the core is the target of the therapy, so one method for selecting a core is to identify the shortest overlapping sequence common to multiple epitopes. The selected core may be the most frequent core among the selected top-ranked epitopes. Once a suitable core for a particular type of HLA data is selected, the method can be repeated as necessary for other types of HLA data to identify at least one acceptable core for the other HLA data in step S612. If the cores identified in step S614 are the same (which is unlikely), a single core is output in step S616. Otherwise, two cores are output as targets in step S616.

[0124] Returning to the example of hepatitis B, in all three subjects, AYPAVSTF (SEQ ID NO: 278) is a dense "core" on HBV-infected cells, which induces one TCR on T cells. This highly specific and dense "core" among subjects is a specific example of one target for TCR-based T cell therapy. This dense target is considered suitable for killing HBV-infected cells in the three subjects exemplified here.

[0125] Over the past decade, T cell therapy has emerged as an important novel therapeutic agent for the treatment of tumors and infectious diseases. TCR-based T cell therapy utilizes engineered T cells that recognize 9-11 amino acid-long epitopes presented by HLA molecules. Because the epitopes are derived from intracellularly expressed proteins, TCR-based T cell therapy differs from traditional antibody-based or chimeric antigen receptor (CAR) therapies, which can only bind to cell surface-expressed proteins. Until now, developers have focused on targeting HLA-restricted epitopes, but this has resulted in limited therapeutic efficacy accompanied by undesirable toxicities due to off-target recognition, which remains a major limitation of TCR-based T cell therapy. Our computer-implemented method has the potential to revolutionize target identification for TCR-based T cell therapy by identifying high-density "cores" on the cell surface that are recognized by TCRs. "Cores" are HLA-unrestricted targets, making them suitable for "core"-specific TCR-based T cell therapy. Potential responders to such therapies can be identified using our computer-implemented method by calculating the density of target "cores" on the surface of patients' diseased cells.

[0126] Deciding on a treatment plan As shown above, the HBV-specific T cell antigen repertoires of subjects 3819, 3821, and 10881 differ not only in specificity (epitope sequence) but also in total EWS, suggesting different strengths of T cell responses. Ranking data for human papillomavirus and SARS-CoV-2 in the same individuals are also shown. The ranking data can be used in antigen selection for vaccine or T cell therapy development in a similar manner as described above.

[0127] The results below show the top-ranked HPV- and SARS-CoV-2-specific epitopes presented by HLA class I alleles in three subjects. [Table 8] TIFF2025540913000024.tif239159TIFF2025540913000025.tif235159TIFF2025540913 000026.tif234159TIFF2025540913000027.tif237159TIFF2025540913000028.tif15159

[0128] The full SARS-CoV-2 dataset is detailed in the table below. [Table 9] TIFF2025540913000030.tif235159TIFF2025540913000031.tif241159TIFF2025540913000032.tif24115 9TIFF2025540913000033.tif238159TIFF2025540913000034.tif240159TIFF2025540913000035.tif50159

[0129] These ranked lists can be used to develop vaccines or T cell therapies against HPV infection, as described above. This method involves identifying the seven amino acid peptide YLHPSYY (SEQ ID NO: 1084) as a dense core, which is present in the top-ranked epitopes of YLHPSYYML (SEQ ID NO: 370) for all three subjects. However, subject 10881 has four strong epitopes YLHPSYYML (SEQ ID NO: 370), FYLHPSYYM (SEQ ID NO: 582), FYLHPSYY (SEQ ID NO: 638), and FYLHPSYYML (SEQ ID NO: 645) (total EWS = 1.87 + 1.70 + 1.03 + 1.00 = 5.60), while subject 3819 has a significantly weaker total EWS of 2.09 (1.37 + 0.72) and subject 3821 has a significantly weaker total EWS of 1.81. These data suggest that the 10-amino acid HPV vaccine antigen FYLHPSYYML (SEQ ID NO: 645) induces CD8+ T cell responses in all three subjects, but its potency is strongest in subject 10881 and weakest in subject 3821. The YLHPSYY (SEQ ID NO: 1084) core would be an excellent target for TCR T cell therapy to kill HPV-infected cells in these subjects. Similarly, for SARS-CoV-2, the 11-amino acid peptide SVYYTSNPTTF (SEQ ID NO: 963) can be selected as a vaccine antigen to induce CD8+ T cell responses in all three subjects, and the 9-amino acid core YYTSNPTTF (SEQ ID NO: 680) can be selected as a target for T cell therapy.

[0130] Determining the strength of the T cell response Interestingly, the EWS values ​​of the top-ranked epitopes were in a similar range, regardless of their location within the proteome. The mean potency of the top-ranked epitopes can be proposed as a new definition of the strength of an individual's T cell response. The strength of the T cell response can be estimated from the mean potency of the top-ranked epitopes, e.g., the mean EWS of the top 100 epitopes. For example, in the case of HBV infection, the virus may induce a weak T cell response in subject 3819 (mean EWS = 0.81) and a strong T cell response in subject 10881 (mean EWS = 1.32). As another example, in the SARS-CoV-2 data mentioned above, subject 3819 had a mean EWS of 1.1, indicating a weak T cell response ranging from 0.95 to 1.90. Subject 3821 exhibited an intermediate T cell response with a mean EWS of 1.62, ranging from 1.31 to 2.56. The mean EWS for subject 10881 was 1.95, ranging from 1.88 to 3.81. In other words, subject 10881 demonstrated the strongest T cell response based on the mean EWS for the top 100 epitopes. These results cannot be directly confirmed experimentally because the limited blood volume required for multiple tests makes it impossible to measure the efficacy of all potential epitopes of the virus in each individual subject.

[0131] It should be noted that the relative strength of T cell responses to different viruses is similar across individuals. As noted above, the same individual produced weak, intermediate, and strong T cell responses to hepatitis B, human papillomavirus, and SARS-CoV-2. For example, HBV, HPV, and SARS-CoV-2 elicited strong T cell responses in subject 10881 and weak T cell responses in subject 3819. This data suggests that a subject's HLA genotype determines the strength of T cell responses in viral infections and tumors, among other diseases that are associated with their clinical responses. The strength of T cell responses to viral infections is estimated by cell surface epitope density, a phenotype determined by an individual's HLA genotype. Subjects with strong T cell responses rapidly kill diseased cells, resulting in mild or no disease. On the other hand, subjects with weak T cell responses may develop severe or chronic disease due to their reduced ability to kill diseased cells. The computer-implemented methods according to the present invention are useful for predicting the strength of an individual's T cell response and thereby predicting the clinical outcome of an infectious or malignant disease. These predictions may influence the preventative measures that a physician or patient should take to reduce the risk of severe or chronic disease.

[0132] The results further demonstrate that T cell responses are unique to each individual due to the extreme interindividual diversity of HLA genotypes. The computer-implemented method described herein is the first to prioritize individual epitopes based on cell surface density estimates, making it an essential step for identifying T cell antigens in the development of novel therapeutics targeting high-density cell surface molecules. For example, high-density epitopes are specific and abundant targets for T cells present on the surface of infected and tumor cells.

[0133] System Example 6 shows a schematic diagram of a system 10 for implementing the method described above. There are standard components of a hardware solution implemented, such as a user interface 20, a processor 30, a memory 40, and an interface 42 for connecting to an optional external database 50. It will be understood that the system may also include other standard components that are not shown for ease of explanation.

[0134] The user interface 20 includes an input device 22, such as a touchscreen, keyboard, mouse, voice recognition input device, or other suitable device. A user 70 can use the input device 22 to input either or both of the subject's HLA genotype data and / or data on proteins expressed in the subject's diseased cells. The user interface 20 may also include an output device 24, such as a suitable display device, for displaying information to the user 70. The output device 24 may display output of the computer-implemented methods described above. For example, the output may be a ranked list of the subject's individual T cell antigens, optionally including T cell antigen potency values. If appropriate, the output may be a personalized vaccine or treatment plan recommendation for the individual.

[0135] The processor 30 processes the received inputs to generate desired outputs, as described above. For example, certain processing is performed in dedicated modules, such as a scoring module 32 that generates scores as described above and a ranking module 36 that ranks the scores. The scoring and ranking modules can use any suitable technology. For example, the ranking module may include an AI core that implements a diagnostic algorithm that predicts the efficacy of a T cell response based on subject data. The modules, particularly the scoring module 32, can access a database 42 for scores associated with specific T cell antigens and cores. This database can be stored locally in the memory 40. Alternatively, the data can be stored in an external database 50, and the machine learning module 32 can communicate with such external database 52.

[0136] The processor 30 also includes a management module 34. Such a management module may be included to manage workflow and track subject data (e.g., to ensure CDISC-compliant data structure and storage). The management module 34 may be used to manage vaccine manufacturing and delivery. The management module 34 may be configured to generate subject-specific reports (e.g., T cell antigen diagnosis, T cell response efficacy). The management module 34 may therefore also communicate with an internal database 42 and / or an external database 50. It should be understood that in both internal and external databases, data may be split across multiple databases and stored in suitable storage devices (e.g., the cloud in the case of external databases). The external database may be located remotely (i.e., at a different location) from the system.

[0137] As used herein, terms such as “component,” “module,” “processor,” or “unit” include, but are not limited to, hardware devices that perform particular tasks or provide related functionality (e.g., circuitry in the form of discrete or integrated components, graphics processing units (GPUs), field programmable gate arrays (FPGAs), or application-specific integrated circuits (ASICs)). In some embodiments, the described elements may be configured to reside on tangible, persistent, addressable storage media and configured to execute on one or more processors. These functional elements may, in some embodiments, include, for example, components, processes, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables, such as software components, object-oriented software components, class components, and task components. While example embodiments are described with reference to components, modules, and units discussed herein, such functional elements may be combined into fewer elements or divided into additional elements.

[0138] Database As described above, various scores can be calculated or determined as part of the VERDI method. To facilitate the determination of these scores, various databases for storing information can be used. This allows scores to be calculated quickly. For example, Figure 7a shows an example of an epitope database containing the sequences (8–20 amino acids in length) of all putative epitopes from proteins (e.g., all proteins expressed from the SARS-CoV-2 genome). Epitope databases from different proteins associated with different diseases are typically stored separately. HLA data is divided into an HLA class I database and an HLA class II database, each containing data representing the sequences of HLA alleles (e.g., 4- to 8-digit numbers, as described below). The HLA database in Figure 7a, designated MHC1HLA, contains the ID and name of each class I HLA allele. The population database shown in Figure 7a may contain individual IDs, their six HLA class I alleles, and their 12 HLA class II alleles.

[0139] A database named "Peptides" includes the ID of each peptide, its amino acid sequence (code), the ID of the left subepitope, and the ID of the right subepitope. The left and right subepitopes are each one amino acid shorter than the original peptide. For example, for the sequence TLDSKTQSL (SEQ ID NO: 672), the left subepitope is TLDSKTQS (SEQ ID NO: 1115) and the right subepitope is LDSKTQSL (SEQ ID NO: 1116). A peptide database can be built based on a single protein of interest (e.g., a SARS-CoV-2 protein). The protein is cut or split into multiple smaller sequences, and the database can include all of these smaller sequences (sometimes called putative epitopes). If the SARS-CoV-2 genome encodes 29 proteins, there are approximately 70,000 or 100,000 smaller sequences derived from the original viral protein, all of which must be paired with HLA class I or HLA class II alleles, respectively. Putative epitopes may also originate from different protein products translated from alternative splice regions or alternative reading frames.

[0140] As described above, an EL score can be calculated for each known HLA allele-peptide pair. Each score is stored along with the peptide ID and HLA allele ID, as shown in Figure 7a. The database, designated "Class1Ligandx" for example, is called a ligand dataset. Figure 7a shows only four of these "Class1Ligandx" datasets to illustrate the details of each database and the flow of data between them. In summary, these intermediate datasets are created by pairing a peptide with x amino acids from the peptide database with one of the Class-I HLAs in the MHC1 HLAs database and generating a score for each pair of peptide and Class-I HLA.

[0141] Once individual data are acquired, EL scores for the paired sequences and HLA alleles (both class I and class II) can be obtained for each individual. As shown in Figure 7a, EL scores can be obtained from the appropriate ligand database constructed as described above. EL scores can be stored for each individual in a second intermediate dataset named, for example, "CL_GenoTypeTreex." Each dataset includes the subject's ID, the ID of the paired peptide and class I HLA, and the scores for the paired peptide and class I HLA. In this example, x is a number between 8 and 14 for HLA class I alleles. Similar databases are generated for HLA class II alleles. These databases can be considered subject ligand databases.

[0142] Figure 7b shows how the subject ligand dataset is used to obtain a set of output datasets labeled "Cl-PersonsEpitopeTreex," where x is a number between 8 and 14 for class I and between 11 and 20 for class II. The table labeled "Cl-PersonsEpitopeTreex" contains the efficacy of each peptide for each subject and can therefore be called the "personal epitope efficacy database."

[0143] The table labeled "Cl-PersonsEpitopeTreex" contains the subject IDs (anonymized) and the IDs of the epitopes whose potency is of interest. The ability of an epitope to activate a T cell depends on the number of epitopes present on the cell surface (the number of epitopes that simultaneously or sequentially signal the TCR). The cell surface density of each epitope can be quantified using the EpitopeWeightScore (EWS) described above. EWS is the number of contributing HLA-epitope pairs, each weighted by their associated EL score (or associated / contributing EL score (EL)). EWS may also be weighted by the expression level and stability of different HLA alleles. For example, as described in "Variations in HLA-B cell surface expression, half-life, and extracellular antigen receptivity" published by Yarzabek et al. in eLife Immunology and Inflammation in July 2018, the HLA-B*0.8-01 allele has higher cell surface expression and cell surface stability in lymphocytes compared to other alleles (HLA-B*51-01).

[0144] The EL scores used to calculate the EWS can be obtained from the Cl_GenoTypeTreex table; therefore, as shown in Figure 7b, there is an arrow indicating data flow from Cl_GenoTypeTree8 to Cl-PersonsEpitopeTree8 (and so on). The table labeled "Cl-PersonsEpitopeTreex" contains the number of HLAs for which the peptide's EL score exceeds the EL threshold. Similarly, as shown in Figure 7b, data flow can occur from the table labeled "Cl-PersonsEpitopeTreex" containing the EpitopeWeightScore of peptides one shorter than the peptide under consideration. In other words, the EWS of the previous table can be used to calculate the subepitope score.

[0145] The table labeled "Cl-PopulationEpitopeTreex" contains the efficacy of each peptide for the entire population and can therefore be referred to as the global peptide efficacy database. The data in the table labeled "Cl-PersonsEpitopeTreex" can be used, along with the input population data, to create the data in the table labeled "Cl-PopulationEpitopeTreex." This table contains the IDs of the peptides whose efficacy is of interest. This table contains the various scores summed across the entire training population, and therefore the number of people contributing to the total.

[0146] For the shortest peptides (e.g., 8 in length in this example), the scores include: SumEpitopeWeightScore (SEWS) is the sum of the EpitopeWeightScores (EWS) of the peptides across all subjects in the population, and SumOverallWeightScore (SOWS) is calculated as follows: For longer peptides, in addition to these sums, the following additional scores are also included: SumSubitopeWeightScore_1 (SSWS1) is the sum of the SubitopeWeightScore_1 (SWS1) of the peptides across all subjects in the population, and SumSubitopeWeightScore_2 (SSWS2) is the sum of the SubitopeWeightScore_1 (SWS2) of the peptides across all subjects in the population. These calculations may be performed using the following formulas:

number

[0147] Figure 7b shows that the data in these output tables can be adjusted using filter parameters. These include ELScoreFilter and EWSFilter. As explained above, EWS is calculated by summing only those peptides whose EL scores exceed an EL threshold, and this EL threshold can be set using ELScoreFilter. The EL threshold is set to remove peptides that are unlikely to bind to HLA (peptides that are very unlikely to be transported to the cell surface). For EL scores between 0 and 1, an EL threshold of 0 sums all scores. Similarly, an EWS threshold can be set within EWSFilter. EWSFilter is set to exclude epitopes that are unlikely to provoke a T cell response (those that are present in low amounts on the cell surface to activate T cells). As mentioned above, this threshold is set to target the number of peptides that should be included.

[0148] Figures 7a and 7b show detailed tables for each Class I HLA under consideration. The same data flow and resulting data tables can be generated for Class II HLA analysis. Figure 7c shows a data flow that combines the information from Figures 7a and 7b, but omits the detailed data within each table for clarity. It should be understood that the data in each table is identical to Figures 7a and 7b, except that "II" is used instead of "I" to indicate that Class II HLA is being considered. Similarly, Figures 7a and 7b can be combined with the same data flow shown in Figure 7c. Similar to Figures 7a and 7b, the tables in Figure 7c are labeled "Class2Ligandx," "CII_GenoTypeTreex," "CII_PersonsEpitopeTreex," and "CII_PopulationEpitopeTreex," where x is the number of amino acids in each peptide. In Figures 7a and 7b, x varies between 8 and 14, while in Figure 7c, x varies between 11 and 20.

[0149] Validation model training and inference As shown in FIG. 5a, the first step is to obtain a ranked list of epitopes. As described above, ranking can be based on EWS scores. To further validate the method described above, we developed a validation system, shown schematically in FIG. 8a. As shown, there are two separate inputs: the HLA genotype data of the test subject and the input epitope (including its sub-epitopes). These separate inputs are input as pairs to a scoring module 800. The scoring module generates multiple scores for each pair, as described above in connection with FIG. 4. Although multiple arrows are shown, three are merely illustrative; typically, six scores are calculated when considering HLA class I molecules. For example, each score can be derived from the EL scores generated by NetMCHPan4.1 or a similar module. These scores are output in a feature vector 802.

[0150] Features related to the original input epitope are also input into the feature vector. One feature, called the "log fold change," quantifies the potency of the epitope compared to a baseline that indicates the proliferation of antigen-specific T cells in the absence of the epitope or after infection. Other features include, for example, the major cores of each of the HLA-A, HLA-B, and HLA-C class I alleles. The core is the pattern or sequence of amino acids within the epitope that is recognized by the TCR.

[0151] The feature vector is then input to a machine learning module 804 that is trained as described below. The machine learning module 804 may use suitable techniques, such as gradient boosting decision trees (GBDTs), ridge regression, logistic regression, and balanced random forest techniques (e.g., as described in Jerome H Friedman, "Stochastic gradient boosting", Computational Statistics and Data Analysis, vol. 38, no. 4, pp. 367-378, 2002; Vovk V. "Kernel ridge regression in Empirical inference", Springer, 2013. p. 105-16; Tolles et al., "Logistic Regression: Relating Patient Characteristics to Outcomes", JAMA. 2016, 316(5):533-534. doi:10.1001 / jama.2016.765 or https: / / statistics.berkeley.edu / sites / default / files / tech-reports / 666.pdf). The machine learning module 804 is used to predict the probability that an input epitope is immunogenic for an individual for whom HLA genotype data was obtained. This method can be run repeatedly for multiple input epitopes, which can be ranked based on their probabilities. The reported output is the specificity of each epitope (i.e., the amino acid sequence within the epitope) and the potency of each epitope (e.g., a value between 0 and 100 of the probability that the epitope will elicit a T cell response). The top-ranked epitopes define an individual's T cell antigen repertoire and can be used to predict an individual's response to SARS-CoV-2 or specific diseases, such as cancer.

[0152] Figure 8b illustrates the training method for such a machine learning model and its use in the inference phase. In the first step of the training phase, S800, cohort data to be used to train the model is acquired. The cohort data may include complete HLA genotypes of individuals. Examples of such data are collected by ImmPort, and the present technology used the ImmPort database (e.g., https: / / www.immport.org / shared / home) to access complete HLA genotypes of individuals from the European SDY614 and US SDY28 populations. Cohort data also includes data assessing the efficacy of an individual's T cell response to a particular disease.

[0153] For example, a model can be trained to accurately rank SARS-CoV-2-specific T cell antigens by potency. Examples of cohort data suitable for training models to make such predictions include the ABF data described by Saini et al. in "SARS-CoV-2 Genome-Wide T-Cell Epitope Mapping Reveals Immunodominance and Substantial CD8+ T Cell Activation in COVID-19 Patients," published in Sci Immunol on April 14, 2021. The ABF cohort includes 79 individuals characterized by four-digit genotypes of six HLA class I alleles, representing the sequence of the HLA epitope-binding pocket. Another example of cohort data is the adaptive data described by Snyder et al. in "Magnitude and Dynamics of the T-Cell Response to SARS-CoV-2 Infection at Both Individual and Population Levels," published in MedRxiv in 2020. The adapted data includes 114 individuals with four-digit genotypes.

[0154] The next step, S802, is to prepare training data and validation data from the cohort data. The ABF cohort data can be used as training data and cross-validation data. The set of clinical outcomes recorded in this data includes healthy, high-risk healthy, outpatient, and inpatient. In preparing the data, epitopes can first be selected and paired with individual data. The selected epitopes are those that have strong binding to one or more major HLA class I alleles in each subject. Binding strength can be calculated using an appropriate method, such as NetMHCpan 4.1 prediction, as described above in connection with calculating the EWS score. The score used for the feature vectors in the inference phase is calculated for each subject and each selected peptide. T cell responses can be measured using labeled peptide-MHC-I multimers to quantify the efficacy of a CD8+ T cell antigen compared to no antigen or baseline, as a "log fold change (LFC)" that indicates the proliferation of antigen-specific T cells in the body after SARS-CoV-2 infection. The "log fold change" value can be included in the training data. All published efficacy data for epitopes for each HLA allele-matched individual were also obtained. 2,204 epitopes were selected, and efficacy data for an average of 920 epitopes [209-1452] could be obtained. As a mere example of the data generated in this step, the following table shows an example of 10 peptides for two individuals: a hospitalized individual and another healthy individual. Note that the first 10 peptides for these two individuals are not identical. [Table 10] TIFF2025540913000038.tif235159TIFF2025540913000039.tif237159TIFF2025540913000040.tif150159

[0155] The adaptive cohort data can be used as validation data. The set of clinical outcomes recorded in this data includes acute Covid-19, non-acute Covid-19, convalescent Covid-19, exposed Covid-19, and healthy (no history of exposure). To prepare the data, epitopes can be selected and paired with individual data, similar to the ABF data. Again using NetMHCpan 4.1 prediction, in this example, 545 distinct HLA class I-binding epitopes are predicted. An average of 169 epitopes (range 18-240) per individual were tested using TCR sequencing. This assay quantifies the efficacy of T cell antigens as "hits" compared to a non-antigen control group. "Hits" represent the copy number of each TCR sequence, which is used to assess the proliferation of antigen-specific T cell clones within an individual after SARS-CoV-2 infection. Because we tested pools of peptides in multiple experiments, we hypothesized that one epitope within the pool would dominate the T cell response due to its strongest binding affinity to the subject's major HLA allele (highlighted in bold). This resulted in an average of 51 epitopes [range 5-121] tested per subject. As a mere example of the data generated in this step, the table below shows an example of 10 peptides in two individuals: an acute COVID-19 patient and a healthy subject. Note that the first 10 peptides in these two individuals are not identical. [Table 11] TIFF2025540913000042.tif218159TIFF2025540913000043.tif235159TIFF2025540913000044.tif26159

[0156] The adaptive data only report immunogenic epitopes (T cell antigens), and no results from experiments without T cell responses exist. Therefore, the "hit" variable described above is used to rank epitopes for each individual. For each peptide selected within a group, a score is calculated for each subject and each selected peptide, which is used in the feature vector for the inference stage. [Table 12]

[0157] Additional potency data can be obtained for each identified peptide as well, as shown in the table below. The primary cores for each of the A, B, and C loci were also obtained but are not shown for ease of presentation (the primary cores for these peptides are the same as the overall primary core): [Table 13] TIFF2025540913000047.tif134159

[0158] Once the training data and validation data are prepared, a model can be trained using the training data in step S804. As shown in step S806, performance or evaluation metrics of the trained model can be calculated as needed using validation (i.e., test) data. As shown in FIG. 8b, depending on the training method selected, the loop of steps S804 and S806 may be performed multiple times. By way of example only, a 5-fold cross-validation procedure may be used, whereby clinical data is split into 5 random splits, dividing the training data and test data into 80% and 20%. The splits may be stratified to maintain the ratio of immunogenic to non-immunogenic antigens in each subject. Grouped data cannot be shared between the training and test datasets.

[0159] Evaluation metrics may include population and / or individual metrics. Population metrics are calculated across the entire dataset and consider each data point in the dataset without regard to individuality. Such metrics provide useful comparative information on model performance, as explained in more detail in the Model Validation section below. Examples of suitable population metrics include receiver operating characteristic (ROC) and precision-recall (PR) curves, calculated using known techniques. PR curves are considered a better alternative to ROC in this class-imbalanced scenario and may provide a more accurate measure of binary classification performance compared to random models. The reported metrics are the area under the ROC curve (AUC-ROC) and the area under the PR curve (AUC-PR), both averaged over all data points.

[0160] The individualized index is specific to an individual and is more consistent with the diagnostic capabilities of the current method. At the individual level, we aim to identify the top N epitopes with the highest T cell responses, where N is defined according to the number of epitopes of interest per individual. For the ABF dataset, N can be set to 20, while for the adaptive dataset, N can be set to 3. The smaller value for the adaptive dataset reflects the fact that fewer epitopes are characterized per individual. For each top N case, one of the performance metrics that can be calculated is the accuracy rank A. R Yes, this is calculated as follows:

number

[0161] In addition to (or instead of) the accuracy rank, we can calculate the precision P and recall R for each individual using the following formulas:

number

[0162] In the adaptive dataset, only positive values ​​are reported. In this case, another metric common in positive unlabeled learning can be used. The positive unlabeled metric PU can be calculated as follows:

number

[0163] Once the model is trained, it is output in step S808 and used to predict individual T cell responses. Referring to FIG. 2a, the inputs to the computer-implemented method are the individual data and epitope data, shown as being acquired in steps S810 and S812, respectively. The data may be acquired simultaneously as shown, or in any order. The next step is to prepare a feature vector in step S814. The feature vector contains as much information as possible from the columns of training data. Thus, the feature vector may include some or all of the following scores calculated as described above: average EWS score, overall EWS score, EWS score for each locus (e.g., A, B, and C for HLA class I), and SWS score for each locus. The feature vector may also include information related to the epitope under consideration, including some or all of the overall major core, major core for each locus, genomic index, start index, end index, number of hits, and major HLA. Once the feature vectors are prepared, the trained model can be used to infer a rank for each epitope for the individual in step S816. The rankings can be displayed to the user (e.g., on a display screen) in step S818 or output to another system for further processing (e.g., to generate vaccine or TCR therapy recommendations).

[0164] Results – Prediction of SARS-CoV-2-specific T cell responses in individuals The rankings predicted by the trained model were compared with those predicted by random selection (referred to as Random) and those predicted by considering only the maximum EL score (referred to as the EL Max model). Figure 9a shows the ROC curves for each of the three prediction methods, showing the true positive rate versus false positive rate, along with the area under the curve (AUC) values. The AUC of the current method (referred to as VERDI) is 0.71, superior to the EL Max model (0.57) and random selection (0.5). In fact, the EL Max model is only slightly better than random guessing. This result is confirmed by the PR curve (recall vs. precision) plotted in Figure 9b. Again, the current method (referred to as VERDI) performs best, with an area under the curve of 0.0452, significantly higher than the EL Max model (0.0135) and random guessing (0.0088).

[0165] Moving on to individual performance metrics, Figure 9c shows a bar graph comparing the average accuracy ranks of the current method and the EL-Max model for the top 10, top 20, top 50, and top 100 epitopes averaged across individuals. The current method (VERDI) significantly outperformed the EL-Max model, which correctly predicted fewer than 10% of the top 10, top 20, and top 50 test epitopes that elicited detectable T cell responses in HLA-matched test subjects. In contrast, the current method was able to predict up to 40% of T cell antigens when all epitopes were considered. These results validate the performance of the proposed method for predicting T cell antigen specificity and potency at the individual level.

[0166] Figures 9d to 9f compare the accuracy, precision, and recall metrics for the top 20 ranked epitopes (candidate T cell antigens) across individuals. Comparisons using only the EL Max score showed limited performance across these metrics for most individuals. In contrast, the proposed method performed adequately across most individuals, albeit with significant inter-individual variability. While we used the same decision threshold to calculate accuracy and precision, we also explored other thresholds for the model with similar results.

[0167] To evaluate the clinical performance of the proposed method outside the ABF cohort, we used an adaptive dataset. The adaptive dataset was independently collected by Adaptive Biotechnologies for the development of the first T cell response diagnostic test. The adaptive dataset includes different epitopes, different cohorts of individuals living in different regions of the world (the United States and Denmark), and different methods of quantifying T cells (TCR sequencing vs. multimer staining). The adaptive dataset also tested a different number of SARS-CoV-2 epitopes per subject, averaging 51 [range 5-121] versus 920 [range 209-1452] in the ABF dataset.

[0168] Figures 10a through 10d show validation results using the adapted dataset. We used the proposed method to rank 70,000 putative epitopes of SARS-CoV-2 for each individual in the adapted dataset. In the adapted dataset, experimental results are reported only for epitopes that induced T cell responses (T cell antigens). Therefore, Figures 10a through 10d focus on the average accuracy rank metrics across individuals. Figure 10a plots the average accuracy ranks across all individuals for the top 1, top 2, top 3, top 5, and top 10 epitopes to compare the proposed method (VERDI) with the EL Max model. Because fewer epitopes were tested per individual in the adapted dataset, the number of ranked epitopes evaluated is smaller than the rankings generated for the ABF dataset. As shown in Figure 10a, the proposed method achieved approximately 30% accuracy for the top 1, top 2, and top 3 ranks. Our method outperformed the EL Max model in all but the top 10 rankings. In the adapted dataset, only positive T cell responses against epitopes predicted to be highly immunogenic were reported, which may result in excessive false positives not being adequately penalized. Therefore, the EL Max model, which generates a large number of false positives, appears to perform better than the proposed method for rankings with large N.

[0169] Figure 10b plots the average PU index of the top 1, top 2, top 3, top 5, and top 10 ranked epitopes across all individuals to compare the proposed method (VERDI) with the EL Max model. The PU index is used to correct for false positives. As shown in Figure 8b, the proposed method outperforms the El Max model in all top-N ranked lists.

[0170] Figures 10c and 10d focus on the top three rankings. Figure 10c plots the number of individuals against the accuracy rank, while Figure 10d plots the number of individuals against the PU index. As shown in Figure 10c, the proposed method predicts at least one of the three most potent T cell antigens identified in FDA-approved T cell response diagnostics in most subjects. Figure 10d shows that the proposed method has good predictive performance in most subjects. Figures 10a to 10d show that the proposed method is a better model for predicting the specificity and potency of T cell responses than the EL max model, even on this challenging adaptive dataset.

[0171] Individuals typically respond to an average of 30 epitopes (see, e.g., Ref. 17 (Tarke)). Therefore, outputting the top 50 most potent epitopes of SARS-CoV-2 for an individual should adequately characterize that individual's SARS-CoV-2 response. Figures 11a and 11b plot the top 50 ranked epitopes as 9-mer cores for two anonymized individuals (IDs: ADAP-142 and ADAP-6359) from the adapted dataset. Figures 11a and 11b plot potency (probability of a T cell response) against specificity (denoted by position in the proteome). HLA genotype data for each individual are shown in the table below: [Table 14]

[0172] Figures 11a and 11b show the highly diverse distribution of T cell antigens in the SARS-CoV-2 proteome. This result is replicated for each of the 2,346 individuals evaluated. Only a small proportion of T cell antigens are shared by a subset of individuals, and no single T cell antigen is immunogenic across all individuals. However, despite variation in T cell antigen specificity, potency across individuals is relatively consistent. The dotted lines in Figures 11a and 11b indicate the average potency (i.e., mean value) of the top 50 antigens, which can be defined as the strength of an individual's response. Individual ADAP-142 had a mean potency of 68%, with minimum and maximum values ​​of 64% and 80%, respectively. Individual ADAP-6359 had a mean potency of 82%, with minimum and maximum potency values ​​of 74% and 96%, respectively. Individual ADAP-142 produced a weak T cell response because no T cell antigens ranked with potency above 80%. In contrast, individual ADAP-6359 has a robust T cell response due to the absence of a ranked T cell antigen with efficacy above 80%.

[0173] Diagnosing the strength of SARS-CoV-2-specific T cell responses is important because it correlates with infection prognosis and vaccine protection (e.g., as described in Reference 3 (Toss)). Figures 11c and 11d plot the average potency of the top 50 ranked epitopes against their prevalence (%) in the population. Figures 11c and 11d resemble Gaussian distributions in separate US and EU populations. Similar results would be expected for the 12 HLA class II molecules. This suggests that the strength of T cell responses to SARS-CoV-2 infection is inherited as a genetic phenotype dependent on HLA genotype. We propose that subjects with weak T cell responses, located in the left tail of the Gaussian curve, will develop symptomatic COVID-19 and require additional intervention.

[0174] In individuals with strong T cell responses, a small number of potent antigens rapidly and vigorously stimulate T cells to proliferate and kill infected cells (e.g., eight T cell antigens with >90% efficacy). After rapidly clearing the infection, some activated T cells form a memory pool, while the remainder undergo apoptosis. In contrast, in individuals with low T cell antigen activity, SARS-CoV-2 elicits a broad T cell response. Low-activity antigens slowly activate T cells, resulting in slow proliferation and slower clearance of infected cells. For example, the individual shown in Figure 11a has 50 T cells with 64% to 80% efficacy. While the virus replicates, additional antigens activate T cells, leading to an expansion of the T cell repertoire. A weak T cell response results in the expression of inhibitory molecules (PD1, Tim3) that promote viral persistence. Individuals with weak T cell responses are more likely to experience high viral loads and severe COVID-19 disease (as suggested in Reference 19 (Cevik)). The extreme heterogeneity of T cell antigen specificity and potency explains the diverse clinical findings in SARS-CoV-2-infected individuals, including expansion, maintenance, and exhaustion of T cell responses. Healthy individuals, such as ADAP-6359, may not require a vaccine to kill infected cells because they have a strong T cell response against diseased cells. However, healthy individuals, such as ADAP-6359, may require a vaccine containing an antigen that enhances the T cell response to kill infected cells. After T cell vaccination, ADAP-6359 possesses long-term memory cells that can rapidly proliferate and destroy infected cells after viral infection. Therefore, the top-ranked epitopes can be used to characterize the potency of an individual's T cell response against diseased cells. T cell antigens may be selected for vaccine development because they induce strong T cell responses in recipients with matching HLA genotypes. These vaccines are effective at attacking diseased cells presenting the same epitope and are useful for the prevention and treatment of SARS-CoV-2, other viral infections, cancer, and several other diseases.

[0175] Personalized vaccines and their uses In the following sections, various aspects of the present disclosure relating to personalized vaccines and their uses are defined in more detail. Each aspect so defined may be combined with other aspects unless otherwise specified. In particular, any feature described as being preferred or advantageous may be combined with any other feature described as being preferred or advantageous.

[0176] Generally, the terms and techniques used in connection with cell and tissue culture, pathology, oncology, molecular biology, immunology, microbiology, genetics, protein and nucleic acid chemistry and hybridization described herein are well known and commonly used in the art.Unless otherwise specified, the methods and techniques of the present disclosure are carried out according to the conventional methods well known in the art and the methods described in various general and more specific references cited and described herein.See, for example, Green and Sambrook et al., Molecular Cloning: A Laboratory Manual, 4th ed., Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY (2012).

[0177] Peptide synthesis and purification techniques are carried out according to manufacturer's specifications, methods commonly practiced by those skilled in the art, or methods described herein. As used herein, terms related to analytical chemistry, synthetic organic chemistry, and medicinal chemistry, as well as experimental procedures and techniques in these fields, are those well known and commonly used by those skilled in the art. Standard techniques are used for chemical synthesis, chemical analysis, pharmaceutical preparation, formulation, administration, and patient treatment.

[0178] In the present invention, unless otherwise specified, the scientific and technical terms used herein have the meanings commonly understood by those skilled in the art. Furthermore, the laboratory procedures used herein for cell culture, molecular genetics, nucleic acid chemistry, and immunology are all standard procedures widely used in the corresponding fields. Meanwhile, in order to better understand the present invention, the definitions and explanations of relevant terms are provided below.

[0179] In the description of this invention and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly dictates otherwise. Also, as used herein, "and / or" refers to and includes all possible combinations of one or more of the associated listed items, and, alternatively, the lack of combinations ("or").

[0180] The term "about," when referring to a measurable value such as peptide amount, dosage, time, temperature, enzyme activity, or other biological activity, is intended to include variations of ±20%, ±10%, ±5%, ±1%, ±0.5%, or ±0.1% of the specified amount.

[0181] The term "multiple" is defined as "at least two," "two or more," or "plurality."

[0182] Limitations of personalized vaccines currently under investigation Existing personalized peptide vaccines in clinical trials consist of a mixture of multiple peptides encoding epitopes and adjuvants. In some cases, the epitopes are expressed using mRNA, DNA, or viral vectors. The development process for currently used personalized vaccines is time-consuming and expensive. This process involves identifying unique mutations, predicting which mutations will trigger an immune response, manufacturing the vaccine, and finally administering it to patients. Personalized peptide vaccines frequently utilize long, overlapping peptides or predicted peptides that bind to specific HLA alleles. However, the process of predicting which epitopes will trigger an immune response is not accurate because any predicted epitope can elicit both CD8 and CD4 T cell responses against the target protein. Even if a personalized vaccine contains multiple predicted epitopes for a patient, not all patients will demonstrate a strong immune response to the vaccine. Furthermore, producing personalized vaccines presents significant logistical and manufacturing challenges. Peptide vaccines involve issues such as peptide solubility, interactions between peptides, quality control, and potential side effects from adjuvants. Therefore, personalized vaccines may not be effective and accessible to all patients. The inventors have aimed to address these limitations.

[0183] Personalized vaccines designed with the VERDI system and computer-implemented methods The present invention is a personalized VERDI vaccine designed by the present system and computer-implemented method. This vaccine is a single peptide encoding at least one cell surface antigen containing a set of highly ranked epitopes presented on the cell surface by the recipient's HLA class I and / or HLA class II molecules, and therefore has the advantage of simultaneously inducing T cell responses from both CD8 and CD4 cells in individuals. The personalized vaccine presented herein is patient-specific and disease-independent. This vaccine can be used to treat diseases such as cancer, viral infections, and bacterial infections.

[0184] Efficacy against infectious diseases or cancers that develop over time is achieved by administering a set of personalized VERDI vaccines (simultaneously and sequentially) that target different antigens expressed in unhealthy cells. However, unlike current vaccines, each vaccine only requires a single dose. Adjuvants are optional and not required for the personalized vaccines of the present invention.

[0185] The VERDI personalized vaccine described herein is a novel invention that can take a variety of forms. One preferred embodiment of the invention comprises a single synthetic peptide chain encoding interchangeable CD8 and CD4 antigens. These antigens consist of a set of highly ranked epitopes presented by HLA molecules on the surface of target cells. These antigens are derived from proteins expressed on target cells, such as tumor cells or infected cells.

[0186] A unique feature of this vaccine is the use of a polyarginine bridge (a peptide consisting of eight or more arginine units) that covalently links the CD8 and CD4 antigens within the peptide chain (as shown in Figures 12a, 13a, and 14a). The VERDI vaccine, with a length of 26–40 amino acids, offers a personalized and innovative solution to induce effective T cell responses using natural peptides.

[0187] The polyarginine cell-penetrating peptide plays a key role in the VERDI vaccine composition: it functions as an immunologically inert bridge, facilitating the uptake of the VERDI vaccine into antigen-presenting dendritic cells and subsequently promoting the induction of potent T cell responses. The introduction of this novel peptide vaccine platform represents an important advance in the field of personalized cancer vaccines, as it addresses limitations of existing peptide vaccines in terms of (i) antigen incorporation, (ii) simultaneous induction of potent CD8 and CD4 T cell responses against a target protein with a single peptide, and (iii) rapid pharmaceutical formulation of the vaccine from two synthetic peptides using the Diels-Alder reaction.

[0188] Further preferred aspects and embodiments of the present invention are described in detail below.

[0189] In one aspect of the invention, there is provided a personalized vaccine or therapeutic composition prepared according to any of the computer-implemented methods described above.

[0190] In a further aspect of the present invention, the personalized vaccine or therapeutic composition comprises at least one peptide antigen derived from a protein expressed in an unhealthy cell of the subject, which peptide antigen comprises a set of overlapping epitopes presentable by the subject's HLA class I and / or class II molecules, induces a CD4 and / or CD8 T cell response, shares a common sequence with multiple epitopes presentable by the subject's HLA class I and / or class II genotype, and optionally comprises a cell membrane-penetrating peptide.

[0191] In one embodiment, the cell membrane-permeable peptide may be located at the N-terminus of at least one peptide antigen. In another embodiment, the cell membrane-permeable peptide may be located at the C-terminus of at least one peptide antigen. In embodiments containing at least two peptide antigens, the cell membrane-permeable peptide may be located at the N-terminus of the peptide formed by at least two antigens, at the C-terminus of the peptide formed by at least two antigens, or between at least two antigens. In some embodiments containing two or more antigens, the vaccine may include two or more cell membrane-permeable peptides.

[0192] In one embodiment, at least one peptide antigen can contain one T cell epitope. In another embodiment, at least one peptide antigen can contain at least two T cell epitopes. The at least two T cell epitopes can be located adjacent to each other, separated by separate peptide regions, or overlapping. In one embodiment, the antigen can contain multiple highly ranked epitopes presented by the subject's HLA molecule, which is advantageous because it can help induce CD4 and / or CD8 T cell responses.

[0193] In one embodiment, the personalized vaccine or therapeutic composition comprises at least two antigens, including at least one antigen that induces a CD4 T cell response and at least one antigen that induces a CD8 T cell response. The antigenic peptides of the present invention take into account the importance of both CD8 and CD4 T cell responses against proteins expressed in unhealthy cells. Individual CD8 and CD4 epitopes carefully matched to the recipient's HLA genotype and disease-specific proteins and identified as being highly ranked are included in the peptide antigens selected for the vaccine composition. Thus, individually inducing both CD8 and CD4 T cell responses against diseased cells represents a significant improvement over the efficacy of current T cell vaccines.

[0194] The peptide antigens of the present invention are advantageous because they can simultaneously induce CD8 and CD4 T cell responses against diseased cells, improving the efficacy of personalized vaccines. Current vaccine peptides are not designed to simultaneously induce CD8 and CD4 T cell responses against proteins expressed on target cells, even though help signals from CD4+ T cells to CD8+ T cells during priming optimize the magnitude and quality of CTL responses. Furthermore, CD4+ T cells can kill tumor cells and infected cells that display specific epitopes recognized by TCRs on their surface.

[0195] In one embodiment, at least one peptide antigen comprises a set of epitopes with a "core" recognized by the TCRs of both the subject's CD4 and CD8 T cells. Epitopes are known to be "promiscuous," i.e., they can bind to more than one HLA class I or HLA class II molecule. However, in the present invention, our system and computer-implemented method are different because they analyze HLA-epitope recognition and TCR-epitope recognition at the cell surface at an individual level. In one embodiment, at least one peptide antigen may comprise 1 to 40 highly ranked epitopes presented by the subject's HLA class I molecules and / or 1 to 40 highly ranked epitopes presented by the subject's HLA class II molecules present on the subject's cell surface.

[0196] In one embodiment where the personalized vaccine comprises at least two peptide antigens, the cell membrane-penetrating peptide is positioned between the at least two antigens. In one aspect of the present invention, an antigenic peptide is provided that comprises the amino acid sequences of at least two peptide antigens and a cell membrane-penetrating peptide. In one embodiment of this aspect, one of the antigens can activate CD4+ helper T cells, and the other antigen can activate CD8+ cytotoxic T cells, and the cell membrane-penetrating peptide is positioned between the CD4 antigen amino acid sequence and the CD8 antigen amino acid sequence (Figures 12a, 13a, and 14a).

[0197] An additional advantage of the present invention is that personalized vaccines consist of or contain a single synthetic peptide chain. This eliminates problems of poor solubility, precipitation, and peptide reactivity, making vaccine preparation and quality control processes more efficient and reliable. Current peptide vaccines can contain 8 to 20 peptides, and mixtures of these different peptides have different solubilities, resulting in some precipitation. Furthermore, peptides in a mixture may react with each other. This makes quality control of peptide mixtures time-consuming and difficult, and confirming their identity is difficult because they cannot be separated, for example, on an HPLC column.

[0198] Current peptide vaccines also contain adjuvants to enhance immunogenicity (e.g., by providing T cell help for CD8 T cell and antibody responses), but these can cause mild to moderate local side effects (e.g., pain, erythema, edema, swelling, warmth, redness, etc.), which can sometimes be severe and require medical treatment, and repeated vaccinations can be poorly tolerated. The personalized peptide vaccines of the present invention are adjuvant-independent, thereby reducing the potential for side effects associated with their use and, because they simultaneously induce strong CD8 and CD4 T cell responses, do not require repeated injections. In some embodiments, the present invention does not contain any adjuvants.

[0199] Cell-penetrating peptides (CPPs) As used herein, the term "cell membrane-permeable peptide" refers to a peptide that can deliver a molecule to which it is bound through the cell membrane into a cell. The cell membrane-permeable peptide ensures that the vaccine quickly penetrates cells, including antigen-presenting dendritic cells, and activates T cell responses. For example, the cell membrane-permeable peptide of the present application has the ability to deliver a biological molecule of interest (e.g., a peptide of interest or a nucleic acid of interest) bound thereto through a membrane into a cell. In the present application, the cell membrane-permeable peptide can be bound to the biological molecule of interest (e.g., a peptide of interest or a nucleic acid of interest) via a covalent or non-covalent bond.

[0200] For example, the cell membrane-permeable peptide of the present application can be linked to a peptide of interest by a covalent bond (optionally via a linker, e.g., a peptide linker). Thus, in certain embodiments, the cell membrane-permeable peptide of the present application can be fused to a peptide of interest via a peptide linker, if desired. Methods for linking a peptide molecule to a peptide or nucleic acid of interest are known to those skilled in the art, for example, by using various known bifunctional linkers.

[0201] Furthermore, the cell membrane-permeable peptide of the present application can be bound to a target biological molecule (e.g., a target peptide) by a non-covalent bond. Thus, in certain embodiments, the cell membrane-permeable peptide of the present application can be bound to a target biological molecule (e.g., a target peptide) through a specific intermolecular interaction / specific bond (e.g., interaction / bonding between an antigen and an antibody, or interaction / bonding between a DNA-binding domain and a DNA molecule).

[0202] The present invention is not limited to any particular CPP or sequence, however, the following specific sequences may be preferred or advantageous because of their sequence, function, tissue specificity, or mode of action.

[0203] In one embodiment, the CPP is HIV-TAT comprising the sequence GRKKRRQRRRPQ (SEQ ID NO: 1029). In one embodiment, the CPP is an 8-polyarginine comprising the sequence RRRRRRRR (SEQ ID NO: 1030). In one embodiment, the CPP is a 9-polyarginine comprising the sequence RRRRRRRRR (SEQ ID NO: 1031). In one embodiment, the CPP is penetratin comprising RQIKIWFQNRRMKWKK (SEQ ID NO: 1032). In one embodiment, the CPP is KLAL comprising the sequence KLALKLALKALKAALKLA (SEQ ID NO: 1033). In one embodiment, the CPP is VP-22 comprising the sequence DAATATRGRSAASRPTERPRAPARSASRPRRPVD (SEQ ID NO: 1034). In one embodiment, the CPP is MPG comprising the sequence GALFLGFLGAAGSTMGAWSQPKKKRKV (SEQ ID NO: 1035). In one embodiment, the CPP is KADY comprising the sequence Ac-GLWRALWRLLRSLWRLLWKAcysteamide (SEQ ID NO: 1036). In one embodiment, the CPP is pVEC comprising the sequence LLIILRRRIRKQAHAHSK-NH2 (SEQ ID NO: 1037). In one embodiment, the CPP is M-918 comprising MVTVLFRRLRIRRASGPPRVRV-NH2 (SEQ ID NO: 1038). In one embodiment, the CPP is KALA comprising WEAKLAKALAKALAKHLAKALAKALKACEA (SEQ ID NO: 1039). In one embodiment, the CPP is PEP-1 comprising Ac-KETWWETWWTEWSQPKKKRKC-cya (SEQ ID NO: 1040). In one embodiment, the CPP is EB1 comprising LIKLWSHLIHIWFQNRRLKWKKK (SEQ ID NO: 1041). In one embodiment, the CPP is transportan comprising the sequence GWTLNSAGYLLGKINLKALAALAKKIL (SEQ ID NO: 1042). In one embodiment, the CPP is p-Antp comprising the sequence RQIKIWFQNRRMKWKK (SEQ ID NO: 1043). In one embodiment, the CPP is hCT(18-32) comprising the sequence KFHTFPQTAIGVGAP-NH2 (SEQ ID NO: 1044). In one embodiment, the CPP is KLA comprising the sequence KLALKLALKALKAALKLA (SEQ ID NO: 1045). In one embodiment, the CPP is AGR, a cancer / tissue-specific CPP for prostate cancer, comprising the sequence CAGRRSAYC (SEQ ID NO: 1046). In one embodiment, the CPP is LyP-2, a cancer / tissue-specific CPP for skin or cervical tumors, comprising the sequence CNRRTKAGC (SEQ ID NO: 1047). In one embodiment, the CPP is REA, a cancer / tissue-specific CPP for prostate, cervical, or breast cancer, and comprises the sequence CREAGRKAC (SEQ ID NO: 148). In one embodiment, the CPP is LSD, a cancer / tissue-specific CPP for melanoma or bone cancer, and comprises the sequence CLSDGKRKC (SEQ ID NO: 1049). In one embodiment, the CPP is HN-1, a cancer / tissue-specific CPP for head and neck squamous cell carcinoma, and comprises the sequence TSPLNIHNGQKL (SEQ ID NO: 1050). In one embodiment, the CPP is CTP, a cancer / tissue-specific CPP for cardiomyocytes, and comprises the sequence APWHLSSQYSRT (SEQ ID NO: 1051). In one embodiment, the CPP is HAP-1, a cancer / tissue-specific CPP for synovial tissue, and comprises the sequence SFHQFARATLAS (SEQ ID NO: 1052). In one embodiment, the CPP is 293P-1, a cancer / tissue-specific CPP against keratocyte growth factor, and comprises the sequence SNNNVRPIHIWP (SEQ ID NO: 1053).

[0204] In a preferred embodiment, the cell membrane-permeable peptide is an immunologically inert cell membrane-permeable peptide, which is advantageous because it does not induce a harmful immune response that may interfere with T cell epitope responses or prevent T cell epitope responses against a CPP other than the intended antigen in the vaccine.

[0205] An example of an immunologically inert cell membrane-penetrating peptide is an at least 8-mer polyarginine. The at least 8-mer polyarginine may be poly-8-arginine or poly-9-arginine. In one embodiment, the at least 8-mer polyarginine may include a tetrazine.

[0206] Preparation and manufacturing of personalized peptide vaccines In one embodiment, the CD4 antigen amino acid sequence is a subject-specific amino acid sequence, and / or the CD8 antigen amino acid sequence is a subject-specific amino acid sequence, e.g., in one embodiment, the CD4 antigen amino acid sequence and / or the CD8 antigen amino acid sequence is specific to a protein specifically expressed in the subject's HLA class I or class II genotype and / or diseased cells.

[0207] In one embodiment, the CD4 antigen, the CD8 antigen, and the cell membrane-penetrating peptide are covalently linked. In one embodiment, the CD4 antigen comprises a highly ranked epitope 9 to 20 amino acids in length. In one embodiment, the CD8 antigen comprises a highly ranked epitope 8 to 15 amino acids in length.

[0208] In one aspect of the present invention, there is provided a pharmaceutical composition comprising a personalized peptide vaccine of an aspect of the present invention, hi one embodiment, the composition further comprises a pharmaceutically acceptable excipient.

[0209] In one aspect of the present invention, a peptide antigen according to an aspect of the present invention or a pharmaceutical composition according to an aspect of the present invention is provided for use as a vaccine. In one embodiment, the CD4 antigen amino acid sequence and the CD8 antigen amino acid sequence are tumor-associated antigens. Tumor-associated antigens may include viral antigens specifically expressed in a patient's tumor, cancer-testis antigens specifically expressed in a patient's tumor, and antigens overexpressed in tumor cells compared to healthy cells. The top-ranked epitopes from these tumor-specific antigens are transported to the cell surface by the patient's HLA, and the "core" of these epitopes is recognized by T cells involved in the cellular immune response.

[0210] Although antigens derived using the methods described herein can be included in other delivery systems, mRNA, DNA, and viral vector-based vaccines must be manufactured industrially and therefore are not affordable and accessible personalized vaccines.

[0211] The personalized VERDI vaccine of the present invention is designed to be prepared for use in pharmacies or clinics. A single peptide antigen in the vaccine can be synthesized and purified as two separate parts and then combined before administration. The present invention also enables the efficient preparation of personalized peptide vaccines with different antigen combinations. The vaccine can contain at least two peptides, each containing or consisting of at least one antigen and a portion of a CPP. When at least two peptides are combined, the vaccine contains at least two antigens separated by an intact CPP between the two antigens. Alternatively, a single peptide antigen in the vaccine can be synthesized and purified and then mixed with one or more excipients before administration. This is an innovative new approach to preparing personalized peptide vaccines for a single immunization in a single individual. This method not only offers superior pharmaceutical quality, but also the same safety as other peptide vaccines, high economic efficiency, ease of production, and time savings.

[0212] Thus, the present invention relates to a method for preparing a personalized vaccine or therapeutic composition, comprising the steps of preparing a first amino acid sequence, wherein the first amino acid sequence comprises a first peptide antigen that induces a CD4 and / or CD8 T cell response and at least a portion of a cell membrane-penetrating peptide, and preparing a second amino acid sequence, wherein the second amino acid sequence comprises a second peptide antigen that induces a CD4 and / or CD8 T cell response and at least a portion of a cell membrane-penetrating peptide, and covalently linking the first and second amino acid sequences to form a personalized vaccine in which the cell membrane-penetrating peptide is positioned between the first antigen and the second antigen.

[0213] In a preferred embodiment, both the first and second antigens are linked (covalently or non-covalently) to four arginines and equipped with tetrazine or norbornene moieties. These chemical modifications allow for covalent conjugation using the inverse electron demand Diels-Alder reaction.

[0214] Because vaccines are personalized and CD4 and CD8 antigens are exchangeable for each vaccine recipient, a general specification for synthetic peptides has been established: the first and second antigens and the lyophilized peptides linked to four arginines can be combined to form the final personalized peptide vaccine product.

[0215] To obtain the conjugated vaccine product, the first and second antigen peptides conjugated to four arginines are dissolved in phosphate-buffered saline (PBS) at a concentration of 1 mg / mL, and the reaction mixture is heated at 40°C for 1–4 hours. The reaction can be followed by the decolorization of the purple color of the tetrazine moiety. This rapid chemical reaction ensures that the two peptide antigens are conjugated in a fast and efficient biorthogonal manner.

[0216] Protocol for in vitro vaccination of HLA-genotyped human subjects In an aspect of the invention, there is provided a method of administering a peptide antigen to a subject, the method comprising administering to the subject a peptide antigen according to an aspect of the invention, or a pharmaceutical composition according to an aspect of the invention. In one embodiment, an antigenic peptide according to an aspect of the invention or a pharmaceutical composition according to an aspect of the invention is administered to a subject in an effective amount.

[0217] In an aspect of the invention, there is provided a method of inducing antigen-specific immunity in a subject, the method comprising administering to the subject a peptide antigen according to an aspect of the invention or a pharmaceutical composition according to an aspect of the invention.

[0218] Directive 2010 / 63 / EU requires the integration of the 3R principles and welfare standards for the treatment of animals in all aspects of the development, manufacturing, and testing of medicines. The 3R principles encourage the reduction, refinement, and replacement of animal testing in medicines development. Inventors have actively pursued alternatives to animal testing and, for the first time, have succeeded in identifying an in vitro human vaccination model that replaces the need for animal testing.

[0219] The use of in vitro human vaccination offers significant advantages over animal models for the development of personalized vaccines. Unlike animal models, in vitro human vaccination takes into account the specific set of HLA class I and class II molecules present in an individual, providing a more accurate prediction of the antigen-specific T cell responses induced by personalized vaccines in that individual. This personalized approach is crucial because T cell responses vary between individuals and are regulated by all human HLA molecules, consisting of six HLA class I molecules and 12 HLA class II molecules. Animal models, such as HLA transgenic mice, poorly mimic the complex genetic background of human subjects and therefore cannot reliably predict antigen-specific T cell responses in individuals.

[0220] The adoption of an in vitro human vaccination model overcomes the limitations of animal models and allows for a more accurate understanding of the impact of personalized vaccines on human T cell responses, which is consistent with the principles of personalized medicine, where individual differences in immune responses play a key role.

[0221] In vitro methods for assessing the efficacy of personalized vaccination in human subjects as an alternative to animal testing are described in detail below.

[0222] In one embodiment, an in vitro method for assessing the efficacy of vaccination of a human subject comprises the following steps. a) preparing monocyte-derived dendritic cells (DCs) from HLA-genotyped individuals; b) incubating DCs with peptide antigens contained in the personalized peptide vaccine; c) co-culturing DCs with autologous PBMCs or isolated T cells in T cell medium; d) Testing antigen-specific CD8 and CD4 T cell responses and / or performing CD8 and CD4 T cell proliferation assays.

[0223] This approach makes it possible to design vaccines that are tailored to an individual's specific immune characteristics.

[0224] Studies on this vaccine composition provide compelling evidence of its potential as the first personalized peptide vaccine platform. This innovative vaccine activates both CD8 and CD4 T cell responses and possesses multiple novel properties that contribute to its efficacy. The inclusion of both CD8 and CD4 T cell antigens allows for personalization based on the unique HLA class I and class II alleles expressed by each individual. Vaccine uptake by dendritic cells ensures efficient antigen processing and presentation, leading to the induction of robust T cell responses. Importantly, the vaccine composition does not contain adjuvants, providing a safe and natural peptide-based vaccine strategy. Overall, this vaccine holds great potential for personalized cancer immunotherapy, offering a new approach to improving patient outcomes and potentially revolutionizing the field of personalized vaccines.

[0225] Treatment with the personalized VERDI vaccine of the present invention The personalized VERDI vaccine represents a novel approach to disease treatment, optimized for each individual. Following administration, the personalized vaccine induces both CD8 and CD4 T cell responses against the target's unhealthy cells. This is important because CD4 helper T cells play a critical role in dendritic cell maturation and the induction of CD8 cytotoxic T cell and antibody responses. The inclusion of both CD8 and CD4 T cell antigens in the vaccine ensures individualization, because T cell responses to antigens derived from unhealthy cells (e.g., cancer or infected cells) are determined by the unique HLA class I and class II alleles expressed by each patient.

[0226] Another important feature of the vaccine is its high intracellular uptake, achieved within just 30 minutes. Efficient uptake of vaccine peptides into dendritic cells allows the vaccine to be processed into epitopes, rapidly saturating HLA class I and class II molecules, leading to excellent antigen presentation to T cells and the induction of CD8 and CD4 T cell responses, respectively.

[0227] The VERDI personalized vaccine does not require the administration of an adjuvant simultaneously or after the vaccine, which is a significant advantage in treating subjects given the known side effects often associated with such adjuvants. Thus, the VERDI personalized vaccine is the safest vaccine platform ever developed.

[0228] The vaccine may be administered intramuscularly, intravenously, intratracheally, intrasynovially, intraperitoneally, subcutaneously, or intraocularly. In one embodiment, the vaccine is administered in an effective amount. In yet another embodiment, the vaccine is administered in a single effective amount. In one embodiment, the subject is a mammal. In yet another embodiment, the subject is a human.

[0229] In one aspect of the invention, a method of treating or preventing a disease is provided, the method comprising administering to a subject a personalized vaccine or therapeutic composition, hi one embodiment, the personalized vaccine or therapeutic composition is administered in combination with an additional therapeutic agent, said administration being simultaneous or sequential.

[0230] In one embodiment, at least two personalized vaccine or therapeutic compositions comprising different peptide antigens are administered simultaneously or sequentially to a subject.

[0231] In one embodiment, the personalized vaccine or therapeutic composition is administered with an adjuvant, either simultaneously or sequentially.

[0232] In one embodiment, the disease is cancer and / or a viral infection and / or an autoimmune disease.

[0233] In a further aspect of the invention, there is provided a method of inducing antigen-specific immunity in a subject, the method comprising administering to the subject a personalized vaccine or therapeutic composition.

[0234] In a further aspect of the invention, there is provided a personalised vaccine or therapeutic composition for use in the prevention or treatment of disease.

[0235] Currently available therapeutic and prophylactic vaccines use repeated administration of vaccines to enhance immune responses. In one embodiment of the present invention, a therapeutic vaccine requires only one dose of a single vaccine to be administered to a patient. A single dose of any one vaccine is preferred because repeated administration of multiple doses can exhaust T cell responses. This is similar to the phenomenon observed in chronic infections and cancers, which limits the body's ability to fight disease, and is one of the reasons why the present invention is superior to current therapeutic vaccines.

[0236] In yet another embodiment, if further vaccine doses are needed, the vaccine is directed to a different target. This is an advantage of personalized vaccines, as they can be tailored to specific target antigens from unhealthy cells. For example, if the unhealthy cells are cancer cells, an initial dose of personalized vaccine can be tailored and administered using antigens expressed in the cancer cells, as determined by transcriptome analysis of a tumor biopsy. If a new tumor is diagnosed in the patient, or if the vaccine fails to completely eliminate the tumor, a new tumor sample can be taken, and a new vaccine can be designed and used to treat the patient. This process can be repeated until the patient is free of detectable tumors. Such flexible and curative treatment as invented herein is only possible with the personalized VERDI vaccine.

[0237] In one embodiment, the disease is cancer, and the cancer is adenoid cystic carcinoma, adrenal tumor, amyloidosis, anal cancer, appendix cancer, astrocytoma, ataxia-telangiectasia, Beckwith-Wiedemann syndrome, bile duct cancer (cholangiocellular carcinoma), Birt-Hogg-Dubé syndrome, bladder cancer, bone cancer (osteosarcoma), brain stem glioma, brain tumor, breast cancer, Carney complex, central nervous system tumor, cervical cancer, colorectal cancer, Cowden syndrome, craniopharyngioma, desmoid tumor, desmoplastic infantile ganglioglioma. , ependymoma, esophageal cancer, Ewing's sarcoma, eye cancer, eyelid cancer, familial adenomatous polyposis, familial malignant melanoma, familial pancreatic cancer, gallbladder cancer, gastrointestinal stromal tumor, germ cell tumor, gestational trophoblastic disease, head and neck cancer, diffuse gastric cancer, leiomyomatosis, renal cell carcinoma, mixed polyposis syndrome, pancreatitis, papillary renal cell carcinoma, juvenile polyposis syndrome, kidney cancer, laryngeal cancer, hypopharyngeal cancer, leukemia, lymphoblastic carcinoma, lymphocytic carcinoma, acute myeloid carcinoma, B-cell prolymphocytic leukemia, hair follicle cell Leukemia, eosinophilic leukemia, Li-Fraumeni syndrome, liver cancer, lung cancer, non-small cell lung cancer, small cell lung cancer, Hodgkin's lymphoma, non-Hodgkin's lymphoma, Lynch syndrome, mastocytosis, medulloblastoma, melanoma, meningioma, mesothelioma, multiple endocrine neoplasia syndrome type 1, multiple endocrine neoplasia syndrome type 2, multiple myeloma, MUTYH (or MYH)-associated polyposis, myelodysplastic syndrome, nasal cavity cancer, paranasal sinus cancer, nasopharyngeal cancer, neuroblastoma, neuroendocrine tumors of the gastrointestinal tract, pulmonary nerve The cancer is selected from the list including endocrine tumors, pancreatic neuroendocrine tumors, neuroendocrine tumors, neurofibromatosis type 1, neurofibromatosis type 2, nevoid basal cell carcinoma syndrome, oral cancer, oropharyngeal cancer, osteosarcoma, ovarian cancer, fallopian tube cancer, peritoneal cancer, pancreatic cancer, parathyroid cancer, penile cancer, Peutz-Jeghers syndrome, pheochromocytoma, paraganglioma, pituitary tumor, pleuropulmonary blastoma, prostate cancer, retinoblastoma, rhabdomyosarcoma, salivary gland cancer, Kaposi's sarcoma, soft tissue sarcoma, skin cancer, small intestine cancer, gastric cancer, testicular cancer, thymoma carcinoma, thymic carcinoma, thyroid cancer, tuberous sclerosis, uterine cancer, vaginal cancer, von Hippel-Lindau syndrome, vulvar cancer, Waldenstrom's hypergammaglobulinemia (lymphoplasmacytic lymphoma), Warner syndrome, Wilms' tumor, and xeroderma pigmentosum. In one embodiment, the personalized VERDI vaccine can be administered in combination with or sequentially with any drugs and biologics at the discretion of the physician.This is similar to prophylactic vaccines, such as the COVID-19 vaccine, which are administered in addition to the patient's current treatment. Specifically, the personalized VERDI vaccine may be administered in combination with at least one anti-cancer therapeutic agent.

[0238] In one embodiment, the anti-cancer therapeutic agent is selected from the list comprising an alkylating agent, a cytotoxic antibiotic, an antimetabolite, an anti-angiogenic agent, a histone deacetylase inhibitor, a hormone, a protein kinase inhibitor, a growth factor, a CAR T cell, a taxane, a topoisomerase inhibitor, a vinca alkaloid, a polyclonal antibody, a monoclonal antibody or fragment thereof, or an immune checkpoint inhibitor.

[0239] The progression of the disease being treated with the vaccine may be monitored after administration by a number of methods familiar to those skilled in the art, including biomarker detection, X-ray scans, MRI scans, CT scans, polymerase chain reaction (PCR), reverse transcription polymerase chain reaction (RT-PCR) testing, branched DNA (bDNA) testing, nucleic acid sequence-based amplification (NASBA) testing, bacterial culture, and other known techniques.

[0240] Individualized VERDI vaccine kit A personalized VERDI vaccine kit may be required to prepare a single dose of personalized peptide vaccine prior to administration.

[0241] Kits containing the personalized vaccine or therapeutic composition optionally further include a pharmaceutically acceptable excipient, and further optionally include instructions for use.

[0242] The kit includes multiple items necessary for preparing a personalized vaccine composition, including two synthetic peptides (each synthetic peptide comprising at least one antigen) and at least a portion of a cell membrane-permeable peptide. Preparation of the personalized vaccine prior to administration includes covalently linking the two synthetic peptides, thereby reconstituting the function of the cell membrane-permeable peptide.

[0243] Additionally, in another aspect of the present invention, kits are provided that include the personalized peptide antigen composition and other components and equipment necessary for the preparation and administration of the personalized vaccine at a pharmacy, such as excipients, adjuvants, syringes, adapters, bacterial filters, plasticware, instructions for use, etc. This list is not limited to these components and includes additional components that will be apparent to one of skill in the art.

[0244] The invention is further defined in the following numbered aspects. 1. A computer-implemented method for identifying, in a subject, at least one antigen that is predicted to induce a T cell response that attacks unhealthy cells in the subject, the method comprising the steps of: obtaining subject data identifying a plurality of major histocompatibility complex (MHC) molecules of the subject; obtaining sequence data for a plurality of epitopes within a protein expressed in the unhealthy cell, each epitope being an amino acid sequence within the protein; obtaining, for each epitope within the plurality of epitopes, a potency score indicative of the likelihood of presenting the epitope on the surface of the identified plurality of MHC molecules; generating a ranked list of epitopes based on the determined potency scores; identifying at least one antigen by selecting at least one epitope that is highly ranked in the ranked list; and Outputting at least one of a ranked list and sequence data for the at least one identified antigen. 2. The method of embodiment 1, further comprising the steps of: selecting a plurality of epitopes that are highly ranked in the ranked list; identifying subtopes of each selected epitope using a directed graph network; and Identifying each subtope common to the selected epitopes. 3. The method of embodiment 2, further comprising identifying at least one antigen by selecting the common subtopes that are highly ranked in the ranked list. 4. The method of aspect 2 or aspect 3, further comprising identifying at least one antigen by selecting the longest common subitope. 5. The method of any one of aspects 2 to 4, further comprising identifying the shortest common subtopes as the target sequence. 6. The method of any one of aspects 1 to 5, wherein obtaining a potency score comprises calculating, for each epitope of the plurality, an epitope weight score by: selecting an epitope from a plurality of epitopes; obtaining, for each identified MHC molecule, a probability score indicating the likelihood that each MHC molecule will transport the selected epitope; and Summing at least some of the probability scores to calculate an epitope weight score for the selected epitope. 7. The method of embodiment 6, wherein the probability score is an eluted ligand score, indicating the likelihood that the epitope will be eluted from a given MHC molecule. 8. The method of embodiment 7, wherein the epitope weight score is calculated by:

number

number

number

number

[0245] All features contained herein may be combined with any of the above aspects in any combination.

[0246] The invention is further illustrated by the following non-limiting examples.

[0247] Examples of personalized vaccines according to the present invention Example 1. Cellular uptake of vaccines composed of model antigens Study Objective: The objective of this study was to demonstrate effective cellular uptake of the vaccine composition.

[0248] Model vaccine design: Two well-characterized T cell antigens were selected as model antigens: CD8 antigen (from pp65 CMV): NLVPMVATV (SEQ ID NO: 1054); CD4 antigen (from tetanus toxoid): QYIKANSKFIGITE (SEQ ID NO: 1055).

[0249] The two peptide vaccines were designed using CD4 and CD8 model antigens: Vaccine 1 (control): KSS-QYIKANSKFIGITE-AAA-LNVPMVATV (SEQ ID NO: 1117), Vaccine 2 (vaccine): KSS-QYIKANSKFIGITE-RRRRRRRR-NVPMVATV (SEQ ID NO: 1118). Synthesis of investigational vaccines: Fluorescently labeled and control vaccines were synthesized for cellular uptake studies. Labeled vaccine (CD4-CD8) Fluorescent substance -KSSQYIKANSKFIGITEAAALNVPMVATV-NH2 (SEQ ID NO: 1057) Labeled vaccine complex (CD4-R8-CD8: Fluorescent substance -KSSQYIKANSKFIGITERRRRRRRRLNVPMVATV-NH2 (SEQ ID NO: 1103) 3. Labeled control vaccine (CD4-R4) Fluorescent substance -KSSQYIKANSKFIGITERRRR-norbornene-NH (SEQ ID NO: 1058) 4. Unlabeled control vaccine conjugate (CD4-R8-CD8): KSSQYIKANSKFIGITERRRRRRRRLNVPMVATV-NH2 (SEQ ID NO: 1056)

[0250] Methods: In this experiment, the uptake of two fluorescently labeled peptide vaccines was examined using cultured cells under a fluorescence microscope. HeLa cells were grown on glass coverslips for 1 day and then treated with the peptide vaccines. Cells were fixed with 4% paraformaldehyde, washed with PBS, and analyzed under a fluorescence microscope. An unlabeled negative control showed no fluorescence, as expected, and is not shown.

[0251] Results: Figure 12b shows the intracellular uptake of the vaccine and control vaccine in HeLa cells after 30 and 120 minutes of incubation. The intracellular uptake was visualized by fluorescence microscopy. The results clearly show a significant increase in vaccine uptake compared to the control vaccine. This increase in intracellular trafficking indicates the potential efficacy of the vaccine in immunization.

[0252] Conclusion: Improved cellular uptake of the vaccine. Experiments revealed noteworthy observations regarding the uptake of the vaccine (CD4-R8-CD8) compared with a peptide vaccine (CD4-CD8) lacking the R8 component. Results showed that the vaccine was rapidly internalized into both the cytoplasm and nucleus of cells within 30 minutes. The vaccine then showed diffuse distribution throughout the cytoplasm and nucleus after 120 minutes. In contrast, the peptide vaccine lacking the R8 component showed little detectable cellular uptake within the same 120-minute time frame and no significant accumulation within the cells. These findings led to the conclusion that cellular uptake of the vaccine was significantly faster than that of the control peptide vaccine.

[0253] The improved cellular uptake of the vaccine suggests increased efficiency in loading peptide antigens into intracellular HLA molecules. This efficient loading process is essential for subsequent T cell-induced immune responses, highlighting the vaccine's potential to effectively induce immune responses against target antigens and generate robust T cell responses.

[0254] Example 2. Cellular uptake of personalized vaccines composed of human papillomavirus (HPV) antigens Study Objective: The objective of this study was to quantify the cellular uptake of an HPV-specific personalized vaccine.

[0255] Vaccine Design: We designed an HPV-specific vaccine that matches a subject's personalized HLA genotype (Figure 13a). This vaccine contains CD8 and CD4 antigens derived from the HPV E7 protein, which is expressed by the high-risk HPV-16 strain. E7 is known to have oncogenic properties and plays a key role in HPV-associated cervical cancer and other HPV-associated malignancies. E7 interacts with host cell proteins, including tumor suppressor proteins such as pRb, leading to cell cycle dysregulation and enhanced cell proliferation. The HPV E7 protein is considered a potential target for therapeutic interventions and diagnostic approaches aimed at combating HPV-associated diseases.

[0256] Synthesis of investigational vaccines: Fluorescently labeled HPV-specific vaccines and a control vaccine (without polyarginine) for cellular uptake studies were synthesized. Labeled HPV-specific vaccine conjugate (CD4-R8-CD8): Fluorescent substance -THVDIRTLEDLLMGTL-RRRRRRRR-RAHYNIVTF-NH2 (SEQ ID NO: 1104) Labeled control vaccine (CD4-CD8): Fluorescent substance -THVDIRTLEDLLMGTL-RAHYNIVTF-NH2 (SEQ ID NO: 1105) Methods: HeLa cells were seeded at a density of 200,000 cells per well in 6-well plates and grown for 1 day. On day 2, cells were treated with peptides at three different concentrations (2.5 μM, 10 μM, and 40 μM) for two different incubation times (15 and 30 minutes). After treatment, cells were detached with trypsin, washed, and resuspended in PBS. Cells were then analyzed using a FACS Canto flow cytometer. A total of 30,000 events were acquired, and the percentage of fluorescent cells and fluorescence intensity (geometric mean and median) were determined by flow analysis.

[0257] Results: Figure 13b shows quantification of cellular uptake of personalized HPV-specific vaccines (blue) compared to controls (red) after 30 minutes of incubation with human cells. Intracellular uptake was assessed by flow cytometry, with fluorescence intensity measured as an indicator of vaccine uptake by cells. Results show significantly higher uptake of the vaccine (blue) compared to the control (red), suggesting its improved intracellular delivery efficiency and potential efficacy in HPV-specific immunization.

[0258] Conclusion: Effective cellular uptake of the HPV-specific personalized vaccine. Experimental results showed significantly higher uptake within 30 minutes of vaccine administration compared to a control peptide vaccine without the R8 component. This observation is particularly important for peptide vaccines because the peptides used in such vaccines are known to degrade rapidly after injection. Experiments revealed that the optimal concentration for vaccine preparation was 40 μM, and that more than 50% of cells internalized the vaccine within 30 minutes. This rapid and efficient cellular uptake is essential in peptide vaccination to ensure proper delivery of antigens to target cells.

[0259] Fluorescence intensity values ​​were analyzed using the median, the central value in a sorted list of intensity values. Using the median allowed for an analysis that took into account the distribution of values ​​and was less sensitive to extreme values ​​or outliers than the mean. This method provided a robust measure of central tendency suitable for distributions of different shapes, including skewed or non-normal distributions. Notably, median analysis revealed that approximately 30% more peptides accumulated in each cell when the vaccine was used compared to the control vaccine. This finding further highlights the improved cellular uptake and potential efficacy of the vaccine in loading epitopes onto patients' HLA molecules.

[0260] Example 3. Cellular uptake of personalized vaccines composed of AKAP-4 tumor-specific antigens Study Objective: The objective of this study was to quantify the cellular uptake of a personalized AKAP-4-specific vaccine.

[0261] Vaccine design: An AKAP-4-specific vaccine was designed to match a specific individual's HLA genotype (Figure 14a). This vaccine contains CD8 and CD4 antigens derived from the AKAP-4 protein, which is expressed in advanced ovarian, lung, colon, pancreatic, and prostate cancers. AKAP-4's specific expression pattern in cancer cells and its lack of expression in healthy cells make it an excellent target for T cell therapy and cancer vaccines. By targeting AKAP-4, these therapeutic approaches may enhance anti-tumor immune responses and potentially improve the prognosis of cancer patients.

[0262] Synthesis of research vaccines: A fluorescently labeled AKAP-4-specific vaccine and a control vaccine (without polyarginine) for cellular uptake studies were synthesized: Labeled HPV-specific vaccine conjugate (CD4-R8-CD8): Fluorescent substance -EEKEIIVIKDTEKKDQS-RRRRRRRR-SQFNVPMLY-NH2 (SEQ ID NO: 1106) Labeled control vaccine (CD4-CD8): Fluorescent substance -EEKEIIVIKDTEKKDQS-SQFNVPMLY-NH2 (SEQ ID NO: 1107)

[0263] Methods: HeLa cells were seeded at a density of 200,000 cells per well in 6-well plates and grown for 1 day. On day 2, cells were treated with peptides at three different concentrations (2.5 μM, 10 μM, and 40 μM) for two different incubation times (15 and 30 minutes). After treatment, cells were detached with trypsin, washed, and resuspended in PBS. Cells were then analyzed using a FACS Canto flow cytometer. A total of 30,000 events were acquired, and the percentage of fluorescent cells and fluorescence intensity (geometric mean and median) were determined by flow analysis.

[0264] Results: Quantification of cellular uptake of the personalized AKAP-4-specific vaccine (blue) compared to the control (red) as shown in Figure 14b. Intracellular uptake was assessed by flow cytometry, measuring fluorescence intensity as an indicator of cellular vaccine uptake. Results show that uptake of the vaccine (blue) was significantly higher than that of the control (red), suggesting its enhanced intracellular trafficking and potential efficacy in AKAP-4-specific immunization.

[0265] Conclusion: Effective cellular uptake of the AKAP-4-specific personalized vaccine. Experimental results further confirmed that vaccines containing AKAP4-specific antigens exhibited significantly higher uptake compared to control peptide vaccines lacking the R8 component. Notably, experiments demonstrated that a concentration of 40 μM was optimal for vaccine formulation, with 45% of cells efficiently ingesting the vaccine within 30 minutes. Interestingly, the HPV-specific vaccine exhibited an even higher cellular uptake rate of 53%, likely due to the CD8 component beginning with "R." This may have resulted in the formation of an R9 bridge, further increasing cellular uptake.

[0266] Example 4. Dendritic cell uptake of vaccines constituting model antigens Study Objective: The objective of this study was to demonstrate improved uptake of vaccine compositions in dendritic cells compared to a control vaccine.

[0267] Vaccine design: Two well-characterized T cell antigens were used as model antigens: CD8 antigen (pp65 CMV-derived) NLVPMVATV (SEQ ID NO: 1054) CD4 antigen (from tetanus toxin) QYIKANSKFIGITE (SEQ ID NO: 1055)

[0268] Synthesis of study vaccines: For cellular uptake studies, fluorescently labeled vaccines and control vaccines (without polyarginine) were synthesized: Labeled model vaccine conjugate (CD4-R8-CD8): Fluorescent substance - KSS-QYIKANSKFIGITE-RRRRRRRR-LNVPMVATV (SEQ ID NO: 1108) Labeled control vaccine (CD4-CD8): Fluorescent substance - KSS-QYIKANSKFIGITE-AAA-LNVPMVATV (SEQ ID NO: 1109) Labeled control vaccine (CD4-R4): Fluorescent substance -KSSQYIKANSKFIGITERRRR-norbornene-NH (SEQ ID NO: 1110)

[0269] Methods: Monocyte-derived dendritic cells were generated by separating monocytes from peripheral blood using density gradient centrifugation. Monocytes were then cultured in complete medium supplemented with IL-4 and GM-CSF for [specify time period] to promote differentiation into dendritic cells. After [specify time period], differentiated dendritic cells were directly cultured on glass coverslips. Cells were treated with the peptide vaccine and fixed with 4% paraformaldehyde. After fixation, cells were washed with PBS and analyzed by fluorescence microscopy.

[0270] Immature DCs were treated in duplicate with 2.5 or 10 μM of fluorescent peptide vaccine. Two wells of untreated DCs served as negative controls. 0.3 mL (2 x 10 5 The cells were removed and washed three times with 5 mL of PBS to remove the peptide and fluorescent dye bound to the surface. The washed cells were then resuspended in 3 mL of PBS (approximately 2 x 10 5 cells) and resuspended in 1.5 mL (1 x 10 5 Cells) were centrifuged in a tabletop centrifuge and spun in 0.1 mL (approximately 1 x 10 cells) of Live Cell Imaging Solution (Invitrogen). 6 The cells were resuspended in 0.1% CO₂Cl (0.1% cells / mL) and imaged.

[0271] Results: Figure 15a shows the intracellular uptake of the vaccine and control vaccine in dendritic cells after 120 minutes of incubation. The intracellular uptake was visualized by fluorescence microscopy. The results show a significant increase in uptake of the vaccine compared to the control vaccine. This vaccine transport into dendritic cells was similar to that in HeLa cells, suggesting the potential for immune effects of the vaccine.

[0272] Conclusion: Effective uptake of vaccine by human primary dendritic cells. The vaccine showed excellent uptake not only in HeLa cells but also in human primary dendritic cells, suggesting the following mechanism of immune response induction: When the injected vaccine is taken up by dendritic cells, vaccine-derived epitopes saturate the patient's HLA molecules. These epitopes are presented at high density on the surface of dendritic cells and can induce both CD8 and CD4 T cell responses. Efficient vaccine uptake by dendritic cells ensures the generation of strong T cell responses.

[0273] The efficient uptake of vaccines by dendritic cells not only clarifies their efficacy but also presents the challenging possibility of injecting natural peptides without adjuvants in future vaccine strategies. The unique immunological capabilities of peptides and dendritic cells hold great potential to revolutionize the field of personalized vaccines. This discovery opens the door to exploring new approaches to exploit the unique properties of vaccines to maximize safety and efficacy.

[0274] Example 5. Personalized VERDI vaccine for patients with advanced ovarian leiomyosarcoma The cancer patient had advanced ovarian leiomyosarcoma (OLMS) with lung metastases and was treated with surgical resection and checkpoint inhibitor (OPDIVO, nivolumab)-based immunotherapy.

[0275] Sequencing of patient tumor and blood samples: Paraffin-embedded tumor samples (FFPE) were obtained from surgical tumor resections. Whole-transcriptome sequencing from patient FFPE samples was performed by two contractors: Lexogen (Vienna Biocenter, AT) and Ibioscience (Pecs, HU). Complete four-digit HLA genotyping from blood samples was performed at the Universitatsklinik fur Transfusionsmedizin und Zelltherapie (AKH, Vienna, AT). [Table 15]

[0276] Results of transcriptome analysis to identify vaccine targets in tumors: The table below summarizes the results of transcriptome analysis of tumor biopsies from OMLS patients, identifying 12 distinct proteins. These targets are some of the overexpressed proteins used so far, primarily cancer-testis antigens, and have previously been demonstrated to be safe for cancer vaccine therapy in the treatment of cancer patients. [Table 16]

[0277] The relative expression of vaccine target proteins is determined by analyzing the transcriptome of patient tumor samples. Expression levels are quantified as the TARGET's TPM (transcripts per million total transcripts) relative to the housekeeping gene GAPDH (glycerol-3-phosphate dehydrogenase) in the tumor sample. The calculation formula is (TARGET's TPM) / (GAPDH's TPM x 1000). This relative expression value provides a clue to the abundance of the TARGET in the tumor compared to the reference gene GAPDH.

[0278] Results of predictive immunodiagnostic testing to support VERDI vaccine design: A predictive diagnosis of the epitope repertoire involved in tumor-specific immune responses in OMLS patients is shown in Figure 15b. This predictive diagnosis highlights a key step in our method. Utilizing our system and computer-implemented method, trained and validated with clinical data, the selection of top-ranked epitopes emerges as a critical step. This predictive diagnosis comprehensively evaluates all possible epitopes derived from target proteins expressed in a patient's tumor. Its significance lies in its role as the foundation for the precision design of personalized peptide vaccines by VERDI. This technology forms the foundation of our innovative personalized vaccine approach.

[0279] Identification of candidate peptides for vaccine design: Figure 16a illustrates the peptide antigen selection method, consisting of a set of highly ranked epitopes presented on the cell surface by the patient's HLA, a key step in creating a potent VERDI vaccine. Antigen selection is based on identifying peptides containing overlapping, highly ranked sets of epitopes presented on the cell surface by class I and class II HLA. Each carefully crafted VERDI vaccine is strategically designed to target specific proteins expressed within the patient's tumor cells and elicit strong CD8 and CD4 T cell responses.

[0280] Once identified, selected peptides undergo rigorous safety testing, a critical defense against potential autoimmunity. Peptides that do not meet strict safety standards or contain sequences that may interfere with automated synthesis are carefully removed from the candidate vaccine list. This curation method ensures that only the most promising and structurally stable candidates advance to the next stage of development, exemplifying our unwavering commitment to patient safety and our pursuit of breakthrough personalized vaccine development.

[0281] [Table 17]

[0282] Table Notes: Most VERDI vaccines target cancer-testis antigens (CTAs). CTAs are proteins present in both testicular cells and certain cancer cells. These antigens are not typically expressed in normal adult tissues other than the testis. However, their expression is activated in cancer cells, making them potential targets for cancer vaccines. CTAs induce a strong immune response against cancer and help the immune system recognize and attack tumor cells. The personalized vaccines designed by VERDI Solutions target CTAs whose expression in patients' tumor cells was confirmed by transcriptome analysis of paraffin-embedded tumor samples. Other targets are overexpressed antigens previously investigated in multiple clinical trials in cancer patients.

[0283] The relative expression of vaccine target proteins is determined by transcriptome analysis of patient tumor samples. Expression levels are quantified as the TARGET's TPM (transcripts per million total transcripts) relative to the housekeeping gene GAPDH (glycerol-3-phosphate dehydrogenase) in the tumor sample. The calculation formula is (TARGET's TPM) / (GAPDH's TPM x 1000). This relative expression value provides a clue to the abundance of the TARGET in the tumor compared to the reference gene GAPDH.

[0284] CD4 and CD8 are indicators of the efficacy of tumor-specific T cell responses induced by personalized peptide vaccines. CD4 helper T cells and CD8 cytotoxic T cells play important roles in orchestrating a strong and effective immune response against cancer. The reported numbers represent the diversity of T cell clones that can be activated by the peptides included in the vaccine. These peptides are fragments of target proteins already expressed in patient tumor samples. By including these specific peptides in the vaccine, the inventors aim to stimulate a broad T cell response in which CD4 and CD8 T cells work in tandem.

[0285] Summary of expected safety and efficacy of the method of the present invention: The 15 personalized vaccines designed for AT-VERDI001 patients contain peptides with the potential to induce CD8 and CD4 T cell responses in patients. Safety by design: (1) All peptides are part of target proteins already expressed in patient tumor samples. (2) We rule out autoimmunity by testing whether 8-amino acid fragments of candidate peptides show homology to other human proteins. (3) Peptide vaccines with different sequences have demonstrated excellent safety and tolerability profiles in thousands of patients, including cancer patients. In a meta-analysis involving 500 patients, only 1.2% of vaccinated patients experienced vaccine-related serious adverse events. Efficacy by design: (1) Target selection using transcriptome analysis of patient tumor samples ensures that target antigens are most likely expressed in tumors, even when RNA expression does not directly correlate with protein expression. (2) VERDI's proprietary predictive diagnostic test ensures that vaccine peptides induce tumor-specific immune responses. (3) Each peptide vaccine has the potential to induce multiple CD8 and CD8 T cell clones against the target protein, thereby increasing the probability that a strong immune response can attack the patient's tumor cells.

[0286] The active pharmaceutical ingredient of the personalized vaccine therapy designed by VERDI for patient AT-VERDI001 comprises a group of 15 peptides strategically designed to induce CD8 and CD4 T cell responses within the patient's immune system.

[0287] Safety is a top priority throughout the entire methodology and is achieved through thorough design practices including: 1. All peptides that make up these vaccines are derived from fragments of target proteins already present in the patient's tumor sample, ensuring a match between the vaccine components and the patient's own biological makeup. 2. To further guard against the possibility of autoimmunity, a rigorous screening protocol will be implemented, which will entail assessing whether the eight amino acid fragment of the candidate peptide shows homology to other human proteins, eliminating the risk of an unintended immune response. 3. Extensive clinical validation has strengthened a solid foundation of safety. Thousands of patients, including those undergoing cancer treatment, have been vaccinated with peptide vaccines of different sequences, demonstrating an overall excellent safety and tolerability profile. (43) A comprehensive meta-analysis of 500 patients found that the incidence of vaccine-related serious adverse events was only 1.2%. (44)

[0288] The effectiveness of our approach is elaborated into all aspects of vaccine design: 1. Targets are precisely selected through transcriptome analysis of patient tumor samples. This method significantly increases the likelihood that the selected target antigens are expressed within the tumor environment, bridging the gap between RNA and protein expression dynamics. 2. VERDI's proprietary predictive diagnostic test is a key feature that validates the vaccine peptide's ability to induce tumor-specific immune responses. 3. Individual peptide vaccines have the unique ability to stimulate multiple CD8 and CD4 T cell clones, increasing the likelihood of a strong immune response that will mount an effective attack against the patient's tumor cells.

[0289] In essence, our personalized vaccine therapy combines safety and efficacy through a multifaceted design strategy that leverages advanced scientific knowledge to increase the likelihood of successful immune intervention.

[0290] Example 6. Personalized peptide vaccine design for patients from Murcia with metastatic signet ring cell adenocarcinoma Signet ring cell carcinoma (SRCC) is an extremely rare and aggressive variant of adenocarcinoma that presents unique diagnostic and therapeutic challenges. It typically arises in the gastrointestinal tract and is characterized by the presence of cells containing abundant intracytoplasmic mucin, with extruded nuclei at the periphery, creating a "signet ring" appearance. Signet ring cell carcinoma is more aggressive than other histologic subtypes of colorectal cancer and is typically detected at an advanced stage due to its endophytic / infiltrative growth pattern. In this report, we describe the diagnostic challenges and therapeutic considerations associated with this rare and aggressive malignancy. Because currently no options exist to augment tumor-specific T cell responses in patients with SRCC, this case report highlights the use of personalized VERDI vaccine as an adjuvant treatment for this malignancy.

[0291] Clinical History: The patient was a 70-year-old man with a history of type 2 diabetes mellitus that was well controlled with metformin therapy, a history of past smoking, and a diagnosis of prostatic syndrome while on alpha-blocker therapy, cholelithiasis, and left renal ureteral stones. In February 2023, he was admitted to the hospital for evaluation of an osteolytic lesion causing pain in the sacrococcygeal region and left lumbar region. Laboratory tests at the time of admission were unremarkable.

[0292] Tumor History: The patient was diagnosed with stage IV signet-ring cell adenocarcinoma (SRCAC) of unknown primary site and multiple bone metastases. He was admitted to the hospital on February 22, 2023, for investigation of an osteolytic lesion. Two months prior, the patient had presented with pain in the sacrococcygeal region, radiating to the left lumbar region, worsening with standing and improving with rest. Laboratory tests revealed elevated blood glucose (118), alkaline phosphatase (711), CEA (2.8), and Ca₂₁₀ (2,199). The PSA level was within normal ranges. PET-CT scan revealed hypermetabolic lesions in the bony structures and prostatic parenchyma, as well as adrenal nodules. Biopsies of multiple bone lesions confirmed stage IV signet-ring cell carcinoma with multiple metastases. Gastroscopy and colonoscopy revealed chronic gastritis, duodenitis, and colonic diverticulosis, but no tumor was found. Biopsy revealed moderate atrophic chronic gastritis, severe inflammatory activity due to Helicobacter pylori, and extensive intestinal and gastric metaplasia. Endoscopic echocardiography revealed no significant abnormalities except for gallbladder stones. Molecular testing revealed no loss of MLH1 or MSH2 expression, and the patient was HER2-negative, with negative CPS PDL-1 and TPS (0). DPyD normal metabolizer.

[0293] In March 2023, the patient underwent a diagnostic laparoscopy and peritoneal implant biopsy, which confirmed peritoneal adenocarcinoma with signet ring cells. Molecular analysis with the OncoDeep test did not detect specific genetic or molecular changes in the tissue sample that would make the cancer suitable for targeted therapy. The lack of these changes limited targeted therapy options for this patient.

[0294] Treatment: In March 2023, the patient began FOLFOX6m as first-line treatment. In May 2023, after four cycles of treatment, PET-CT revealed carcinomatosis of the omentum and mesentery. No obvious hypermetabolic lesions were observed, but a mild metabolic increase was observed. The CA19.9 tumor marker increased from 2,199 to 4,390 over a two-month period, suggesting disease progression.

[0295] On May 29, 2023, the patient began second-line treatment consisting of paclitaxel and ramucirumab for peritoneal progression. Because there was no commercially available treatment to cure the patient, the medical team contacted VERDI Solutions and requested a personalized vaccine design. After obtaining the patient's written consent, VERDI Solutions obtained the patient's HLA genotype and clinical data and designed the personalized vaccine. The patient from Murcia then received 10 doses of personalized VERDI vaccine therapy in combination with paclitaxel and ramucirumab.

[0296] Scheduling the VERDI vaccination as adjuvant therapy presented a therapeutic challenge. Patients were pre-administered with 20 mg of dexamethasone to mitigate potentially life-threatening infusion-associated reactions before paclitaxel and ramucirumab administration. Dexamethasone may suppress the immune system by reducing inflammation, raising concerns about its impact on vaccine efficacy. After careful evaluation, the medical team decided to start paclitaxel and ramucirumab treatment on July 3, 2023. The first four doses of the vaccine were administered on July 13, 2023, and the remaining six doses on July 27, 2023. Chemotherapy was delayed until August 1, 2023, to optimize the treatment sequence.

[0297] Personalized VERDI Vaccines: VERDI's strategy for cancer patients involves designing at least 10 personalized vaccines that match an individual's immunogenetic profile and tumor characteristics. To enable this, the inventors developed and validated a predictive diagnostic test that predicts the top-ranked epitopes from proteins involved in individual T-cell responses. Leveraging the results of this test, VERDI can help physicians design personalized peptide vaccines that are more likely to destroy a patient's tumor cells.

[0298] The VERDI test utilizes two inputs, obtained from sequencing of blood and tumor specimens, to predict a patient's epitope repertoire: four-digit HLA genotype and tumor-specific proteins that will trigger a tumor-specific immune response. For the Murcia patients, HLA genotypes were readily available: HLA-A*24:02, HLA-A*31:01, HLA-B*08:01, HLA-B*51:01, HLA-C*05:01, HLA-C*07:01, HLA-DPA1*01:03, HLA-DPA1*01:03, HLA-DPB1*03:01, HLA-DPB1*04:01, HLA-DQA1*01:03, HLA-DQA1*05:01, HLA-DQB1*02:01, HLA-DQB1*06:03, HLA-DRB1*03:01, HLA-DRB1*13:01, HLA-DRB3*01:01, and HLA-DRB3*01:01. However, because tumor samples were not available from the patient, tumor-specific protein selection was based on data obtained from peer-reviewed literature. Given the patient's active Helicobacter pylori infection and the presumed gastric origin of the tumor, this study focused on identifying cancer-testis antigens (CTAs) associated with Helicobacter pylori-induced gastric cancer. We identified KK-LC-1 as a particularly promising target, as its expression is observed in approximately 80% of H. pylori-associated gastric cancers (45). Furthermore, based on reports from the same paper, we also included GTGIB and SSX4, CTAs expressed in 24% and 16% of H. pylori-positive gastric cancers, in our selection. Furthermore, five additional CTAs were selected based on comprehensive transcriptome analysis of 375 gastric cancer specimens (46). This selection procedure allowed us to construct a panel of tumor-specific proteins potentially expressed in the patient's tumor.

[0299] In the second step, HLA and CTA data are input into the VERDI Test to predict the epitopes most likely to induce CD8 and CD4 T cell responses against signet ring cell adenocarcinoma (SRCAC) in Murcia patients. The VERDI Test provides a selection of predicted immunogenic epitopes, including their sequences and the predicted potency of the T cell response they may provoke (typically ranging from 0.4 to 2). This information can be represented in a comb diagram (Figure 16b), visualizing the predicted antigen-specific T cell responses in Murcia patients. Notably, this shows the variation between immunologically active and inactive regions within the protein, demonstrating that immunologically active regions independently induce CD4 and CD8 responses.

[0300] In the third step, we design a personalized VERDI vaccine based on the results of the VERDI test. Our approach balances efficacy and safety through intentional design. To ensure vaccine efficacy, peptides included in the VERDI vaccine are selected to induce strong CD8 and CD4 T cell responses against antigens expressed in Murcia patient tumors. Regarding safety, care was taken to eliminate any possibility of autoimmunity. Because it is impossible to control the safety and efficacy of the personalized VERDI vaccine in other patients, our vaccine design includes peptides most likely to induce strong responses in Murcia patient T cells, some of which are likely to target tumors and others that are unlikely to target the patient's tumors. As a positive control, we selected two peptides from KKLC1 because this antigen was highly likely to be expressed in Murcia patient tumors. As a negative control, we designed two similar immunogenic peptides from SSX4 because SSX4 was unlikely to be expressed in H. pylori-positive SRCAC. Considering tumor heterogeneity, we included immunogenic peptides from six additional CTAs that are likely to be expressed in gastric cancer. The vaccine peptide selection criteria aimed to maximize the number of epitopes recognized by the same TCR. Next, we subjected all candidate vaccines to immunogenicity studies using digital twins to exclude vaccines that may induce autoimmunity. As a result, we designed 10 peptide vaccines for Murcia patients, each containing a single peptide of 18 to 29 amino acids in length (Figure 16b).

[0301] The table below lists the target antigens, peptide sequences, and immunological properties of the personalized VERDI vaccine administered to Murcia patients as adjuvant immunotherapy for H. pylori -positive signet ring cell adenocarcinoma. [Table 18]

[0302] The inventors decided to utilize a peptide vaccine platform due to its excellent safety and tolerability profile, which has been demonstrated in thousands of patients. Furthermore, the personalized VERDI vaccine, which contains a peptide solution emulsified in Montanide adjuvant (Seppic, France), can be prepared by a pharmacist or physician, facilitating patient access to this treatment. Safety Findings: Treatment, including vaccination, was well tolerated with no adverse events observed. However, on September 1, 2023, a thrombotic event at an atypical site without a history of venous catheterization posed a diagnostic challenge. This event was characterized by marked dilation of the left jugular vein, with no clear contrast filling from the cranial exit of the jugular foramen to the bifurcation of the left brachiocephalic vein. This was accompanied by consistent peripheral inflammatory changes throughout the neck, multiple reactive lymphadenopathy, and significant bulging of the ipsilateral sternocleidomastoid muscle, causing a mild depression of the ipsilateral piriform sinus.

[0303] After reviewing the patient's medical history and literature, we hypothesized that the thrombosis was likely related to the patient's underlying cancer (SRCAC) or angiogenesis inhibitor treatment and was unrelated to peptide vaccination. This hypothesis was supported by the patient's platelet count exceeding 350,000 and meeting the criteria for prophylactic anticoagulation therapy according to the Khorana score. SRCAC has been reported to occasionally present initially with internal jugular vein (IJV) thrombosis due to a hypercoagulable state associated with malignancy. Furthermore, the angiogenesis inhibitor ramucirumab has been associated with a somewhat higher incidence of grade 3 or higher arterial thromboembolism. However, these events were primarily arterial in nature, whereas the patient's thrombosis was venous. Regarding vaccinations, thrombotic side effects are extremely rare and are primarily associated with adenoviral vector-based vaccinations and thrombocytopenia. Therefore, it is unlikely that the peptide vaccine was involved in the thrombosis.

[0304] Immunological Results: A key element of this case is the immune response observed after the personalized VERDI vaccine was administered to the Murcia patient. T cell responses were assessed with the QuantiFERON® ELISA (QFN) assay, designed to detect human interferon-γ (IFN-γ) in plasma after overnight stimulation of whole blood cells with peptide (Qiagen, USA). The QFN test has high sensitivity, with a detection limit as low as 0.065 IU / ml. The amount of IFN-γ serves as a measure of VERDI vaccine-specific T cell responses, including both CD4 and CD8 T cell responses (49,50).

[0305] T cell responses were detected against four specific peptides from among the ten VERDI vaccines administered (Figure 17). In particular, the positive control peptides C1 and C2 derived from the KK-LC-1 antigen served as positive controls and consistently induced T cell responses after a single vaccination. Meanwhile, the two negative control VERDI vaccines targeting SSX4 generated responses near the detection limit of the QFN test. These results strongly suggest that KK-LC-1-expressing cells are present in the tumor cells of Murcia patients, potentially stimulating KKLC1-specific T cells in vivo and causing a booster effect. Meanwhile, the lack of detectable responses against the two SSX4 peptides provides strong evidence that this target is not expressed in the patient's tumor. Furthermore, positive T cell responses against MAGEA4 and MAGA3 were observed, indicating the potential of these vaccines to enhance immune responses against tumor cells. The incorporation of internal positive and negative controls in our study not only revealed potential tumor targets but also supported our hypotheses regarding the safety, immunogenicity, and efficacy of different VERDI vaccines as adjuvant therapy for SRCAC.

[0306] Clinical Progress: The patient's clinical condition remained favorable throughout the treatment period. A significant decrease in carbohydrate antigen 19-9 (CA19-9) levels during second-line treatment (from 4,390 ng / mL on May 29, 2023 to 372 ng / mL on September 7, 2023) indicates a positive response to the administered cancer therapy (Figure 18). This significant decrease in CA19-9, a known biomarker in pancreatic and gastrointestinal cancers, strongly suggests effective disease control and reduced tumor activity. These data clearly demonstrate the importance of regular biomarker monitoring to track disease progression and evaluate treatment efficacy.

[0307] We received the results of a PET-CT scan for tumor response assessment. The scan, performed on September 19, 2023, was compared with the previous scan on July 17, 2023. The scan confirmed known peritoneal carcinomatosis, infiltrating the omentum and mesentery. Metabolic evaluation, although partially limited by heterogeneous hypermetabolism of the intestine, was stable compared with the previous scan. Of note, no pathological lesions were found in the liver, no changes were observed in the nodular thickening of the bilateral adrenal glands, and prostate enlargement was observed but without obvious hypermetabolic lesions. Abdominal and pelvic lymphadenopathy was not observed, and no suspicious nodules were found in the lungs. Regarding bone involvement, mild metabolic increases were noted in some lesions compared with the previous study. No macroscopic malignant disease was found in other parts of the body.

[0308] A noteworthy observation emerged from the discrepancy between PET-CT and biomarker trends. While the former showed a stable trend despite some local fluctuations, the latter showed dramatic improvement, including an improvement and significant reduction in tumor-related bone damage (as indicated by a 35-fold decrease in alkaline phosphatase) and a decline in gastrointestinal tumor activity (a nearly 10-fold decrease in CA19-9) (Figure 18). The decrease in alkaline phosphatase began after second-line chemotherapy and continued after vaccination. In particular, the turning point in the gastrointestinal antigen CA19-9 was particularly noteworthy after vaccination. Given these changes, we considered the possibility of pseudoprogression. Pseudoprogression is a phenomenon that mimics disease progression but is actually a hypermetabolic "flare" caused by T-cell tumor infiltration. It is noteworthy that patients who experience pseudoprogression often demonstrate objective responses during subsequent treatment phases.

[0309] Next, we aimed to verify pseudoprogression through pathological analysis. If tumor tissue was identified, we adopted a strategy of designing additional VERDI vaccines using targets selected from tumor transcriptome analysis. For sample acquisition, we chose CT-guided aspiration of the peritoneal lesion, which we considered more efficient than biopsy from bone sites due to the presence of bone metastases. The administration of ramucirumab, which interferes with proper wound healing, presented a notable challenge in the form of a complex interventional procedure. Two right peritoneal biopsies (RPBs) were obtained on October 4, 2023. Pathological examination of the biopsies revealed 80% signet ring cells, prominent mucus secretion, and a lack of lymphocytes on hematoxylin staining. Peripheral blood samples collected on the same day also showed no vaccine-induced T cell responses, suggesting the absence of specific target cells recognizable by T cells within the tumor (Figure 17).

[0310] Transcriptome analysis of tumor biopsies (see table below) confirmed that no tumor cells expressed any of the CTAs targeted by the personalized vaccine, despite the 96% probability that at least one target CTA is expressed in gastric cancer cells. These results strongly suggest that vaccine-induced T cells killed target-specific cells in the tumors of Murcia patients. These results provide evidence of the efficacy of the personalized VERDI vaccine. A single dose of the personalized VERDI vaccine can completely destroy target tumor cells in patients with metastatic cancer.

[0311] [Table 19]

[0312] Using transcriptome analysis of the Murcia patient, we identified proteins (cancer-testis antigens) expressed in the patient's tumor cells but not in healthy cells. All predicted epitopes, 8-20 amino acids long, were generated from these proteins and used in the design of a second-generation personalized vaccine according to the present invention. The following table shows the characteristics of the second-generation personalized vaccine. These antigens vary in length and amino acid sequence. All selected peptide antigens contain multiple highly ranked epitopes (13-66) presented on the cell surface by the patient's HLA class I and class II molecules. The range of potency scores for the selected epitopes is shown. The epitopes presented by the patient's HLA class I and class II molecules have been shown to induce CD8 and CD4 responses, respectively. As described herein, the potency of each epitope in the vaccine was calculated and presented to inform physicians of the predicted efficacy of the personalized vaccine in recipients. [Table 20]

[0313] Conclusions: This study reveals a breakthrough in the treatment of metastatic SRCAC, an extremely rare and aggressive subtype of colorectal cancer with a 5-year survival rate of 36% (51). The complexity of the Murcia patient's case not only illuminates the multifaceted nature of this disease, but also introduces a groundbreaking approach to personalized cancer medicine with the VERDI system. The VERDI Test is a novel diagnostic tool that serves as the foundation for this innovation. This test not only predicts peptide fragments that will elicit strong CD4 and CD8 T-cell responses, but also enables clinicians to design personalized peptide vaccines (VERDI vaccines) designed to destroy each patient's unique tumor cells. Incorporating the personalized VERDI vaccine into patient treatment plans for this aggressive adenocarcinoma represents an important advancement.

[0314] Summary: In conclusion, findings from these studies highlight the potential of this vaccine as an innovative approach to personalized vaccination. Its ability to induce CD8 and CD4 T cell responses, its personalized properties, and its efficient dendritic cell uptake position it as a promising strategy in the field of personalized cancer immunotherapy. By harnessing the power of natural peptides and dendritic cells, this vaccine offers new possibilities to improve patient outcomes and revolutionize the landscape of cancer treatment.

[0315] Example 7. Example of a method for manufacturing the VERDI personalized vaccine The VERDI vaccine can be produced from "raw peptide materials" using an inverse electron demand Diels-Alder (IEDA) bond. Lyophilized raw peptide materials bearing a tetrazine or norbornene moiety are dissolved in phosphate-buffered saline (PBS) at a concentration of 1 mg / mL. The reaction mixture is heated at 40°C for 1 hour to obtain the conjugated VERDI vaccine product (API). The VERDI vaccine is then filter-sterilized and loaded into syringes for subcutaneous injection. After the reaction, the purple color of the tetrazine moiety is decolorized. This rapid chemical reaction ensures that the two peptide antigens are conjugated in a fast and efficient biorthogonal manner. [Table 21]

[0316] The raw material is a lyophilized peptide provided to the pharmacy by VERDI Solutions GmbH along with a Certificate of Analysis and detailed analytical documentation. All raw materials are of the highest quality and free of TSE contamination (supplier name and address) All solvents used in the synthesis and purification steps are Ph.Eur. reagent grade and inhibit bacterial growth. The water used in the purification step is Type I purified water.

[0317] Synthesis and purification of raw peptide materials The peptide chain was manually extended on a 0.1 mmol scale using a Rink amide linker and Fmoc protection scheme on TentaGel® RAM resin (0.19 mmol / g). The coupling was performed in two steps. 1. Three equivalents of an Fmoc-protected amino acid, three equivalents of the uronium coupling agent O-(7-azabenzotriazol-1-yl)-N,N,N',N'-tetramethyluronium hexafluorophosphate (HATU), and six equivalents of N,N-diisopropylethylamine (DIPEA) are used in N,N-dimethylformamide (DMF) as a solvent and shaken for three hours. 2. The second coupling is carried out with 1 equivalent of amino acid, 1 equivalent of HATU, and 2 equivalents of DIPEA.

[0318] After the coupling step, the resin was washed three times with DMF, once with MeOH, and three times with DCM. No truncated sequences were observed under these coupling conditions. Deprotection was performed in two steps using 2% DBU and 2% piperidine in DMF, with reaction times of 5 and 15 minutes. Cleavage was performed using TFA / water / dl-dithiothreitol (DTT) / TIS (90:5:2.5:2.5) at 0°C for 1 hour.

[0319] Purification was performed by RP-HPLC using a Phenomenex Luna C18 100Å 10 μm column (10 mm × 250 mm). The HPLC equipment was from JASCO. Solvent system: 0.1% TFA in water; 0.1% TFA in 80% acetonitrile in water; a linear gradient was used for 60 min at a flow rate of 4.0 mL / min, with detection at 206 nm.

[0320] The purity of the fractions was determined by analytical RP-HPLC-MS on an Agilent 1200 HPLC system equipped with a Bruker HCT II ion trap MS and a Phenomenex Luna C18 100 Å 5 μm column (4.6 mm x 250 mm). Pure fractions were pooled and lyophilized. The purified peptides were characterized by MS, a Bruker HCT II ion trap mass spectrometer equipped with an electrospray ion source.

[0321] While several preferred embodiments of the present invention have been shown and described, it will be apparent to those skilled in the art that various changes and modifications can be made thereto without departing from the scope of the invention as defined in the appended claims. The invention is not limited to the details of the above-described embodiments. The invention extends to any new or novel combination of features disclosed in this specification (including the accompanying claims, abstract and drawings), or any new or novel combination of steps of the methods or processes disclosed herein.

[0322] Attention is drawn to all papers and documents related to this application that are filed contemporaneously or prior to this application and that are publicly available herewith, the contents of which are incorporated herein by reference in their entirety.

[0323] All features disclosed herein (including the accompanying claims, abstract, and drawings) and / or all steps of any method or process disclosed herein may be combined in any combination, except combinations in which at least some of such features and / or steps are mutually exclusive. Each feature disclosed herein (including the accompanying claims, abstract, and drawings) may be replaced by an alternative feature serving the same, equivalent, or similar purpose, unless expressly stated otherwise. Thus, unless expressly stated otherwise, each disclosed feature is merely one example of a generic series of equivalent or similar features.

[0324] Example 8: Industrial applicability of the invention for the treatment of cancer patients Our invention offers a new approach to personalizing cancer treatment that challenges the traditional pharmaceutical model. Our innovative approach involves developing personalized vaccines that are tailored to each cancer patient's unique genetic profile and tumor antigen characteristics.

[0325] Our process begins with the collection of tumor biopsy-based transcriptome sequencing and four-digit HLA genotype data, which are readily commercially available in Europe, facilitating smooth implementation. Leveraging these patient-specific datasets, VERDI provides physicians with cloud-based, tumor-specific T cell response predictive diagnosis and personalized peptide vaccine design services (Figure 19a).

[0326] VERDI's vaccine design software serves as a clinical decision support tool, providing physicians with evidence-based, personalized vaccine recommendations. Objective evidence is generated by our clinically validated predictive diagnostic tools from patients' HLA genotypes and tumor mRNA data. It supports decision-making by providing a ranked list of personalized vaccines based on patient-specific data, enabling healthcare professionals to integrate their expertise with other treatment options to make informed choices in patient care.

[0327] The formulated peptide vaccines are dispensed at pharmacies, ensuring patient accessibility and affordability. Our "Efficacy by Design" approach ensures that personalized vaccines designed with VERDI have a high probability of inducing tumor-specific T cell responses, enhancing the efficacy of traditional treatments and increasing the likelihood of long-term remission.

[0328] Our proprietary and breakthrough technology includes predictive diagnostic software that leverages machine learning. This software is trained and validated on clinical data to accurately predict individual T cell responses and generate a ranked list of epitopes likely to elicit an immune response. As a result, our personalized VERDI vaccine is designed based on objective clinical evidence, enabling an optimized approach to each individual's immune response (Figure 19b).

[0329] To improve the quality of input data for predictive diagnosis, our in-house developed transcriptome analysis software processes RNASEQ data extracted from tumor biopsies and appropriately ranks a patient's tumor-specific antigens according to their relative expression levels, allowing us to consider well-expressed tumor antigens for the design of the VERDI vaccine (Figure 20a).

[0330] Authorized users can securely access predictive diagnostic and vaccine design results through our user-friendly, web-based application. The platform not only ensures efficiency, but also adheres to GDPR and other relevant regulations, making data security a top priority. Our commitment to privacy and compliance is seamlessly integrated into every aspect of the user experience, providing our customers with a reliable and responsible service (Figure 20b).

[0331] Current innovative medical products are a) only available to a subset of eligible patients and b) accessible only through clinical trials of drugs centrally manufactured by large pharmaceutical companies.Clinical trials of "state-of-the-art one-size-fits-all drugs" are conducted in patient populations most likely to respond positively to the investigational drug.

[0332] Our business innovation is that all cancer patients can already receive personalized vaccines manufactured at their local pharmacies thanks to special measures under the Pharmaceutical Affairs Law, which allows us to gather real-world evidence before starting clinical trials and significantly accelerates EMA market approval by having access to at least several hundred patients before clinical trials begin.

[0333] The timing to bring this innovation to market is ideal because:

[0334] Trending: Just as our reflection in the mirror is unique, so too does our DNA reveal our individuality. In an era where individualism is celebrated, why does medicine still adhere to a one-size-fits-all strategy? A shift towards personalized medicine is called for.

[0335] The immunotherapy revolution: The ongoing immunotherapy revolution is highlighting the limitations of existing solutions and paving the way for personalized approaches like ours.

[0336] Advances in Vaccine Development: Major vaccine manufacturers are investing in the development of personalized RNA vaccines, heralding a new era in therapeutic vaccine development.

[0337] Growing interest in personalized cancer treatment: New initiatives such as the German Center for Personalized Medicine reflect growing interest in personalized cancer treatment.

[0338] Common Sense Oncology Initiative: Leading oncologists worldwide are driving change in cancer clinical trials and care, perfectly aligned with VERDI's mission.

[0339] Urgent medical need: The fact that the first cancer patient has been treated with a personalized vaccine designed by VERDI highlights the urgent medical need and stakeholder acceptance.

[0340] COVID pandemic: The pandemic has increased public knowledge and acceptance of vaccines, creating a favorable environment for personalized vaccine therapy for cancer patients.

[0341] Advances in AI: As noted by reviewers, significant advances in AI further support the timely adoption of the company's AI-powered solutions to enable personalized cancer care.

[0342] Societal Needs: The urgency of the development of the proposed project is justified by its alignment with current scientific and technological trends as well as societal needs.

[0343] Our business model is massively scalable because the components needed to manufacture our personalized vaccine are not manufactured in-house. Instead, the personalized vaccine is manufactured in local pharmacies and the components for the vaccine kit are sourced locally from biotech companies, completely eliminating our dependency on big pharma companies.

[0344] Our web application manages the entire patient treatment process, starting from diagnosis, vaccine design, formulation, personalized manufacturing at the nearest or fastest provider, delivery from a central source, distribution of pharmacy kits, dispensing at the pharmacy, and finally patient treatment at the oncologist's office. Each party can track each step of the process, just like tracking a package with DHL, UPS, or FedEx. This is the foundation for the scalability of our business.

[0345] References 1.Sette and Crotty,Adaptive immunity to SARS-CoV-2 and COVID-19, Cell(2021),https: / / doi.org / 10.1016 / j.cell.2021.01.007 2.Swadling,L.,Diniz,M.O.,Schmidt,N.M.et al.Pre-existing polymerase-specific T cells expand in abortive seronegative SARS-CoV-2.Nature(2021).https: / / doi.org / 10.1038 / s41586-021-04186-8 1 3.Moss,P.The T cell immune response against SARS-CoV-2.Nat Immunol 23,186-193(2022).https: / / doi.org / 10.1038 / s41590-021-01122-w 4.Grifoni A,Sidney J,Vita R,Peters B,Crotty S,Weiskopf D,Sette A.SARS-CoV-2 human T cell epitopes:Adaptive immune response against COVID-19.Cell Host Microbe.2021 Jul 14;29(7):1076-1092.doi:10.1016 / j.chom.2021.05.010.Epub 2021 May 21.PMID:34237248;PMCID:PMC8139264. 5.Bukhari,S.N.H.;Jain,A.;Haq,E.;Mehbodniya,A.;Webber,J.Machine Learning Techniques for the Prediction of B-Cell and T-Cell Epitopes as Potential Vaccine Targets with a Specific Focus on SARS-CoV-2 Pathogen:A Review.Pathogens 2022, 11,146.https: / / doi.org / 10.3390 / pathogens11020146 6.Reynisson B,Alvarez B,Paul S,Peters B,Nielsen M.NetMHCpan-4.1 and NetMHCIIpan-4.0:improved predictions of MHC antigen presentation by concurrent motif deconvolution and integration of MS MHC eluted ligand data.Nucleic Acids Res.2020;48:449-454.doi:10.1093 / nar / gkaa379 7.Saini SK,Hersby DS,Tamhane T,Povlsen HR,Amaya Hernandez SP,Nielsen M,Gang AO,Hadrup SR.SARS-CoV-2 genome-wide T cell epitope mapping reveals immunodominance and substantial CD8+ T cell activation in COVID-19 patients.Sci Immunol.2021 Apr 14;6(58):eabf7550.doi:10.1126 / sciimmunol.abf7550.PMID:33853928;PMCID:PMC8139428. 8.Snyder,T.M.,Gittelman,R.M.,Klinger,M.,May,D.H.,Osborne,E.J.,Taniguchi,R.,… Robins,H.S.(2020).Magnitude and Dynamics of the T-Cell Response to SARS-CoV-2 Infection at Both Individual and Population Levels.MedRxiv:The Preprint Server for Health Sciences.https: / / doi.org / 10.1101 / 2020.07.31.20165647 9.Jaskie K,Spanias A.Positive Unlabeled Learning.Synthesis Lectures on Artificial Intelligence and Machine Learning.2022 Apr 19;16(1):2-152. 10.Stone JD,Chervin AS,Kranz DM.T-cell receptor binding affinities and kinetics:impact on T-cell activity and specificity.Immunology.2009;126(2):165-176.doi:10.1111 / j.1365-2567.2008.03015.x 11.Hennecke and Wiley,T Cell Receptor-MHC Interactions up Close,Cell(2001),https: / / doi.org / 10.1016 / S0092-8674(01)00185-4 12.https: / / www.fda.gov / media / 146479 / download 13.Blumenthal at al.2016.Trafficking of MHC molecules to the cell surface creates dynamic protein patches.J.Cell Sci.129:3342-3350. 14.Anikeeva,N.,Fisher,N.,O.,Blanchette,C.,D.& Sykulev,Y.(2019).Extent of MHC Clustering Regulates Selectivity and Effectiveness of T Cell Responses.The Journal of Immunology,202,591-597.Available at:https: / / doi.org / 10.4049 / jimmunol.1801196 15.Assarsson,E.,Sidney,J.,Oseroff,C.,Pasquetto,V.,Bui,H.-H.,Frahm,N.,… Sette,A.(2007).A Quantitative Analysis of the Variables Affecting the Repertoire of T Cell Specificities Recognized after Vaccinia Virus Infection.The Journal of Immunology,178(12),7890-7901.https: / / doi.org / 10.4049 / jimmunol 178.12.7890 16.Jiang,S.et al.(2022)‘Identification of a promiscuous conserved CTL epitope within the SARS-CoV-2 spike protein’,Emerging Microbes ￥& Infections.Taylor & Francis,11(1),pp.730-740.doi:10.1080 / 22221751.2022.2043727 17.Tarke A,Sidney J,Kidd CK,et al.Comprehensive analysis of T cell immunodominance and immunoprevalence of SARS-CoV-2 epitopes in COVID-19 cases.Cell Reports 18. Aghbash PS, Eslami N, Shamekh A, Entezari-Maleki T, Baghi HB Epub 2021 Jan 27.PMID:33508291;PMCID:PMC7838580. 19.Cevik M,Kuppalli K,Kindrachuk J,Peiris M.Virology, transmission, and pathogenesis of SARS-CoV-2

Claims

1. 1. A computer-implemented method for identifying in a subject at least one antigen that is predicted to induce a T cell response that attacks unhealthy cells in said subject, comprising: obtaining subject data identifying a plurality of major histocompatibility complex (MHC) molecules of said subject; obtaining sequence data for a plurality of epitopes within a protein expressed in the unhealthy cell, each epitope being an amino acid sequence within the protein; obtaining, for each epitope within the plurality of epitopes, a potency score indicative of the likelihood that each of the identified plurality of MHC molecules will present the epitope on its surface; generating a ranked list of epitopes based on the determined potency scores; identifying at least one antigen by selecting at least one epitope that is highly ranked in the ranked list; and outputting at least one of the ranked list and sequence data for the at least one identified antigen; A method comprising:

2. selecting a plurality of epitopes that are highly ranked in said ranked list; identifying subtopes of each of the selected epitopes using a directed graph network; and identifying each subtopes common to said selected epitopes; The method of claim 1 further comprising:

3. 3. The method of claim 2, further comprising identifying at least one antigen by selecting a common subtopes that are highly ranked in the ranked list.

4. 4. The method of claim 2 or claim 3, further comprising identifying at least one of said antigens by selecting the longest common subitope.

5. The method of any one of claims 2 to 4, further comprising identifying the shortest common subtopes as the target sequence.

6. The potency score is obtained by, for each epitope within the plurality of epitopes: selecting an epitope from said plurality of epitopes; obtaining, for each of the identified MHC molecules, a probability score indicative of the likelihood that each MHC molecule will carry the selected epitope; and summing at least some of the probability scores to calculate the epitope weight score for the selected epitope; The method of any one of claims 1 to 5, comprising the step of calculating an epitope weight score by:

7. 7. The method of claim 6, wherein the probability score is an eluted ligand score that indicates the likelihood that an epitope will be eluted from a particular MHC molecule.

8. The epitope weight score is: [Equation 1] where x is the peptide, i is an integer from 1 to n, n is the number of identified MHC molecules, and ELS （ｘ，ｉ） is the probability score for each MHC molecule and epitope pair (i.e., the eluted ligand score), ELT is the eluted ligand threshold, w ｉ The method of claim 7 , wherein: is a weighting parameter.

9. 9. The method of any one of claims 1 to 8, when dependent on any one of claims 2 to 5, further comprising the step of calculating for each epitope at least one additional score indicating whether the subtopes of each epitope are capable of eliciting the same T cell response.

10. 1. A method for designing a personalized vaccine to induce a T cell response that attacks unhealthy cells in a subject, comprising: obtaining subject data identifying a plurality of major histocompatibility complex (MHC) molecules for the subject, the subject data identifying at least one of a set of HLA class I molecules and a set of HLA class II molecules; obtaining sequence data for a plurality of epitopes within a protein expressed in the unhealthy cell, each epitope being an amino acid sequence within the protein; obtaining, for each epitope within the plurality of epitopes, a first potency score indicative of the likelihood that each epitope will be presented on the surface of the identified set of HLA class I molecules; generating a first ranked list of epitopes based on the determined first potency score; selecting a plurality of epitopes that are highly ranked in said first ranked list; using a directed graph network to identify the subtopes of each of the selected epitopes; identifying each subtopes common to two or more of said selected epitopes; identifying at least one first antigen by selecting at least one subtopes that is itself a highly ranked epitope in said first ranked list; and outputting the identified at least one first antigen for use in the personalized vaccine; A method comprising:

11. obtaining, for each epitope within the plurality of epitopes, a second efficacy score indicative of the likelihood that each epitope will be presented on the surface of the identified set of HLA class II molecules; generating a second ranked list of epitopes based on the determined second potency score; selecting a plurality of epitopes that are highly ranked in said second ranked list; using a directed graph network to identify the subtopes of each of the selected epitopes; identifying each subtopes common to two or more of said selected epitopes; identifying at least one second antigen by selecting at least one subtopes that is itself a highly ranked epitope in said second ranked list; and outputting the identified at least one second antigen for the personalized vaccine; The method of claim 10 further comprising:

12. determining whether the identified first antigen is a highly ranked epitope in the second ranked list; and designing the personalized vaccine based on the first antigen if the first antigen is a highly ranked epitope in the second ranked list; The method of claim 11 further comprising:

13. 13. The method of any one of claims 1 to 12, wherein the antigens are highly ranked in the ranked list, are capable of being presented by MHC class I and / or MHC class II molecules on the surface of cells of the subject, induce a CD4 and / or CD8 T cell response, and comprise multiple epitopes sharing at least one common sequence capable of eliciting the same T cell response, and optionally multiple overlapping epitopes.

14. 14. A personalized vaccine composition comprising a peptide antigen derived from a protein expressed in an unhealthy cell of a recipient, the peptide antigen comprising a plurality of top-ranked epitopes selected according to the method of any one of claims 1 to 13, and optionally comprising a cell membrane-penetrating peptide and / or an excipient.

15. 15. The composition of claim 14, further comprising at least two peptide antigens comprising a plurality of top-ranked epitopes selected by the method of any one of claims 1 to 13, derived from proteins expressed in the unhealthy cells of the recipient.

16. The personalized vaccine composition of claim 14 or claim 15, wherein the cell membrane-penetrating peptide is positioned between two of the antigens.

17. 17. A method for determining the efficacy of an immune response as a means of safety and efficacy control in a recipient of a personalized vaccine composition according to any one of claims 14 to 16, comprising at least one antigen identified using the method of claims 1 to 13, comprising: generating all potential epitopes from said sequence of said selected antigen; creating a ranked list of these epitopes based on said potency scores; verifying that epitopes derived from proteins expressed in unhealthy cells of the recipient evoke a strong immune response as indicated by their high potency scores, thereby ensuring efficacy; ensuring that the epitopes derived from the cell membrane-penetrating peptides and / or excipients are immunologically inert, as indicated by their low potency scores, thereby ensuring safety; confirming that epitopes with high potency scores are not components of proteins expressed in healthy cells, thereby further contributing to the safety of the personalized vaccine; A method comprising:

18. 17. A method of treating or preventing a disease in a subject, comprising administering at least one personalized vaccine composition according to any one of claims 14 to 16, wherein at least one personalized vaccine composition is administered alone or in combination with an additional therapeutic agent, said administration occurring simultaneously or sequentially, and optionally wherein at least two personalized vaccine compositions are administered.

19. 19. The method of claim 18, wherein the disease is cancer or an autoimmune disease, or a viral infection.

20. A method for inducing an antigen-specific immune response in a subject, comprising administering to the subject a personalized vaccine composition according to any one of claims 14 to 16.

21. A kit comprising a number of items necessary for preparing an individualized vaccine composition according to any one of claims 14 to 16 for personal use, Two synthetic peptides, each synthetic peptide comprising at least one selected antigen according to any one of claims 1 to 13, and further comprising at least a portion of a cell membrane-penetrating peptide; and a means for covalently linking the two synthetic peptides during the preparation of the personalized vaccine, thereby reconstituting the function of the cell membrane-penetrating peptide; Includes a kit.

22. A method for preparing a personalized peptide vaccine or therapeutic composition, comprising: preparing a first amino acid sequence; preparing a second amino acid sequence, both of which comprise an antigen comprising a set of top-ranked epitopes derived from a protein expressed in the unhealthy cells of the individual; and at least a portion of a cell membrane-penetrating peptide; and covalently linking the first amino acid sequence and the second amino acid sequence to form a personalized vaccine and reconstitute the function of the cell membrane-penetrating peptide located between the first antigen and the second antigen.

23. 1. A method of treating a patient with cancer and / or infectious disease with a personalized vaccine that induces an immune response against unhealthy cells, comprising: a) obtaining at least one sample from a subject containing said unhealthy cells; b) analyzing at least one of said samples to identify expression of MHC class I and / or MHC class II molecules in said subject and at least one protein expressed in said unhealthy cells; c) generating a ranked list of epitopes from at least one of said proteins based on their ability to be presented by an MHC molecule of interest and to induce a CD4 and / or CD8 T cell response using the method of any one of claims 1 to 13; d) selecting from said ranked list a plurality of top-ranked epitopes that share at least one common sequence capable of eliciting a T cell response; e) incorporating said selected epitopes into at least one first generation personalized vaccine design; f) preparing said personalized vaccine, including, if necessary, preparing said vaccine at a pharmacy under a compounding exemption; g) administering said first generation personalized vaccine to said subject; h) monitoring the subject's response to the vaccine; and i) if the subject has not been rid of the unhealthy cells, preparing at least one second-generation personalized vaccine that is different from the first-generation personalized vaccine according to steps a-f; A method comprising:

Images