Ranking neoantigens for personalized cancer vaccines

Abstract

Disclosed herein is a method for ranking tumor-specific neoantigens from a subject's tumor that are suitable for a subject-specific immunogenic composition. Suitable tumor-specific neoantigens are those that are likely to be presented on the cell surface of the tumor, likely to be immunogenic, predicted to be expressed in sufficient amounts to elicit an immune response in the subject, optionally exhibit sufficient diversity throughout the tumor, and have relatively high manufacturability. The method takes a group of neoantigens (peptide vaccine candidates) and ranks the neoantigens such that the top ranked group of neoantigens simultaneously promotes cell surface presentation of key neoantigens to class I and class II MHC molecules. The top ranked neoantigens can then be further refined according to manufacturability and / or other criteria.

Description

[Technical field]

[0001] This application claims priority to U.S. Provisional Patent Application No. 63 / 146,392, filed February 5, 2021, the entire contents of which are incorporated herein by reference.

[0002] Sequence Listing Reference This application contains a sequence listing in computer readable form, which is incorporated herein by reference. The ASCII copy, created on February 3, 2022, has the filename 146401_091686_SL.txt and is 14,005 bytes in size. [Background technology]

[0003] Cancer is the leading cause of death worldwide, with approximately one in four deaths due to cancer. Siegel et al., CA: A Cancer Journal for Clinicians, 68:7-30 (2018). In 2018, 18.1 million new cases of cancer and 9.6 million deaths due to cancer-related causes were reported. Bray et al., CA: A Cancer Journal for Clinicians, 68(6):394-424. There are currently many standard cancer treatments, including ablative techniques (e.g., surgery and radiation) and chemical techniques (e.g., chemotherapy drugs). Unfortunately, such treatments are often associated with significant risks, toxic side effects, prohibitive costs, and uncertain efficacy.

[0004] Cancer immunotherapy (e.g., cancer vaccines) has emerged as a promising cancer treatment. The goal of cancer immunotherapy is to activate the immune system to selectively destroy cancer while leaving normal tissues intact. Traditional cancer vaccines typically target tumor-associated antigens, which are typically present in normal tissues but overexpressed in cancer. However, because these antigens are often present in normal tissues, immune tolerance may prevent immune activation. Some clinical trials targeting tumor-associated antigens have failed to demonstrate sustained beneficial effects compared to standard of care. Li et al., Ann Oncol., 28 (Suppl 12): xii11-xii17 (2017).

[0005] Neoantigens can be very attractive targets for cancer immunotherapy. Neoantigens are non-self proteins with individual specificity. Neoantigens originate from random somatic mutations in the genome of tumor cells and are not expressed on the surface of normal cells. Ibid. Because neoantigens are expressed only on tumor cells and do not induce central immune tolerance, cancer vaccines targeting cancer neoantigens may have advantages such as reduced central immune tolerance and improved safety profile. Ibid.

[0006] The cancer mutational landscape is complex, and tumor mutations are unique to each subject. Most somatic mutations detected by sequencing do not result in effective neoantigens. Only a small fraction of tumor DNA or tumor cell mutations are transcribed, translated, and processed into tumor-specific neoantigens with sufficient accuracy to design a potentially effective vaccine. Furthermore, not all neoantigens are immunogenic. In fact, only about 1%–2% of T cells spontaneously recognize endogenous neoantigens. See Karpanen et al., Front Immunol., 8:1718 (2017). Furthermore, the cost and time required to produce a neoantigen vaccine is enormous.

[0007] Thus, efficiently and accurately predicting, prioritizing, and selecting neoantigen candidates for immunogenic compositions remains a challenge. Thus, there is a significant unmet need for integrated methods to characterize tumor genomic material to identify neoantigens, identify which neoantigens are targeted by the immune system, and select which neoantigens are likely suitable for an effective immunogenic composition.

[0008] Summary of the Invention The present disclosure relates to a novel method for ranking one or more suitable tumor-specific neoantigens from a subject's tumor for a personalized (i.e., subject-specific) immunogenic composition. The present disclosure also relates to a method for treating cancer in a subject in need thereof by administering an immunogenic composition comprising selected tumor-specific neoantigens using a novel approach for ranking tumor-specific neoantigens and formulating an immunogenic composition comprising selected tumor-specific neoantigens based on the ranking technique. Suitable tumor-specific neoantigens are neoantigens that are likely to be present on the cell surface of the tumor, likely to be immunogenic, predicted to be expressed in sufficient amounts to induce an immune response in the subject, and optionally exhibit sufficient diversity throughout the tumor, and have relatively high manufacturability. The method takes a group of neoantigens (peptide vaccine candidates) and ranks the neoantigens such that the top ranked group of neoantigens simultaneously promotes cell surface presentation of key neoantigens to class I and class II MHC molecules. The top ranked group of neoantigens can then be further refined according to manufacturability and / or other criteria.

[0009] The approach begins with obtaining sequence data from the tumor. The sequence data is used to obtain data representing the polypeptide sequences of one or more tumor-specific neoantigens. The sequence data may be nucleotide sequence data, polypeptide sequence data, exome sequence data, transcriptome sequence data, or whole genome nucleotide sequence data. Suitable tumor-specific neoantigens are neoantigens that are likely to be present on the cell surface of the tumor, likely to be immunogenic, predicted to be expressed in sufficient amounts to induce an immune response in a subject, and optionally exhibit sufficient diversity throughout the tumor, and have relatively high manufacturability. The method obtains a group of neoantigens (peptide vaccine candidates) and ranks the neoantigens such that the top ranked group of neoantigens simultaneously promotes cell surface presentation of key neoantigens to class I and class II MHC molecules. The top ranked neoantigens can then be further refined according to manufacturability and / or other criteria. As described in more detail below, the ranking is primarily based on a calculation of the immunogenicity of the neoantigens. For short neoantigens, the immunogenicity of the short neoantigen is determined based at least in part on the probability that at least one allele in the subject's plurality of HLA class I alleles presents the short neoantigen and does not present a germline sibling of the short neoantigen. Similarly, for long neoantigens, the immunogenicity of the long neoantigen is determined based at least in part on the probability that at least one allele in the subject's plurality of HLA class II alleles presents the long neoantigen and does not present a germline sibling of the long neoantigen. These probabilities are determined using output provided by one or more machine learning platforms / models, which are trained to determine the probability that a given allele presents a particular antigen.

[0010] The immunogenic composition formulated at least in part based on the technology of the present invention may comprise at least about 10 tumor-specific neoantigens, or at least about 20 tumor-specific neoantigens. The tumor-specific neoantigens may be encoded by short or long polypeptides. The immunogenic composition may comprise a nucleotide sequence, a polypeptide sequence, RNA, DNA, a cell, a plasmid, a vector, a dendritic cell, or a long synthetic peptide. The immunogenic composition may further comprise an adjuvant.

[0011] The present disclosure also relates to a method of treating cancer in a subject in need thereof, comprising administering an individualized immunogenic composition comprising one or more tumor-specific neoantigens selected using the methods described herein. The methods disclosed herein can be suitable for treating a variety of cancers. The tumors include melanoma, breast cancer, ovarian cancer, prostate cancer, kidney cancer, gastric cancer, colon cancer, testicular cancer, head and neck cancer, pancreatic cancer, brain cancer, B-cell lymphoma, acute myeloid leukemia, chronic myeloid tumor, leukemia, chronic lymphocytic leukemia, T-cell lymphocytic leukemia, bladder cancer, or lung cancer. Preferably, the cancer is melanoma, breast cancer, lung cancer, or bladder cancer.

[0012] Various embodiments according to the present disclosure will now be described with reference to the drawings. [Brief description of the drawings]

[0013] [Figure 1] 1 illustrates an exemplary provider network (or "service provider system") environment, according to some embodiments. [Diagram 2] 1 is a block diagram of an example provider network that offers storage and hardware virtualization services to customers, according to some embodiments. [Diagram 3] 1 illustrates a system that implements some or all of the techniques described herein, according to some embodiments. [Figure 4]According to an exemplary embodiment, a method for ranking tumor-specific neoantigens from a subject's tumor with respect to a subject-specific immunogenic composition is described. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0014] The present disclosure relates to a novel approach to ranking tumor-specific neoantigens for inclusion in a potent personalized cancer immunogenic composition (e.g., a subject-specific immunogenic composition). The present disclosure also relates to a method of treating cancer in a subject in need thereof by administering an immunogenic composition comprising tumor-specific neoantigens formed using the novel approach to ranking tumor-specific neoantigens and formulating an immunogenic composition comprising the selected tumor-specific neoantigens.

[0015] All patents and publications cited herein are hereby incorporated by reference in their entirety. In the event that an incorporated material conflicts with this specification, this specification takes precedence over such material. Additionally, the citation of a reference herein does not imply that the reference is prior art to the present disclosure. When a range of values ​​is expressed, embodiments using a specific value within that range are included. Furthermore, reference to values ​​stated within a range is intended to include every value within that range. All ranges are inclusive of endpoints and are combinable. When values ​​are expressed as approximations, it is understood that the particular value forms another embodiment by use of the antecedent "about." Reference to a particular numerical value includes at least that particular value unless the context clearly dictates otherwise. The use of "or" means "and / or" unless the specific context dictates otherwise.

[0016] In the present specification and claims, various terms are used in connection with aspects of the present invention. Unless otherwise indicated, such terms shall be given their ordinary meaning in the art. Other terms specifically defined shall be interpreted in a manner consistent with the definitions set forth herein. The techniques and procedures described or referred to herein are generally well understood by those of skill in the art and are employed according to well-known conventional methodologies, e.g., Sambrook et al., Molecular Cloning: A Laboratory Manual 4th ed. (2012) Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, and the like. Where appropriate, procedures involving the use of commercially available kits and reagents are generally carried out according to manufacturer-defined protocols and conditions, unless otherwise indicated.

[0017] As used herein, the singular forms "a," "an," and "the" include the plural forms unless the context clearly dictates otherwise. Terms such as "include," "such as," and the like are intended to mean inclusion without limitation, unless otherwise specified.

[0018] Unless otherwise indicated, the terms "at least," "less than," "about," or similar terms used as antecedents to a series or range should be understood to refer to every element in that series or range. Those skilled in the art will recognize, or be able to ascertain using no more than routine experimentation, many equivalents to the specific embodiments of the invention described herein. Such equivalents are intended to be encompassed by the following claims.

[0019] "Cancer" refers to a physiological condition in a subject in which a population of cells is characterized by uncontrolled proliferation, immortality, metastatic potential, rapid growth, proliferation rate, and / or specific morphological characteristics. Cancer often takes the form of a tumor or mass, but may exist alone in a subject's body or circulate in the bloodstream as separate cells, such as leukemia or lymphoma cells. The term "cancer" encompasses all types of cancer and metastases, including hematological malignancies, solid tumors, sarcomas, carcinomas, other solid tumors, and non-solid tumors. Examples of cancer include, but are not limited to, carcinomas, lymphomas, blastomas, sarcomas, and leukemias. More specific examples of such cancers include squamous cell carcinoma, small cell lung cancer, non-small cell lung cancer, lung adenocarcinoma, lung squamous cell carcinoma, peritoneal carcinoma, hepatocellular carcinoma, gastrointestinal cancer, pancreatic cancer, glioblastoma, cervical cancer, ovarian cancer, liver cancer, bladder cancer, liver cancer, breast cancer (e.g., triple-negative breast cancer, hormone receptor positive breast cancer), osteosarcoma, melanoma, colon cancer, colorectal cancer, endometrial (e.g., serous) or uterine cancer, salivary gland cancer, kidney cancer, liver cancer, prostate cancer, vulvar cancer, thyroid cancer, liver cancer, and various types of head and neck cancer. In some embodiments, triple-negative breast cancer (TNBC) refers to breast cancer that does not express the estrogen receptor (ER), progesterone receptor (PR), or Her2 / neu genes. Hormone receptor positive breast cancer refers to breast cancer that is positive for at least one of ER or PR and negative for Her2 / neu (HER2).

[0020] As used herein, a "neoantigen" refers to an antigen that has at least one change that distinguishes it from the corresponding parent antigen, for example, through a mutation in a tumor cell or a tumor cell-specific post-translational modification. Mutations or alterations include frameshifts, indels, missense or nonsense substitutions, splice site changes, genomic rearrangements or gene fusions, or alterations in genomic expression that give rise to neoantigens. Mutations may include splice mutations. Tumor cell-specific post-translational modifications may include aberrant phosphorylation. Tumor cell-specific post-translational modifications may also include splice antigens generated by the proteasome. See Lipe et al., Science, 354(6310):354:358 (2016). Point mutations typically account for about 95% of tumor mutations, with indels and frameshift mutations accounting for the remainder. See Snyder et al., N Engl J Med., 371:2189-2199 (2014).

[0021] As used herein, a "tumor-specific neoantigen" refers to a neoantigen that is present in a subject's tumor cells or tissues, but not present in the subject's normal cells or tissues.

[0022] "Germline sibling" herein refers to a germline antigen that represents the non-mutated peptide equivalent of the corresponding neoantigen.

[0023] As used herein, "next-generation sequencing" or "NGS" refers to a sequencing technology that offers increased throughput compared to traditional approaches (e.g., Sanger sequencing) and the ability to generate hundreds of thousands of sequence reads at a time.

[0024] "Neural network" in this specification refers to a machine learning model for classification or regression, typically composed of multiple layers of linear transformations followed by element-wise nonlinearities trained by stochastic gradient descent and backpropagation.

[0025] As used herein, a "subject" refers to any animal, such as a mammal, including but not limited to a human, a non-human primate, a rodent, etc. In some embodiments, the mammal is a mouse. In some embodiments, the mammal is a human.

[0026] As used herein, a "cancer cell" refers to any cell that is or originates from a cancer cell. A "tumor cell" also refers to a cell that exhibits cancer-like properties, such as uncontrolled replication, resistance to anti-proliferative signals, the ability to metastasize, or the loss of the ability to undergo programmed cell death.

[0027] Further description of the method and guidance for carrying out the method are provided herein.

[0028] I. Methods for Ranking Tumor-Specific Neoantigens Disclosed herein is a method for ranking tumor-specific neoantigens from a subject's tumor that are suitable for a subject-specific immunogenic composition. Suitable tumor-specific neoantigens are those that are likely to be presented on the cell surface of the tumor, likely to be immunogenic, predicted to be expressed in sufficient amounts to elicit an immune response in the subject, optionally exhibit sufficient diversity throughout the tumor, and have relatively high manufacturability. The method takes a group of neoantigens (peptide vaccine candidates) and ranks the neoantigens such that the top ranked group of neoantigens simultaneously promotes cell surface presentation of key neoantigens to class I and class II MHC molecules. The top ranked group of neoantigens can then be further refined according to manufacturability and / or other criteria.

[0029] Tumor and subject sequence data are used to rank tumor-specific neoantigens derived from a subject's tumor. Tumor sequence data is used to obtain data representing the polypeptide sequences of one or more tumor-specific neoantigens. Typically, sequence data representing the polypeptide sequences of one or more tumor-specific neoantigens can be obtained by subjecting a tumor sample to sequence analysis. In some embodiments, obtaining sequence data includes receiving or accessing stored data from a previously performed sequencing run. The sequence data can be exome sequence data, transcriptome sequence data, whole genome nucleotide sequence data, nucleotide sequence data, or polypeptide sequence data. Various methods of obtaining tumor and subject sequence data can be used in the methods described herein. Some exemplary sequencing methods are described in more detail below.

[0030] Once sequence data representing the polypeptide sequences of one or more tumor-specific neoantigens is obtained, the sequence data can be analyzed in conjunction with the subject's MHC molecules to identify and rank candidate neoantigens for inclusion in the subject's immunogenic composition.

[0031] In one embodiment, given a set of HLA I and HLAI I alleles of a subject and a list of somatic mutations of a tumor, a top-ranked set of about 30 long peptide candidates and about 15 short peptide candidates are identified and subjected to manufacturability analysis. A starting set of peptides is identified using sliding windows for each somatic mutation. These are scored using MHC class I and class II machine learning models described below. The 15 short peptides and the 30 long peptides contain at least one MHC class I epitope, and the long peptides may also contain one or more MHC class II epitopes. Then, 9 of the 30 long peptide candidates and 10 of the 15 short peptide candidates are selected for addition to the immunogenic composition based on manufacturability. In other embodiments, various numbers of the top-ranked long and / or short peptide candidates are provided for manufacturability analysis. In some embodiments, 20 to 100 (e.g., 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58 , 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98, 99, 100) candidates are provided. In other embodiments, more or fewer top ranked candidates may be provided for manufacturability analysis.

[0032] Similarly, various numbers of long and / or short peptide candidates may be ultimately selected for addition to the immunogenic composition. Long neoantigens typically have more manufacturing constraints than short neoantigens, and therefore more long neoantigens are required. In some embodiments, long neoantigens are neoantigens having about 15-30 amino acids (e.g., 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 amino acids) and short neoantigens are neoantigens having about 8-11 amino acids (e.g., 8, 9, 10, or 11 amino acids). In other embodiments / implementations of the present technology, long and short neoantigens having different numbers of amino acids may be defined.

[0033] FIG. 4 illustrates an exemplary method (400) for ranking tumor-specific neoantigens from a tumor of a subject with respect to a subject-specific immunogenic composition. First, a plurality of somatic mutations present in the tumor is identified (410). Then, for each individual somatic mutation, a first plurality of short neoantigens and a first plurality of long neoantigens associated with the somatic mutation are identified or otherwise obtained (420). The first plurality of short neoantigens may include short polypeptides comprising at least one MHC class I epitope associated with the subject. The first plurality of long neoantigens may include long polypeptides comprising at least one MHC class I epitope and at least one MHC class II epitope associated with the subject.

[0034] From the first plurality of short neoantigens, the short neoantigen with the highest immunogenicity score is selected or determined (430) and added to the short neoantigen candidate list (440). The selected short neoantigen is also referred to as the best short neoantigen for the particular somatic mutation. Similarly, from the first plurality of long neoantigens, the long neoantigen with the highest immunogenicity score is selected or determined (460) and added to the long neoantigen candidate list (470). The selected long neoantigen is also referred to as the best long neoantigen for the particular somatic mutation. The immunogenicity score may be any form of assessment or value, such as a numerical or non-numerical value, used to represent the quality of the neoantigen based on one or more data with respect to one or more criteria. The immunogenicity score of a neoantigen may be determined according to the techniques described in more detail below. The step of selecting the best short neoantigen and the best long neoantigen is performed for each somatic mutation in the plurality of somatic mutations, and the completed short neoantigen candidate list includes the best short neoantigen for each somatic mutation, and the completed long neoantigen candidate list includes the best long neoantigen for each somatic mutation. In other words, each short neoantigen in the short neoantigen candidate list may be the best short neoantigen for each somatic mutation in the plurality of identified somatic mutations. Similarly, each long neoantigen in the long neoantigen candidate list may be the best long neoantigen for each somatic mutation in the plurality of identified somatic mutations. For each somatic mutation, the best short neoantigen and the best long neoantigen are identified. The short neoantigen candidate list and the long neoantigen candidate list are then sorted and ranked in descending order of immunogenicity score (450, 480), respectively.

[0035] In some embodiments, the sorted list of long neoantigen candidates is trimmed to a predetermined number of top ranked long neoantigen candidates. For example, the list may be trimmed to the top 30 long neoantigen candidates. In other words, a predetermined number of top ranked long neoantigens from the sorted list are selected for manufacturability analysis and determination. In some embodiments, the trimmed list of long neoantigen candidates (i.e., a predetermined number of the top ranked long neoantigens) is provided to a manufacturer to determine manufacturability. The manufacturability of a particular neoantigen may be expressed as a numerical or non-numerical score, value, classification, etc. The manufacturability may be based on one or more criteria or data, which may be calculated, weighted, or otherwise processed in various ways. The manufacturability determination may be based on an analysis performed on the actual neoantigen or may be based on a standard. The manufacturer then selects a subgroup of long neoantigen candidates from the trimmed list of long neoantigen candidates based on manufacturability (i.e., the top group in manufacturability scores). For example, a subgroup may include the top 9 long neoantigens with the highest manufacturability scores. As used herein, "manufacturer" refers to the entity that performs the manufacturability analysis and selects the subgroup, but may be a similar entity or a third party responsible for the remaining technology.

[0036] Once the subgroup of long neoantigen candidates based on manufacturability is obtained, neoantigens in the short neoantigen candidate list that are included in any of the neoantigens in the long neoantigen subgroup are removed from the short neoantigen candidate list to avoid duplication. In other words, for each long neoantigen in the long neoantigen candidate subgroup, its mutation is identified and the corresponding short neoantigen is removed from the short neoantigen candidate list. The short neoantigens remaining in the short neoantigen candidate list are trimmed to a predetermined number based on the immunogenicity score. For example, the list may be trimmed to about 15 neoantigens. These short candidates are then subjected to a manufacturability determination to obtain a subgroup of short neoantigen candidates selected for their manufacturability. The subgroup of short neoantigen candidates and the subgroup of long neoantigen candidates are used to form or generate a subject-specific immunogenic composition that can be administered to a subject.

[0037] Obtaining the first multiple short neoantigens: To obtain the first plurality of short neoantigens defined for an individual somatic mutation, first, the longest neoantigen sequence containing the mutated amino acid, neoT, is identified. The germline sibling of this neoantigen, neoG, is also identified. Then, all neoantigen sequences containing the mutation with a minimum (e.g., 8) to a maximum (e.g., 11) number of amino acids are identified using a sliding window spanning the longest neoantigen sequence. This results in the first plurality of short neoantigens, neoT_1. For example, in an embodiment where the minimum number of amino acids is 8 and the maximum number is 11, all neoantigen sequences containing the mutation and having a length of 8, 9, 10, and 11 amino acids within the longest neoantigen, neoT, are identified and designated as members of the first plurality of short neoantigens, neoT_1. In some embodiments, the individual alleles in the subject's multiple HLA class I alleles a1 are identified.

number

number

number

number

[0038] In some embodiments, the probability may be determined based at least in part on data from an MHC class I machine learning model trained to determine the probability that a given allele in a plurality of HLA class I alleles presents a particular antigen.

[0039] The first plurality of short neoantigens may be further filtered to not include neoantigens that are nested within a neoantigen from the first plurality of short neoantigens or that are nested within another neoantigen, which may be done by identifying pairs of neoantigens where one sequence of the pair is nested within the other sequence, which may have a higher probability score, as calculated using formula (1) above.

number

[0040] Obtaining the first multiple long neoantigens: Once the first plurality of short neoantigens is defined, a short subsequence T1 is identified from the longest neoantigen sequence neoT. The short subsequence T1 is identified as the shortest subsequence of the longest neoantigen sequence neoT that includes all neoantigens in the first plurality of short neoantigens neoT_1. As described above, the filtered first plurality of short neoantigens neoT_1_filt does not include neoantigens included in the first plurality of short neoantigens or includes other neoantigens. An extended sequence T1_long can be identified. The extended sequence T1_long is obtained by adding amino acids on both sides of the short subsequence T1 according to the longest neoantigen neoT, so that there is a first maximum number of amino acids adjacent to each side of the mutated amino acid. For example, the first maximum number is 29. In some embodiments, the second maximum number of amino acids may be between 9 and 50 (e.g., 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50). Thus, in such embodiments, the extended sequence T1_long comprises the short subsequence T1 and 29 amino acids flanking either side of the mutated amino acid.

[0041] All possible subsequences from the long subsequence having a length between the length of the short subsequence [length(T1)] and the length of a second maximum number of amino acids may be identified and designated as the initial plurality of long neoantigens neoT_2. For example, the second maximum number is 30. In some embodiments, the second maximum number of amino acids may be 9 to 50 (e.g., 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50). In some embodiments, the initial plurality of long neoantigens may be filtered based on one or more manufacturability criteria.

[0042] Obtaining immunogenicity scores for neoantigens: The immunogenicity score of the individual short neoantigens used to select and rank (sort) the short neoantigens can be determined based at least in part on the probability that at least one allele in the subject's plurality of HLA class I alleles will present the individual short neoantigen and not present a germline sibling of the individual short neoantigen, which can be expressed as:

number

[0043] In some embodiments, the above calculated value is referred to as the pan-HLAI allele score for each short neoantigen in the initial plurality of short neoantigens before filtering for nested pairs.

number

[0044] Further, in some embodiments, the allele score of the peptide sequence comprising the variant amino acid and the MHC class I epitope for an individual MHC class I allele is determined based on the probability that the individual allele represents at least one neoantigen and does not represent a germline sibling of at least one neoantigen in the initial plurality of short neoantigens after filtering for nested pairs, which can be expressed as follows:

number

[0045]

number

[0046] In some embodiments, the pan-allelic HLA class I score is determined based at least in part on the probability that at least one allele in the plurality of HLA class I alleles of the subject presents at least one neoantigen in the short neoantigen group and does not present a germline sibling of the at least one neoantigen, which can be expressed as:

number

[0047] The immunogenicity score of each individual long neoantigen used to select and rank (sort) the long neoantigens can be determined based at least in part on the probability that at least one allele in the plurality of HLA class II alleles of the subject presents the individual neoantigen and does not present a germline sibling of the individual neoantigen.

number

number

number

number

number

[0048] Therefore, the probability that at least one allele in a subject's multiple HLAII alleles will present an individual neoantigen and not present a germline sibling of the individual neoantigen can be expressed as follows:

number

[0049] The probability is determined based at least in part on data from an MHC class II machine learning model trained to determine the probability that a given allele in a plurality of HLA class II alleles will present a particular antigen.

[0050] The probability that a variant peptide sequence will generate an immune response against one or more HLA class I alleles can be expressed as:

number

number

number

[0051] Short peptides can be directly incorporated into the vaccine and are expected to compete with endogenously expressed peptides for binding to MHC-I. Thus, for these peptides, the score S may be adjusted to also include the predicted binding probability of the peptide to a given MHC-I molecule. The revised score for short peptides can be expressed as:

number

[0052] Allele-specific CD8+ immunogenicity can be expressed as follows:

number

[0053] DistToSelf can be expressed as follows:

number

[0054] The following function can be used to constrain the probability of immunogenicity of a variant peptide close to the germline peptide:

number

number

[0055] In some embodiments, a method of ranking tumor-specific neoantigens from a tumor of a subject for a subject-specific immunogenic composition includes identifying a plurality of somatic mutations present in the tumor, determining a best short neoantigen from the first plurality of short neoantigens for each somatic mutation in the plurality of somatic mutations based at least in part on a quality score of the best short neoantigen, and determining a best long neoantigen from the first plurality of long neoantigens based at least in part on a quality score of the best long neoantigen. Add the best short neoantigen for each somatic mutation to a short neoantigen candidate list, and add the best long neoantigen for each somatic mutation to a long neoantigen candidate list. The lists are then ranked in descending order of quality score, respectively. In some embodiments, the quality score is based at least in part on at least one of a predicted probability of presentation, a predicted binding affinity, and a predicted immunogenic response. In some embodiments, the quality score is based at least in part on a predicted probability of presentation. In some embodiments, the quality score is based at least in part on a predicted binding affinity. In some embodiments, the predicted binding affinity is determined at least in part based on data from an MHC class II learning model trained to determine the binding affinity between class II alleles and a given peptide. In some embodiments, the quality score is based at least in part on a predicted immunogenic response. In some embodiments, the quality score is based at least in part on a combination of predicted presentation probability, predicted binding affinity, and predicted presentation probability. In some embodiments, the predicted presentation probability, predicted binding affinity, and predicted presentation probability are determined by one or more machine learning models.

[0056] In some embodiments, peptides may be filtered for consideration or inclusion in the final subject-specific immunogenic composition based on any of the following criteria: 1) RNA abundance (measured in transcripts per million, TPM) of the gene to which the somatic mutation belongs. For example, RNA abundance can be determined by multiplying the RNA TPM value of the gene to which the mutation belongs by the ratio of the number of reads overlapping the mutant locus containing the mutant allele to the sum of (a) the number of reads overlapping the mutant locus containing the mutant allele and (b) the total number of reads overlapping the mutant locus. 2) Whether the somatic mutation is in an essential or driver gene. A driver gene is a gene whose mutation can cause tumor growth. An essential gene is a gene that is essential for the survival of the organism. 3) Is the peptide predicted to be able to pass quality control thresholds for synthesis and solubility? 4) How foreign (i.e., different) is the mutant peptide to the corresponding germline peptide? In some embodiments, a minimum number of mutated amino acids may be required for a peptide to be considered or included, and more exotic peptides may be prioritized over less exotic peptides. 5) The confidence factor that a particular mutation is present in a particular subject. For example, a rare somatic mutation is given a lower confidence score than a more frequently occurring mutation. 6) Whether a candidate peptide contains a particular amino acid, such as cysteine.

[0057] Somatic variants can mutate zero, one, or more amino acids. For example, silent mutations mutate zero amino acids, while single nucleotide variants typically mutate one amino acid, and frameshift or stop-loss mutations can mutate multiple amino acids. If an RNA top read is found to be assembled upstream of the mutated locus, the longest consensus mRNA sequence that overlaps with the mutated amino acid is assembled. The assembly of the mRNA sequence is stopped when the RNA read coverage ends or a new stop codon is found. If no RNA top read is found, or no mutated amino acid is found beyond the length of the requested protein sequence, the assembly of the mRNA sequence is stopped.

[0058] The predicted presentation data consists entirely of "positive" samples that may be presented on the cell surface. Thus, one or more probabilistic negative mining strategies can be used during training to train predictors that may require "negative" samples that cannot be presented on the cell surface. Such a process may include HLA allele shuffling, where given a positive sample (e.g., a peptide and corresponding HLA allele), different alleles that do not belong to the epitype of the positive allele can be randomly sampled to replace the given allele. Each HLA allele may be classified into one or more HLA epitypes until only unclassified HLA alleles remain. Unclassified HLA alleles can be mapped to one or more "unclassified" epitype classes, and these groups can be treated similarly to classified epitype classes.

[0059] Additionally, predictors can be trained using peptide shuffling, where, given a positive sample consisting of a peptide and a corresponding HLA allele, the peptide is replaced with an amino acid subsequence of the same length randomly sampled from the protein source of the peptide.

[0060] Random peptides may be generated to aid in training the predictor. According to this embodiment, random peptides sampled from the amino acid data distribution can be generated, with a qualitative affinity target below a predefined threshold and a negatively represented target. The length of the random peptide is determined such that there is an equal number of non-binding data points per peptide length for each allele. A predefined ratio (e.g., 10:1) of negatively represented samples to positively represented samples can be sampled, and a sample weight can be applied to the negative samples for balance loss. For each negative sample, the sampling method can be selected randomly with a uniform distribution.

[0061] Sequencing methods Various sequencing methods are known in the art, including but not limited to PCR-based methods, including real-time PCR, whole exome sequencing, deep sequencing, high-throughput sequencing, or combinations thereof. In some embodiments, the aforementioned techniques and procedures are carried out according to the methods described in, for example, Sambrook et al., Molecular Cloning: A Laboratory Manual 4th ed. (2012) Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Other references include Austell et al., Current Protocols in Molecular Biology, ed., Greene Publishing and Wiley-Interscience New York (1992) (regularly updated).

[0062] Sequencing methods include high-throughput sequencing, single-cell RNA sequencing, RNA sequencing, pyrosequencing, sequencing by synthesis, single molecule sequencing, nanopore sequencing, semiconductor sequencing, sequencing by synthesis, sequencing by ligation, sequencing by hybridization, RNA-Sew (Illumina), Digital Gene Expression (Helicos), next-generation sequencing, single molecule sequencing by synthesis (SMSS) (Helicos), massively parallel sequencing, clonal single molecule arrays (Solexa), shotgun sequencing, Maxam-Hillberry or Sanger sequencing, whole genome sequencing, whole exome sequencing, primer walking, sequencing using PacBio, SOLid, IonTorrent, or Nanopore platforms, or any other sequencing method known in the art. The sequencing method used herein to obtain sequence data is preferably high-throughput sequencing. High-throughput sequencing techniques can sequence multiple nucleic acid molecules in parallel, and can sequence millions of nucleic acid molecules at a time. See Churko et al., Circ. Res. 112(12):1613-1623 (2013).

[0063] In some cases, high-throughput sequencing can be replaced by next-generation sequencing. There are many different next-generation platforms that use different sequencing technologies (e.g., using HiSeq or MiSeq instruments available from Illumina, San Diego, CA). Any of these platforms can be used to sequence the genetic material disclosed herein. Next-generation sequencing is based on sequencing a large number of independent reads, each representing 10-1000 bases of nucleic acid. Sequencing by synthesis is a common technique used in next-generation sequencing. In general, sequencing involves hybridizing a primer to a template to form a template / primer duplex, and contacting the duplex with a polymerase in the presence of detectably labeled nucleotides under conditions that allow the polymerase to add nucleotides to the primer in a template-dependent manner. The signal from the detectable label is then used to identify the incorporated base, and the steps are repeated in order to determine the linear order of the nucleotides in the template. Detectable labels include radiolabels, fluorescent labels, enzymatic labels, and the like. Many techniques are known for detecting sequences, such as the Illumina NextSeq platform by cycle-end sequencing.

[0064] Machine learning models Once sequence data representing the polypeptide sequences of one or more tumor-specific neoantigens is obtained, the sequence data can be input into a machine learning platform (i.e., a model) along with the MHC molecules of a subject. The machine learning platform can generate one or more numerical probability scores that predict whether the one or more tumor-specific neoantigens are immunogenic (e.g., whether they will elicit an immune response in a subject).

[0065] MHC molecules transport and present peptides on the cell surface. MHC molecules are classified into class I and class II MHC molecules. MHC class I is present on the surface of almost all cells in the body, including most tumor cells. MHC class I proteins are usually loaded with antigens derived from endogenous proteins or pathogens present inside the cell and presented to cytotoxic T lymphocytes (i.e., CD8+). MHC class I molecules can include HLA-A, HLA-B, or HLA-C. Class II MHC molecules are present only on dendritic cells, B lymphocytes, macrophages, and other antigen-presenting cells. They mainly present peptides and are processed by helper T (Th) cells (i.e., CD4+) from external antigen sources, i.e., outside the cell. MHC class II molecules can include HLA-DPA1, HLA-DPB1, HLA-DQA1, HLA-DQB1, HLA-DRA, and HLA-DRB1. In some cases, MHC class II molecules can also be expressed on cancer cells.

[0066] MHC class I molecules and / or MHC class II molecules can be input to the machine learning platform. Typically, either MHC class I molecules or MHC class II molecules are input to the machine learning platform. In some embodiments, MHC class I molecules are input to the machine learning platform. In other embodiments, MHC class II molecules are input to the machine learning platform. In some embodiments, the MHC class I machine learning platform may be trained based on MHC class I training data. In some embodiments, the MHC class II machine learning platform may be trained based on MHC class II training data. In some embodiments, the same machine learning platform may be trained based on both MHC class I and MHC class II training data. In some embodiments, the machine learning platform may include an MHC class I model and an MHC class II mode.

[0067] MHC class I molecules bind short peptides. MHC class I molecules can accommodate peptides that are typically about 8 to about 10 amino acids long. In one embodiment, the sequence data encoding one or more tumor-specific neoantigens are short peptides that are typically about 8 to about 10 amino acids long. MHC class II molecules bind longer peptides. MHC class II molecules can accommodate peptides that are typically about 13 to about 25 amino acids long. In one embodiment, the sequence data encoding one or more tumor-specific neoantigens are long peptides that are typically about 13 to about 25 amino acids long.

[0068] The sequence data encoding one or more tumor-specific neoantigens may be about 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, or 30 amino acids in length.

[0069] The machine learning platform can predict the likelihood that one or more tumor-specific neoantigens are immunogenic (e.g., elicit an immune response).

[0070] Immunogenic tumor-specific neoantigens are not expressed in normal tissues. They can be presented to CD4+ and CD8+ T cells by antigen-presenting cells to elicit an immune response. In an embodiment, the immune response in the subject elicited by one or more tumor-specific neoantigens comprises the presentation of one or more tumor-specific neoantigens on the surface of tumor cells. More specifically, the immune response in the subject elicited by one or more tumor-specific neoantigens comprises the presentation of one or more tumor-specific neoantigens by one or more MHC molecules on tumor cells. The immune response elicited by one or more tumor-specific neoantigens is expected to be a T cell-mediated response. The immune response in the subject elicited by one or more tumor-specific neoantigens can involve one or more tumor-specific neoantigens that can be presented to T cells by antigen-presenting cells such as dendritic cells. Preferably, the one or more tumor-specific neoantigens can activate CD8+ T cells and / or CD4+ T cells.

[0071] In some embodiments, the machine learning platform can predict the likelihood that one or more tumor-specific neoantigens will activate CD8+ T cells. In other embodiments, the machine learning platform can predict the likelihood that one or more tumor-specific neoantigens will activate CD4+ T cells. In some cases, the machine learning platform can predict the antibody titers that one or more tumor-specific neoantigens can elicit. In other examples, the machine learning platform can predict the frequency of CD8+ activation by one or more tumor-specific neoantigens.

[0072] The machine learning platform may include a model trained based on training data. The training data may be obtained from different groups of subjects. The training data may include data obtained from healthy subjects as well as subjects suffering from cancer. The training data may include various data that may be used to generate a probability score indicating whether one or more tumor-specific neoantigens will induce an immune response in a subject. Exemplary training data may include data representing nucleotide or polypeptide sequences from normal tissues and / or cells, data representing nucleotide or polypeptide sequences from tumor tissues, data representing MHC peptidome sequences from normal and tumor tissues, peptide-MHC binding affinity measurements, or combinations thereof. The reference data may further include mass spectrometry data, DNA sequence data, RNA sequence data, clinical data from healthy or cancer subjects, cytokine profiling data, T cell toxicity assay data, peptide-MHC monomer or multimer data, synthetic proteins, normal and tumor human cell lines, fresh and frozen primary samples, proteomic data of single allele cell lines engineered to express a given MHC allele and then exposed to T cell assays, etc.

[0073] In some exemplary embodiments, binding affinity predictions for various samples, including corresponding "weak" labels for samples with unknown binding affinity predictions, may be extracted and added to a binding affinity training dataset. Samples with binding affinity predictions above a predefined threshold are removed, leaving a refined dataset for use in further training processes.

[0074] The machine learning platform can be a supervised learning platform, an unsupervised learning platform, or a semi-supervised learning platform. The machine learning platform can use a sequence-based approach to generate a numerical probability that one or more tumor-specific neoantigens can induce an immune response (e.g., induce a high / low antibody response or a CD8+ response). Sequence-based predictions include artificial neural networks (deep or otherwise), support vector machines, k-nearest neighbors, logistic multiple network constrained regression (LogMiNeR), regression trees, random forests, adaboost, XGBoost, or hidden Markov models, among others. These platforms require a training dataset that contains known MHC-binding peptides.

[0075] According to some embodiments, masked language modeling may be implemented in a pre-training stage, where a determined subset of peptide sequences is masked via a tokenization process, after which the classifier can predict the original token values ​​based on the existing unmasked tokens.

[0076] According to another example, the next peptide in the sequence is determined according to a pre-training process where the input sequence may be the concatenation of two peptide sequences instead of the peptide and allele sequences in the main training phase. The two peptide sequences are separated using special separation tokens, where each segment may have a different segment index and embedding. The segment sequences are provided as input to the network, and each token may indicate whether it belongs to the first sequence, the second sequence, or is a special token. The tokens can be used to train a classifier to predict whether the second peptide is the next peptide to occur in the protein or not. The peptides may be provided by two consecutive peptides of the same length from a human protein, or may be randomly sampled from different proteins.

[0077] A number of predictive programs have been used to predict whether tumor-specific neoantigens can be presented on MHC molecules and induce an immune response. Exemplary prediction programs include, for example, HLAminer (Warren et al., Genome Med., 4:95 (2012); HLA type predicted by orienting the assembly of shotgun sequence data and comparing it with the reference allele sequence database), VariantEffect Predictor Tool (McLaren et al., Genome Biol., 17:122 (2016)), NetMHCpan (Andreatta et al., Bioinformatics., 32:511-517 (2016); sequence comparison method based on artificial neural network, and predict the affinity of peptide-MHC-I type molecule), UCSC browser (Kent et al., Genome Res., 12:996-1006 (2002)), CloudNeo pipeline (Bais et al., Bioinformatics, 33:3110-2 (2017)), OptiType (Szolek et al., al., Bioinformatics, 30:3310-316 (2014)), ATHLATES (Liu C et al., Nucleic Acids Res. 41:e142 (2013)), pVAC-Seq (Hundal et al., Genome Med. 8:11 (2016), MuPeXI (Bjerregaard et al., Cancer Immunol Immunother., 66:1123-30 (2017)), Strelka (Saunders et al., Bioinformatics. 28:1811-7 (2012)), Strelka2 (Kim et al., Nat Methods.2018;15:591-4.), VarScan2 (Koboldt et al., Genome Res., 22:568-76 (2012)), Somaticseq (Fang L et al., Genome Biol., 16:197 (2015)), SMMPMBEC (Kim et al., BMC Bioinformatics., 10:394 (2009)), NeoPredPipe (Schenck RO, BMC Bioinformatics., 20:264 (2019)), Weka (Witten et al., Data mining: practical machine-learning tools and techniques. 4. th ed. Elsevier, ISBN: 97801280435578 (eBook) (2017), or Orange (Demsar et al., Orange: Data Mining Toolbox in Python., J. Mach Learn Res., 14:2349-2353 (2013). Any known predictive program can be used as a machine learning platform to generate a numerical probability score indicating whether a neoantigen will induce an immune response or not.

[0078] Depending on the machine learning platform used, additional filters can be applied to prioritize tumor-specific neoantigen candidates, including removing hypothesized (Riken) proteins, using antigen processing algorithms to remove epitopes that are unlikely to be proteolytically generated by the constitutive proteasome or immune proteasome, and prioritizing neoantigens with higher predicted binding affinity than the corresponding wild-type sequence.

[0079] The numerical probability score may be a number between 0 and 1. In embodiments, the numerical probability score may be 0, 0.0001, 0.0002, 0.0003, 0.0004, 0.0005, 0.0006, 0.0007, 0.0008, 0.0009, 0.001, 0.002, 0.003, 0.004, 0.005, 0.006, 0.007, 0.008, 0.009, 0.01, 0.02, 0.03, 0.04, 0.05, 0.06, 0.07, 0.08, 0.09, 0.10, 0.20, 0.30, 0.40, 0.50, 0.60, 0.70, 0.80, 0.90, or 1. A tumor-specific neoantigen with a high numerical probability score compared to a low numerical probability score indicates that the tumor-specific neoantigen will induce a greater immune response in a subject and is therefore more likely to be a suitable candidate for an immunogenic composition. For example, a tumor-specific neoantigen with a numerical probability score of 1 is more likely to induce a greater immune response in a subject than a tumor-specific neoantigen with a numerical probability score of 0.05. Similarly, a tumor-specific neoantigen with a numerical probability score of 0.5 is more likely to induce a greater immune response in a subject than a tumor-specific neoantigen with a numerical probability score of 0.1.

[0080] Higher numerical probability scores are preferred over lower numerical probability scores. Preferably, the tumor-specific neoantigens have a numerical probability score of at least 0.8, 0.81, 0.82, 0.83, 0.84, 0.85, 0.86, 0.87, 0.88, 0.89, 0.9, 0.95, 0.96, 0.97, 0.98, 0.99, or 1, indicating a high likelihood of eliciting an immune response in a subject.

[0081] Although a higher probability score is preferable, a lower probability score may still indicate that the tumor-specific neoantigen is capable of eliciting a sufficient immune response and is therefore likely to be a suitable candidate.

[0082] In some cases, the machine learning platform described herein can also predict the likelihood that one or more tumor-specific neoantigens are presented by MHC molecules on tumor cells. The machine learning platform can predict the likelihood that one or more tumor-specific neoantigens are presented by MHC class I molecules or MHC class II molecules.

[0083] The method for selecting one or more tumor-specific neoantigens may further comprise a step of measuring in silico the affinity of one or more tumor-specific neoantigens to bind to MHC molecules in a subject. A tumor-specific neoantigen having a binding affinity to an MHC molecule of less than about 1000 nM indicates that the one or more tumor-specific neoantigens may be suitable for an immunogenic composition. A tumor-specific neoantigen having a binding affinity to an MHC molecule of less than about 500 nM, less than about 400 nM, less than about 300 nM, less than about 200 nM, less than about 100 nM, or less than about 50 nM indicates that the one or more tumor-specific neoantigens may be suitable for an immunogenic composition. The affinity of one or more tumor-specific neoantigens to bind to an MHC molecule in a subject can predict the immunogenicity of the tumor-specific neoantigen. Alternatively, the median affinity is an effective way to predict the immunogenicity of the tumor-specific neoantigen. Median affinity can be calculated using epitope prediction algorithms such as NetMHCpan, ANN, SMM, and SMMPMBEC.

[0084] The RNA expression of one or more tumor-specific neoantigens is also quantified. The RNA expression of one or more tumor-specific neoantigens is quantified to identify one or more neoantigens that induce an immune response in the subject. There are various methods for measuring RNA expression. Known techniques that can measure RNA expression include RNA-seq, in situ hybridization (e.g., FISH), Northern blot, DNA microarray, tiling array, and quantitative polymerase chain reaction (qPCR). Other techniques known in the art can be used to quantify RNA expression. RNA includes messenger RNA (mRNA), short interfering RNA (siRNA), microRNA (miRNA), circular RNA (circRNA), transfer RNA (tRNA), ribosomal RNA (rRNA), small nucleolar RNA (snRNA), Piwi-interacting RNA (piRNA), long non-coding RNA (long ncRNA), subgenomic RNA (sgRNA), RNA from integrated or non-integrated viruses, or other RNA. Preferably, mRNA expression is measured.

[0085] The present technique can further reduce the possibility of selecting tumor-specific neoantigens that may induce an autoimmune response in normal tissues. Tumor-specific neoantigens with sequences similar to normal antigens are expected to induce an autoimmune response in normal tissues. For example, tumor-specific neoantigens that are at least 90%, at least 91%, at least 92%, at least 93%, at least 94%, at least 95%, at least 96%, at least 97%, at least 98%, or at least 99% similar to normal antigens may induce an autoimmune response. Tumor-specific neoantigens predicted to induce an autoimmune response are not prioritized as immunogenic compositions. Tumor-specific neoantigens predicted to induce an autoimmune response are typically not selected as immunogenic compositions. The method can further include measuring the ability of one or more tumor-specific neoantigens to induce immune tolerance. Tumor-specific neoantigens predicted to induce immune tolerance are not prioritized as immunogenic compositions. Tumor-specific neoantigens predicted to induce immune tolerance are not prioritized as immunogenic compositions.

[0086] Finally, one or more tumor-specific neoantigens are selected for formulation into a subject-specific immunogenic composition based on the tumor-specific score. In embodiments, at least about 1, at least about 2, at least about 3, at least about 4, at least about 5, at least about 6, at least about 7, at least about 8, at least about 9, at least about 10, at least about 11, at least about 12, at least about 13, at least about 14, at least about 15, at least about 16, at least about 17, at least about 18, at least about 19, at least about 20, at least about 25, at least about 30, at least about 35, at least about 40, at least about 50 or more tumor-specific neoantigens are selected for the immunogenic composition. Typically, at least about 10 tumor-specific neoantigens are selected. In other examples, at least about 20 tumor-specific neoantigens are selected.

[0087] II. Treatment method The present disclosure also relates to a method of treating cancer in a subject in need thereof comprising administering an individualized immunogenic composition comprising one or more tumor-specific neoantigens selected using the methods described herein.

[0088] The cancer may be any solid tumor or any hematological tumor. The methods disclosed herein are preferably suitable for solid tumors. The tumor may be a primary tumor (e.g., a tumor at the original site where the tumor first developed). Solid tumors include, but are not limited to, breast cancer tumors, ovarian cancer tumors, prostate cancer tumors, lung cancer tumors, kidney cancer tumors, stomach cancer tumors, testicular cancer tumors, head and neck cancer tumors, pancreatic cancer tumors, brain cancer tumors, melanoma tumors, etc. Hematological tumors include, but are not limited to, tumors derived from lymphomas (e.g., B-cell lymphomas) and leukemias (e.g., acute myeloid leukemia, chronic myeloid leukemia, chronic lymphocytic leukemia, and T-cell lymphocytic leukemia).

[0089] The methods disclosed herein can be used with any suitable cancerous tumor, including hematological malignancies, solid tumors, sarcomas, carcinomas, other solid tumors, and non-solid tumors. Examples of suitable cancers include, for example, acute lymphoblastic leukemia (ALL), acute myeloid leukemia (AML), adrenocortical carcinoma, anal cancer, appendix cancer, astrocytoma, basal cell carcinoma, brain tumor, bile duct cancer, bladder cancer, bone cancer, breast cancer, bronchial tumor, cancer of unknown primary, cardiac tumor, cervical cancer, chordoma, colon cancer, colorectal cancer, craniopharyngioma, ductal carcinoma, embryonal tumor, endometrial cancer, ependymoma, esophageal cancer, esthesioneuroblastoma, fibrohistiocyte tumor, and fibroblastic leukemia. melanoma, Ewing's sarcoma, eye cancer, germ cell tumor, gallbladder cancer, gastric cancer, gastrointestinal carcinoid tumor, gastrointestinal stromal tumor, gestational trophoblastic disease, glioma, head and neck cancer, hepatocellular carcinoma, histiocytosis, Hodgkin's lymphoma, hypopharyngeal cancer, intraocular melanoma, islet cell tumor, Kaposi's sarcoma, kidney cancer, Langerhans cell histiocytosis, laryngeal cancer, lip and oral cancer, liver cancer, lobular carcinoma in situ, lung cancer, macroglobulinemia, malignant fibrous histiocytoma, melanoma tumor, Merkel cell carcinoma, mesothelioma, metastatic squamous cell carcinoma with occult primary midline tract cancer involving the NUT gene, oral cavity cancer, multiple endocrine neoplasia syndrome, multiple myeloma, mycosis fungoides, myelodysplastic syndrome, myelodysplastic / myeloproliferative neoplasms, nasal cavity and paranasal sinus cancer, nasopharyngeal carcinoma, neuroblastoma, non-small cell lung cancer, oropharyngeal cancer, osteosarcoma, ovarian cancer, pancreatic cancer, papillomatosis, paraganglioma, parathyroid cancer, penile cancer, pharyngeal cancer, pheochromocytoma , pituitary tumor, pleuropulmonary blastoma, primary central nervous system lymphoma, prostate cancer, rectal cancer, renal cell carcinoma, renal pelvis and ureter cancer, retinoblastoma, rhabdoid tumor, salivary gland cancer, Sezary syndrome, skin cancer, small cell lung cancer, small intestine cancer, soft tissue sarcoma, spinal tumor, gastric cancer, T-cell lymphoma, teratoma, testicular cancer, pharyngeal cancer, thymoma, thymic cancer, thyroid cancer, urethral cancer, uterine cancer, vaginal cancer, vulvar cancer, Wilms' tumor, etc. Preferably, the cancer is melanoma, breast cancer, ovarian cancer, prostate cancer, kidney cancer, gastric cancer, colon cancer, testicular cancer, head and neck cancer, pancreatic cancer, brain cancer, B-cell lymphoma, acute myeloid leukemia, chronic myeloid neoplasm, leukemia, chronic lymphocytic leukemia, T-cell lymphocytic leukemia, bladder cancer, or lung cancer. Melanoma is of particular interest. Breast, lung and bladder cancers are also of particular interest.

[0090] The immunogenic composition stimulates the immune system of the subject, particularly the response of specific CD8+ T cells or CD4+ T cells. Interferon gamma produced by CD8+ and T helper CD4+ cells regulates the expression of PD-L1. The expression of PD-L1 on tumor cells is upregulated when attacked by T cells. Thus, tumor vaccines may induce the production of specific T cells and simultaneously upregulate the expression of PD-L1, which may limit the effectiveness of the immunogenic composition. Furthermore, when the immune system is activated, the expression of the T cell surface reporter CTLA-4 increases accordingly. CTLA-4 exerts an immunosuppressive effect by binding to the ligand B7-1 / B7-2 on antigen-presenting cells. Therefore, in some cases, the subject may further be administered an anti-immunosuppressant or an immunostimulant, such as a checkpoint inhibitor. Checkpoint inhibitors include, but are not limited to, anti-CTL4-A antibodies, anti-PD-1 antibodies, anti-PD-L1 antibodies, and the like. These checkpoint inhibitors bind to immune checkpoint proteins on T cells, relieving tumor cell inhibition of T cell function. Blockade of CTLA-4 or PD-L1 with antibodies may enhance a patient's immune response against cancer cells. CTLA-4 has been shown to be effective when followed in a vaccination regimen.

[0091] The immunogenic composition comprising one or more tumor-specific neoantigens can be administered to a subject diagnosed with cancer, a subject already suffering from cancer, a subject with recurrent cancer (i.e., recurrence), or a subject at risk of developing cancer. The immunogenic composition comprising one or more tumor-specific neoantigens can be administered to a subject who is resistant to other forms of cancer treatment (e.g., chemotherapy, immunotherapy, radiation). The immunogenic composition comprising one or more tumor-specific neoantigens can be administered to a subject prior to other standard cancer treatments (e.g., chemotherapy, immunotherapy, radiation). The immunogenic composition comprising one or more tumor-specific neoantigens can be administered to a subject in conjunction with, after, or in combination with other standard cancer treatments (e.g., chemotherapy, immunotherapy, radiation).

[0092] The subject may be a human, dog, cat, horse, or any animal in which a tumor-specific response is desired.

[0093] The immunogenic composition is administered to the subject in an amount sufficient to induce an immune response against tumor-specific neoantigens and destroy or at least partially prevent symptoms and / or complications. In embodiments, the immunogenic composition can provide a long-lasting immune response. The long-lasting immune response can be established by administering a booster immunogenic composition to the subject. The immune response against the immunogenic composition can be extended by administering a booster dose to the subject. In embodiments, at least one, at least two, at least three or more booster doses may be administered to alleviate the cancer. The first booster dose can increase the immune response by at least 50%, at least 100%, at least 200%, at least 300%, at least 400%, at least 500%, or at least 1000%. The second booster dose can increase the immune response by at least 50%, at least 100%, at least 200%, at least 300%, at least 400%, at least 500%, or at least 1000%. The third booster dose can increase the immune response by at least 50%, at least 100%, at least 200%, at least 300%, at least 400%, at least 500%, or at least 1000%.

[0094] An amount sufficient to induce an immune response is defined as a "therapeutically effective amount." Amounts effective for this use will depend, for example, on the composition, the method of administration, the stage and severity of the disease being treated, the weight and general health of the patient, and the judgment of the prescribing physician. It should be noted that immunogenic compositions can generally be used in severe disease conditions, i.e., situations that are life-threatening or potentially life-threatening, particularly when cancer has metastasized. In such cases, the attending physician may determine that it is possible and desirable to administer substantial excesses of these immunogenic compositions, taking into account the minimization of foreign material and the relatively non-toxic nature of neoantigens.

[0095] The immunogenic composition comprising one or more tumor-specific neoantigens can be administered to a subject alone or in combination with other therapeutic agents. The therapeutic agent can be, for example, a chemotherapeutic agent, radiation therapy, or immunotherapy. The appropriate therapeutic treatment can be administered for the particular cancer. Exemplary chemotherapeutic agents include aldesleukin, altretamine, amifostine, asparaginase, bleomycin, capecitabine, carboplatin, carmustine, cladribine, cisapride, cisplatin, cyclophosphamide, cytarabine, dacarbazine (DTIC), dactinomycin, docetaxel, doxorubicin, dronabinol, epoetin alfa, etoposide, filgrastim, fludarabine, fluorouracil, gemcitabine, granisetron, hydroxyurea, idarubicin, ifosfamide, These include, but are not limited to, interferon alpha, irinotecan, lansoprazole, levamisole, leucovorin, megestrol, mesna, methotrexate, metoclopramide, mitomycin, mitotane, mitoxantrone, omeprazole, ondansetron, paclitaxel (Taxol®), pilocarpine, prochlorperazine, rituximab, tamoxifen, taxol, topotecan hydrochloride, trastuzumab, vinblastine, vincristine, vinorelbine tartrate, and the like. The subject may also be administered a small molecule or targeted therapy (e.g., a kinase inhibitor). The subject may also be administered an anti-CTLA antibody, an anti-PD-1 antibody, or an anti-PD-L1 antibody. Blocking CTLA-4 or PD-L1 with antibodies may enhance the patient's immune response to cancer cells.

[0096] III. Immunogenic composition The present invention further relates to personalized (i.e., subject-specific) immunogenic compositions (e.g., cancer vaccines) that comprise one or more tumor-specific antigens selected using the methods described herein. Such immunogenic compositions can be formulated according to standard procedures in the art. The immunogenic compositions can induce specific immune responses.

[0097] The immunogenic composition can be formulated to tailor the selection and number of tumor-specific neoantigens to the particular cancer of the subject. For example, the selection of tumor-specific neoantigens can depend on the particular type of cancer, the state of the cancer, the immune status of the subject, and the MHC type of the subject.

[0098] The immunogenic composition can comprise at least 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 37, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, or more tumor-specific neoantigens. The immunogenic composition may comprise about 10-20 tumor-specific neoantigens, about 10-30 tumor-specific neoantigens, about 10-40 tumor-specific neoantigens, about 10-50 tumor-specific neoantigens, about 10-60 tumor-specific neoantigens, about 10-70 tumor-specific neoantigens, about 10-80 tumor-specific neoantigens, about 10-90 tumor-specific neoantigens, or about 10-100 tumor-specific neoantigens. Preferably, the immunogenic composition comprises at least about 10 tumor-specific neoantigens. Also preferred is an immunogenic composition comprising at least about 20 tumor-specific neoantigens.

[0099] The immunogenic composition may further comprise natural or synthetic antigens, which may enhance the immune response. Exemplary natural or synthetic antigens include, but are not limited to, pan-DR epitopes (PADRE) and tetanus toxin antigens.

[0100] The immunogenic composition may be in any form, for example, a synthetic long peptide, RNA, DNA, a cell, a dendritic cell, a nucleotide sequence, a polypeptide sequence, a plasmid, or a vector.

[0101] Tumor-specific neoantigens can also be included in viral vector-based vaccine platforms such as vaccinia, fowlpox, self-replicating alphaviruses, Maraba viruses, adenoviruses (see, e.g., Tatsis et al., Molecular Therapy, 10:616-629 (2004)), or lentiviruses (see, e.g., Hu et al., Immunol Rev., 239(1): 45-61 (2011), Sakma et al, Biochem J., 443(3):603-18 (2012)), including but not limited to second-generation, third-generation, or hybrid second / third-generation lentiviruses, any generation of recombinant lentiviruses designed to target a specific cell type or receptor. Depending on the packaging capacity of the viral vector-based vaccine platform described above, this approach can deliver one or more nucleotide sequences encoding one or more tumor-specific neoantigen peptides. The sequences may be adjacent to non-mutated sequences, separated by linkers, or preceded by one or more sequences that target intracellular compartments (see, e.g., Gros et al., Nat Med., 22 (4):433-8 (2016), Stronen et al., Science., 352(6291): 1337-1341 (2016), Lu et al., Clin Cancer Res., 20(13):3401-3410 (2014)). Upon introduction into a host, the infected cells express one or more tumor-specific neoantigens, thereby eliciting a host immune (e.g., CD8+ or CD4+) response against the one or more tumor-specific neoantigens. Vaccinia vectors and methods useful for immunization protocols are described, for example, in U.S. Pat. No. 4,722,848. Another vector is BCG (Bacille Calmette Guerin). BCG vectors are described in Stover et al. (Nature 351:456-460 (1991)).A wide variety of other vaccine vectors useful for therapeutic administration or immunization of neoantigens, as will be apparent to those of skill in the art from the disclosure herein, can also be used.

[0102] The immunogenic compositions can include individualized components depending on the personal needs of a particular subject.

[0103] The immunogenic composition herein may further comprise an adjuvant. An adjuvant is a substance that, when mixed with the immunogenic composition, increases or otherwise enhances and / or boosts the immune response to tumor-specific neoantigens, but does not generate an immune response to tumor-specific neoantigens when administered alone. The adjuvant preferably generates an immune response to neoantigens and does not cause allergic or other adverse reactions. It is contemplated herein that the immunogenic composition can be administered before, simultaneously with, or after administration of the immunogenic composition.

[0104] Adjuvants can enhance the immune response by any mechanism including, for example, lymphocyte recruitment, stimulation of B cells and / or T cells, stimulation of macrophages. When the immunogenic compositions of the invention include an adjuvant or are administered with one or more adjuvants, adjuvants that can be used include, but are not limited to, mineral salt or mineral salt gel adjuvants, particle adjuvants, particulate adjuvants, mucosal adjuvants, and immunostimulatory adjuvants. Exemplary adjuvants include aluminum salts (alum) (such as aluminum hydroxide, aluminum phosphate, and aluminum sulfate), 3-De-O-acylated monophosphoryl lipid A (MPL) (see GB2220211), MF59 (Novartis), AS03 (Glaxo SmithKline), AS04 (Glaxo SmithKline), polysorbate 80 (Tween 80; ICL Americas), imidazopyridine compounds (see International Application No. PCT / US2007 / 064857, published as International Publication No. WO2007 / 109812), imidazoquinoxaline compounds (see International Application No. PCT / US2007 / 064858, published as International Publication No. WO2007 / 109813), saponins such as QS21 (Kensil et al, in Vaccine Design: The Subunit and Adjuvant Approach (eds. Powell & Newman, Plenum Press, NY, 1995), U.S. Pat. No. 5,057,540). In some embodiments, the adjuvant is Freund's adjuvant (complete or incomplete). Other adjuvants are oil-in-water emulsions (such as squalene or peanut oil), optionally combined with an immunostimulant such as monophosphoryl lipid A (see Stoute et al, N. Engl. J. Med. 336, 86-91 (1997)).

[0105] CpG immunostimulatory oligonucleotides have also been reported to enhance the effect of adjuvants in a vaccine setting. Other TLR binding molecules, such as RNA that binds to TLR7, TLR8 and / or TLR9, can also be used.

[0106] Other examples of useful adjuvants include, but are not limited to, chemically modified CpG (e.g., CpR, Idera), poly(I:C) (e.g., polyi:CI2U), polyICLC, non-CpG bacterial DNA or RNA, immunologically active small molecules and antibodies that act therapeutically and / or as adjuvants, such as cyclophosphamide, sunitomib, bevacizumab, Celebrex (celecoxib), NCX-4016, sildenafil, tadalafil, vardenafil, sorafinib, XL-999, CP-547632, pazopam, ZD2171, AZD2171, ipilimumab, tremelimumab, SC58175, etc. In an embodiment, polyICLC is a preferred adjuvant.

[0107] The immunogenic composition may comprise one or more tumor-specific neoantigens described herein alone or together with a pharma- ceutically acceptable carrier. A suspension or dispersion of one or more tumor-specific neoantigens may be used, in particular an isotonic aqueous suspension, dispersion, or amphipathic solvent. The immunogenic composition may be sterilized and / or may contain excipients, such as preservatives, stabilizers, wetting agents and / or emulsifiers, solubilizers, salts for adjusting osmotic pressure and / or buffers, and may be prepared by means of known methods, such as conventional dispersion and suspension processes. In certain embodiments, such dispersions or suspensions may contain viscosity modifiers. The suspensions or dispersions are stored at a temperature of about 2°C to 8°C, and for preferential long-term storage, they may also be frozen and thawed immediately before use. For injection, the vaccine or immunogenic preparation may be formulated in an aqueous solution, preferably a physiologically compatible buffer such as Hank's solution, Ringer's solution, or physiological saline buffer. The solution may contain formulating agents such as suspending agents, stabilizing agents and / or dispersing agents.

[0108] In certain embodiments, the compositions described herein further comprise a preservative, such as the mercury derivative thimerosal. In certain embodiments, the pharmaceutical compositions described herein comprise 0.001%-0.01% thimerosal. In other embodiments, the pharmaceutical compositions described herein do not comprise a preservative.

[0109] The excipient can be present independent of the adjuvant. The function of the excipient is, for example, to increase the molecular weight of the immunogenic composition, to increase activity or immunogenicity, to confer stability, to increase biological activity, or to increase serum half-life. The excipient can also be used to aid in the presentation of one or more tumor-specific neoantigens to T cells (e.g., CD4+ or CD8+ T cells). The excipient can be a carrier protein such as, but not limited to, keyhole limpet hemocyanin, serum proteins such as transferrin, bovine serum albumin, human serum albumin, thyroglobulin or ovalbumin, immunoglobulins, or hormones such as insulin or palmitic acid. For human immunization, the carrier is generally a physiologically acceptable carrier that is accepted by the human body and is safe. Alternatively, the carrier can be dextran, e.g., sepharose.

[0110] Cytotoxic T cells recognize antigens in the form of peptides bound to MHC molecules, not intact foreign antigens themselves. MHC molecules themselves are present on the cell surface of antigen-presenting cells. Thus, in the presence of a trimeric complex of peptide antigen, MHC molecule, and antigen-presenting cell (APC), activation of cytotoxic T cells is possible. Beyond using one or more tumor-specific antigens to activate cytotoxic T cells, adding additional APCs containing the respective MHC molecules may enhance the immune response. Thus, in some embodiments, the immunogenic composition further comprises at least one APC.

[0111] The immunogenic composition can include an acceptable carrier (e.g., an aqueous carrier). A variety of aqueous carriers can be used, such as water, buffered water, 0.9% saline, 0.3% glycine, hyaluronic acid, and the like. These compositions can be sterilized by conventional and well-known sterilization techniques or can be sterile filtered. The resulting aqueous solution can be packaged for use as is, or lyophilized, and the lyophilized preparation is combined with a sterile solution before administration. The composition can contain pharma- ceutically acceptable auxiliary substances necessary to approximate physiological conditions, such as pH adjusting agents, buffers, tonics, wetting agents, and the like, such as sodium acetate, sodium lactate, sodium chloride, potassium chloride, calcium chloride, sorbitan monolaurate, triethanolamine oleate, and the like.

[0112] Neoantigens can also be administered via liposomes to target specific cellular tissues, such as lymphoid tissues. Liposomes also help extend half-life. Liposomes include emulsions, foams, micelles, insoluble monolayers, liquid crystals, phospholipid dispersions, lamellar layers, and the like. In these formulations, the neoantigen to be delivered is incorporated as part of a liposome, either alone or in combination with a molecule that binds to a receptor that is prevalent on lymphoid cells, such as a monoclonal antibody that binds to the CD45 antigen, or in combination with other therapeutic or immunogenic compositions. Thus, liposomes loaded with the desired neoantigen can be directed to the site of lymphoid cells, where they deliver the selected immunogenic composition. Liposomes can be formed from standard vesicle-forming lipids, which typically include neutral or negatively charged phospholipids and a sterol, such as cholesterol. The lipid is generally selected based on, for example, the size of the liposome, acid lability, and stability of the liposome in the bloodstream. Various methods are available for preparing liposomes, including those described in Szoka et al., An. Rev. Biophys. Bioeng. 9;467 (1980), U.S. Pat. Nos. 4,235,871, 4,501,728, 4,501,728, 4,837,028, and 5,019,369.

[0113] To target immune cells, ligands incorporated into the liposomes may include, for example, antibodies or fragments thereof specific for cell surface determinants of the desired immune system cells. The liposomal suspensions can be administered intravenously, topically, etc., in doses that vary depending, inter alia, on the method of administration, the peptide being delivered, and the stage of the disease being treated.

[0114] As an alternative method of targeting immune cells, components of the immunogenic composition, such as antigens (i.e., tumor-specific neoantigens), ligands, or adjuvants (e.g., TLRs), can be incorporated into polylactic-co-glycolic acid microspheres, which can entrap components of the immunogenic composition as endosomal delivery devices.

[0115] Nucleic acids encoding tumor-specific neoantigens as described herein can also be administered to patients for therapeutic or immunization purposes. Numerous methods have been employed to deliver nucleic acids to patients. For example, the nucleic acid can be delivered directly as "naked DNA". This approach is described, for example, in Wolff et al., Science 247: 1465-1468 (1990), as well as U.S. Pat. Nos. 5,580,859 and 5,589,466. Nucleic acids can also be administered using ballistic delivery, as described, for example, in U.S. Pat. No. 5,204,253. Particles consisting of DNA alone can also be administered. Alternatively, DNA can be attached to particles, such as gold particles. Approaches for nucleic acid sequence delivery include viral vectors, mRNA vectors, and DNA vectors, with or without the use of electroporation techniques. Nucleic acids can also be delivered complexed with cationic compounds, such as cationic lipids.

[0116] The immunogenic compositions provided herein can be administered to a subject by methods including, but not limited to, oral, intradermal, intratumoral, intramuscular, intraperitoneal, intravenous, topical, subcutaneous, transdermal, intranasal, inhalation routes, and by scarification (e.g., scratching the top layer of the skin using a bifurcated needle). The immunogenic compositions can be administered to the tumor site to induce a local immune response against the tumor.

[0117] The dosage of one or more tumor-specific neoantigens will depend on the type of composition, the subject's age, weight, body surface area, individual condition, individual pharmacokinetic data, and mode of administration.

[0118] Also disclosed herein are methods of producing immunogenic compositions comprising one or more tumor-specific neoantigens selected by the methods disclosed herein. The immunogenic compositions described herein can be produced using methods known in the art. For example, a method of producing a tumor-specific neoantigen or vector (e.g., a vector comprising at least one sequence encoding one or more tumor-specific neoantigens) disclosed herein includes culturing a host cell under conditions suitable for expressing the neoantigen or vector. The host cell comprises at least one polynucleotide encoding the neoantigen or vector and purifying the neoantigen or vector. Standard purification methods include chromatographic techniques, electrophoretic, immunological, precipitation, dialysis, filtration, concentration, and chromatofocusing techniques.

[0119] The host cell may include Chinese Hamster Ovary (CHO) cells, NS0 cells, yeast, or HEK293 cells. The host cell may be transformed with one or more polynucleotides comprising at least one nucleic acid sequence encoding one or more tumor-specific neoantigens or vectors disclosed herein. In certain embodiments, the isolated polynucleotide is a cDNA.

[0120] IV. Sample The methods disclosed herein include ranking one or more tumor-specific neoantigens derived from a tumor. The method of ranking one or more tumor-specific neoantigens includes obtaining sequence data derived from the tumor. Such sequence data can be obtained from a tumor sample of a subject. The tumor sample can be obtained from a tumor biopsy.

[0121] The tumor sample may be obtained from a human or non-human subject. Preferably, the tumor sample is obtained from a human. The tumor sample may be obtained from a variety of biological sources, including cancerous tumors. The tumor may be from a tumor site or from circulating tumor cells in the blood. Examples of samples include, but are not limited to, bodily fluids, tissue biopsies, blood samples, serum plasma, stool, skin samples, and the like. The source of the sample may be a solid tissue sample, such as a tumor tissue biopsy. The tissue biopsy sample may be, for example, from a biopsy of the lung, prostate, colon, skin, breast tissue, or lymph node. The sample may be, for example, a sample of bone marrow, including bone marrow aspirates and bone marrow biopsies. The sample may also be a liquid biopsy sample, such as circulating tumor cells, cell-free circulating tumor DNA, or exosomes. The blood sample may be whole blood, partially purified blood, or a portion of whole blood or partially purified blood, such as peripheral blood mononuclear cells (PBMCs).

[0122] The tumor sample described herein may be obtained directly from a subject, may be derived from a subject, or may be derived from a sample obtained from a subject, such as cultured cells derived from a body fluid or tissue sample. The tumor biopsy sample may be a fresh sample. The fresh sample may be fixed using a known fixative (e.g., formalin, Zenker fixative, or B-5 fixative) after removal from the subject. The tumor biopsy sample may also be a preserved sample, such as a frozen sample, a cryopreserved sample, of cells obtained directly from a subject or cells derived from cells obtained from a subject. Preferably, the tumor sample obtained from a subject is a fresh tumor biopsy sample.

[0123] Tumor samples can be obtained from subjects by means including but not limited to tumor biopsy, needle aspiration, scraping, surgical resection, surgical incision, venipuncture, or means known in the art.Tumor biopsy is the preferred method for tumor collection.Tumor biopsy can be performed at any cancer site, such as primary tumor or secondary tumor.Tumor biopsy from primary tumor is generally preferred.Those skilled in the art will be aware of other suitable techniques for obtaining tumor samples.

[0124] Tumor samples can be obtained from a subject in a single procedure. Tumor samples can also be obtained repeatedly from a subject over a period of time. For example, tumor samples may be obtained daily, weekly, monthly, semi-annually, or annually. Obtaining multiple samples over a period of time can aid in identifying and selecting new tumor-specific neoantigens. Tumor samples can be obtained from the same tumor or from different tumors.

[0125] Tumor samples can be obtained from the primary tumor, one or more metastases, and / or individual sites of tumor growth (e.g., bone marrow from different skeletal parts such as hips, bones, or vertebrae). Tumor samples can be taken from the same site or from different sites.

[0126] All or part of the above may be implemented on a computing environment such as that shown in Figures 1-3. Figure 1 illustrates an exemplary provider network (or "service provider system") environment, according to some embodiments. The provider network 900 may offer resource virtualization to customers via one or more virtualization services 910, which allows customers to purchase, rent, or otherwise acquire instances 912 of virtualized resources, including but not limited to compute and storage resources, implemented on devices in the provider network(s) in one or more data centers. A local Internet Protocol (IP) address 916 may be associated with the resource instance 912, the local IP address being an internal network address of the resource instance 912 on the provider network 900. In some embodiments, the provider network 900 also offers a public IP address 914 and / or a public IP address range (e.g., Internet Protocol version 4 (IPv4) or Internet Protocol version 6 (IPv6) address) that the customer may acquire from the provider 900.

[0127] Conventionally, the provider network 900, via the virtualization service 910, allows a customer of the service provider (e.g., a customer operating one or more customer networks 950A-950C including one or more customer devices 952) to dynamically map at least some public IP addresses 914 assigned or allocated to the customer to a particular resource instance 912 assigned to the customer. The provider network 900 also allows the customer to remap a public IP address 914 previously mapped to one virtualized computing resource instance 912 assigned to the customer to another virtualized computing resource instance 912 also assigned to the customer. The customer of the service provider, such as the operator of the customer network(s) 950A-950C, uses the virtualized computing resource instances 912 and public IP addresses 914 provided by the service provider, for example, to implement customer-specific applications and present the customer applications over an intermediate network 940, such as the Internet. Another network entity 920 on the intermediate network 940 may then generate traffic to a destination public IP address 914 exposed by customer network(s) 950A-950C, which is routed to the service provider data center, where it is routed through the network fabric to the local IP address 916 of the virtualized computing resource instance 912 that is now mapped to the destination public IP address 914. Similarly, response traffic from the virtualized computing resource instance 912 is routed back to the intermediate network 940 through the network fabric to the source entity 920.

[0128] A local IP address herein refers to an internal or "private" network address of, for example, a resource instance within a provider network. A local IP address may be within an address block reserved by the Internet Engineering Task Force (IETF) Request for Comments (RFC) 1918 and / or within an address block of an address format specified by IETF RFC 4193, and is changeable within the provider network. Network traffic originating from outside the provider network is not routed directly to a local IP address; instead, the traffic uses a public IP address that is mapped to the local IP address of the resource instance. The provider network includes a network device or appliance that provides network address translation (NAT) or similar functionality to perform mapping from public IP addresses to local IP addresses and vice versa.

[0129] A public IP address is an Internet-visible network address that is assigned to a resource instance by a service provider or customer. Traffic routed to a public IP address is translated, for example via a 1:1 NAT, and forwarded to the respective local IP address of the resource instance.

[0130] Some public IP addresses may be assigned to particular resource instances by the provider network infrastructure, and these public IP addresses may be referred to as standard public IP addresses, or simply standard IP addresses. In some embodiments, mapping the standard IP addresses to the local IP addresses of a resource instance is the default launch configuration for all resource instance types.

[0131] At least some public IP addresses may be assigned or obtained by customers of the provider network 900, who then assign their assigned public IP addresses to specific resource instances assigned to the customer. These public IP addresses are referred to as customer public IP addresses, or simply customer IP addresses. Instead of being assigned to resource instances by the provider network 900 as is the case with standard IP addresses, customer IP addresses are assigned to resource instances by customers, for example, via an API provided by the service provider. Unlike standard IP addresses, customer IP addresses are assigned to customer accounts and can be remapped to other resource instances by each customer as needed or desired. Customer IP addresses are associated with a customer account, not a specific resource instance, and the customer controls the IP address until the customer chooses to release the IP address. Unlike traditional static IP addresses, customer IP addresses allow customers to mask resource instance or availability zone failures by remapping a customer's public IP address to any resource instance associated with the customer's account. The customer IP address may, for example, enable a customer to address problems with the customer's resource instances or software by remapping the customer IP address to an alternative resource instance.

[0132] 2 is a block diagram of an exemplary provider network that provides storage and hardware virtualization services to customers, according to some embodiments. The hardware virtualization service 1020 provides a number of computing resources 1024 (e.g., VMs) to the customers. The computing resources 1024 may be, for example, rented or leased to customers of the provider network 1000 (e.g., to customers implementing customer network 1050). Each computing resource 1024 is provided with one or more local IP addresses. The provider network 1000 is configured to route packets from the local IP addresses of the computing resources 1024 to public Internet destinations and from public Internet sources to the local IP addresses of the computing resources 1024.

[0133] The provider network 1000 provides customer networks 1050, connected to an intermediate network 1040, e.g., via a local network 1056, with the ability to implement virtual computing systems 1092 via a hardware virtualization service 1020 connected to the intermediate network 1040 and the provider network 1000. In some embodiments, the hardware virtualization service 1020 may provide one or more APIs 1002, e.g., web services interfaces, through which the customer networks 1050 access the functionality provided by the hardware virtualization service 1020, e.g., via a console 1094 (e.g., a web-based application, a standalone application, a mobile application, etc.). In some embodiments, each virtual computing system 1092 of the customer network 1050 corresponds to a computational resource 1024 leased, rented, or otherwise provided in the provider network 1000 to the customer network 1050.

[0134] From an instance of a virtual computing system 1092 and / or another customer device 1090 (e.g., via a console 1094), a customer accesses functionality of the storage service 1010, e.g., via one or more APIs 1002, to access and store data in storage resources 1018A-1018N of a virtual data store 1016 (e.g., folders or "buckets," virtualized volumes, databases, etc.) provided by the provider network 1000. In some embodiments, a virtualized data store gateway (not shown) may be provided in the customer network 1050, which may cache at least some data locally, e.g., frequently accessed or important data, and communicate with the storage service 1010 over one or more communication channels to upload new or changed data from the local cache so that a primary store of data (the virtual data store 1016) is maintained. In some embodiments, a user can mount and access volumes of the virtual data store 1016 via a virtual computing system 1092 and / or on another customer device 1090 via a storage service 1010 acting as a storage virtualization service, and these volumes are presented to the user as local (virtualized) storage 1098.

[0135] 2, the virtualization service(s) may also be accessed from resource instances in provider network 1000 via API(s) 1002. For example, a customer, appliance service provider, or other entity may access a virtualization service in its own virtual network on provider network 1000 via API 1002 to request allocation of one or more resource instances in the virtual network or in another virtual network.

[0136] Exemplary System In some embodiments, a system for implementing some or all of the techniques described herein may include a general-purpose computer system including or configured to access one or more computer-accessible media, such as computer system 1100 illustrated in Figure 3. In the illustrated embodiment, computer system 1100 includes one or more processors 1110 coupled to system memory 1120 via an input / output (I / O) interface 1130. Computer system 1100 further includes a network interface 1140 connected to I / O interface 1130. Although Figure 3 illustrates computer system 1100 as a single computing device, in various embodiments computer system 1100 includes one computing device or any number of computing devices configured to work together as a single computer system 1100.

[0137] In various embodiments, computer system 1100 is a uniprocessor system including one processor 1110, or a multiprocessor system including several (e.g., two, four, eight, or another suitable number) processors 1110. Processor 1110 may be any suitable processor capable of executing instructions. For example, in various embodiments, processor 1110 may be a general-purpose or embedded processor implementing any of a variety of instruction set architectures (ISAs), such as the x86, ARM, PowerPC, SPARC, or MIPS-ISA, or any other suitable ISA. In a multiprocessor system, each of processors 1110 typically, though not necessarily, implements the same ISA.

[0138] The system memory 1120 stores instructions and data accessible by the processor(s) 1110. In various embodiments, the system memory 1120 is implemented with any suitable memory technology, such as random access memory (RAM), static RAM (SRAM), synchronous dynamic RAM (SDRAM), non-volatile / flash memory, or any other type of memory. In the illustrated embodiment, program instructions and data implementing one or more desired functions, such as the methods, techniques, and data described above, are shown stored in the system memory 1120 as enzyme substrate prediction service code 1125 and data 1126.

[0139] In one embodiment, the I / O interface 1130 is configured to coordinate I / O traffic between the processor 1110, the system memory 1120, and any peripheral devices in the device, including the network interface 1140 or other peripheral interfaces. In some embodiments, the I / O interface 1130 performs any necessary protocol conversions, timing conversions, or other data conversions to convert data signals from one component (e.g., the system memory 1120) into a format suitable for use by another component (e.g., the processor 1110). In some embodiments, the I / O interface 1130 includes support for devices attached via various types of peripheral buses, such as, for example, variations on the Peripheral Component Interconnect (PCI) bus standard or the Universal Serial Bus (USB) standard. In some embodiments, the functionality of the I / O interface 1130 is split into two or more separate components, such as, for example, a northbridge and a southbridge. Also, in some embodiments, some or all of the functionality of the I / O interface 1130, such as the interface to the system memory 1120, is incorporated directly into the processor 1110.

[0140] Network interface 1140 is configured to enable data exchange between computer system 1100 and other devices 1160 connected to network(s) 1150. In various embodiments, network interface 1140 supports communication over any suitable wired or wireless general-purpose data network, such as, for example, an Ethernet network. Additionally, network interface 1140 supports communication over a telecommunications / telephone network, such as an analog voice network or a digital fiber communications network, or over a storage area network (SAN), such as a Fibre Channel SAN, or over other suitable types of networks and / or protocols.

[0141] In some embodiments, computer system 1100 includes one or more offload cards 1170 (including one or more processors 1175 and possibly one or more network interfaces 1140) connected using an I / O interface 1130 (e.g., a bus implementing a version of the Peripheral Component Interconnect Express (PCI-E) standard or another interconnect such as the QuickPath Interconnect (QPI) or UltraPath Interconnect (UPI)). For example, in some embodiments, computer system 1100 may act as a host electronic device (e.g., operating as part of a hardware virtualization service) that hosts compute instances, and one or more offload cards 1170 execute a virtualization manager that may manage the compute instances executing on the host electronic device. As an example, in some embodiments, offload card(s) 1170 may perform compute instance management operations, such as pausing and / or unpausing compute instances, starting and / or terminating compute instances, performing memory transfer / copy operations, etc. These management operations, in some embodiments, are performed by the offload card(s) 1170 in cooperation with (e.g., in response to requests from) hypervisors executed by other processors 1110A-1110N of computer system 1100. However, in some embodiments, the virtualization manager implemented by the offload card(s) 1170 may service requests from other entities (e.g., from the compute instances themselves) and may not cooperate with (or service) any separate hypervisor.

[0142] In some embodiments, the system memory 1120 may be an embodiment of a computer-accessible medium configured to store the program instructions and data described above. However, in other embodiments, the program instructions and / or data may be received, transmitted, or stored in different types of computer-accessible media. In general, computer-accessible media include non-transitory storage or memory media, such as magnetic or optical media, such as disks or DVDs / CDs, connected to the computer system 1100 via the I / O interface 1130. Non-transitory computer-accessible storage media also include any volatile or non-volatile media that may be included as the system memory 1120 or another type of memory in some embodiments of the computer system 1100, such as RAM (e.g., SDRAM, double data rate (DDR) SDRAM, SRAM, etc.), read-only memory (ROM), etc. Additionally, computer-accessible media include transmission media or signals, such as electrical, electromagnetic, or digital signals, conveyed over a communication medium, such as a network and / or a wireless link, such as may be implemented via the network interface 1140.

[0143] Various embodiments discussed or suggested herein can be implemented in a variety of operating environments, which in some cases include one or more user computers, computing devices, or processing devices that can be used to operate any number of applications. User or client devices include any of a number of general purpose personal computers, such as desktop or laptop computers running standard operating systems, as well as cellular, wireless, and handheld devices that can run portable software and support a number of networking and messaging protocols. Such systems also include a number of workstations that run any of a variety of commercially available operating systems and other known applications for purposes such as development and database management. These devices can also include other electronic devices, such as dummy terminals, thin clients, gaming systems, and / or other devices that can communicate over a network.

[0144] Most embodiments utilize at least one network well known to those skilled in the art to support communications using any of a variety of widely available protocols, such as Transmission Control Protocol / Internet Protocol (TCP / IP), File Transfer Protocol (FTP), Universal Plug and Play (UPnP), Network File System (NFS), Common Internet File System (CIFS), Extensible Messaging and Presence Protocol (XMPP), AppleTalk, etc. The network(s) may include, for example, a local area network (LAN), a wide area network (WAN), a virtual private network (VPN), the Internet, an intranet, an extranet, a public switched telephone network (PSTN), an infrared network, a wireless network, and combinations thereof.

[0145] In embodiments utilizing a web server, the web server may run any of a variety of server or mid-tier applications including an HTTP server, a File Transfer Protocol (FTP) server, a Common Gateway Interface (CGI) server, a data server, a Java server, a business application server, etc. The server(s) may also execute programs or scripts in response to requests from user devices, such as by executing one or more web applications, which may be implemented as one or more scripts or programs written in any programming language, such as Java, C, C#, or C++, or a scripting language, such as Perl, Python, PHP, or TCL, as well as combinations thereof. The server(s) may also include database servers, including, but not limited to, those commercially available from Oracle, Microsoft, Sybase, IBM, etc. Database servers may be relational or non-relational (e.g., "NoSQL"), distributed or non-distributed, etc.

[0146] The environments disclosed herein may include various data stores and other memory and storage media, as described above. These may reside in a variety of locations, such as storage media local to (and / or resident within) one or more computers, or remote from any or all of the computers across a network. In a particular set of embodiments, information may reside within a storage area network ("SAN"), familiar to those skilled in the art. Similarly, any necessary files to perform functions belonging to a computer, server, or other network device may be stored locally and / or remotely, as appropriate. Where the system includes computerized devices, each such device may include hardware elements that may be electrically coupled via a bus, including, for example, at least one central processing unit (CPU), at least one input device (e.g., a mouse, keyboard, controller, touch screen, or keypad), and / or at least one output device (e.g., a display device, printer, or speaker). Such a system also includes one or more storage devices, such as disk drives, optical storage devices, and solid-state storage devices, such as random access memory (RAM) or read-only memory (ROM), as well as removable storage devices, memory cards, flash cards, and the like.

[0147] Such devices also include computer-readable storage medium readers, communication devices (e.g., modems, network cards (wireless or wired), infrared communication devices, etc.), and working memories, as described above. The computer-readable storage medium reader can be connected to or configured to accept computer-readable storage media, representing remote, local, fixed and / or removable storage devices and storage media for temporarily and / or more permanently containing, storing, transmitting and retrieving computer-readable information. The systems and various devices also typically include several software applications, modules, services or other elements located in at least one working memory device, including an operating system and application programs such as client applications or web browsers. It should be understood that alternative embodiments include variations or modifications of the above. For example, customized hardware may also be used and / or certain elements may be implemented in hardware, software (including portable software such as applets), or both. Additionally, connections to other computing devices, such as network input / output devices, may be used.

[0148] Storage media and computer readable media for containing the code or portions of code may include any suitable media known or used in the art, including storage media and communication media, such as, but not limited to, volatile and non-volatile media, removable and non-removable media implemented in any manner or technology for storage and / or transmission of information, such as computer readable instructions, data structures, program modules or other data, including RAM, ROM, Electrically Erasable Programmable Read Only Memory ("EEPROM"), flash memory or other memory technology, compact disk read only memory ("CD-ROM"), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and that can be accessed by the system devices. Based on the disclosure and teachings provided herein, one of ordinary skill in the art will recognize other manners and / or methods of implementing the various embodiments.

[0149] Various embodiments have been described above. For purposes of explanation, specific configurations and details are set forth to provide a thorough understanding of the embodiments. However, it should be understood by those skilled in the art that the embodiments may be practiced without the specific details. Furthermore, well-known features may be omitted or simplified so as not to obscure the described embodiments.

[0150] Bracketed text and dashed (e.g., large dashes, small dashes, dots, dots) blocks are used herein to describe optional operations that add additional features to some embodiments. However, such notations should not be understood to mean that these are the only options or optional operations and / or that solid blocks are not optional in a particular embodiment.

[0151] A reference number with a suffix letter may be used to indicate that there may be one or more instances of the referenced entity in various embodiments, and that if multiple instances are present, each need not be identical, but may instead share some common characteristics or operate in the same manner. Furthermore, a particular suffix used does not imply that a particular number of entities are present, unless specifically indicated to the contrary. Thus, two entities using the same or different suffix letters may or may not have the same number of instances in various embodiments.

[0152] References to "one embodiment," "embodiment," "exemplary embodiment," etc. indicate that the embodiment being described may include a particular feature, structure, or characteristic, but not every embodiment necessarily includes the particular feature, structure, or characteristic. Moreover, such phrases do not necessarily refer to the same embodiment. Moreover, when a particular feature, structure, or characteristic is described in connection with an embodiment, it is believed to be within the knowledge of one of ordinary skill in the art to include such feature, structure, or characteristic in connection with other embodiments, whether or not explicitly stated.

[0153] Additionally, in the various embodiments described above, unless otherwise indicated, disjunctive language, such as the phrase "at least one of A, B, or C," is intended to be understood to mean either A, B, or C, or any combination thereof (e.g., A, B, and / or C). Thus, disjunctive language is not intended to imply, and should not be understood to imply, that at least one A, at least one B, or at least one C, respectively, must be present in a given embodiment.

[0154] Accordingly, the specification and drawings are to be regarded in an illustrative rather than a restrictive sense, however it will be apparent that various modifications and changes may be made therein without departing from the broader spirit and scope of the present disclosure as set forth in the appended claims.

[0155] Equivalent

[0156] Other suitable modifications and adaptations of the methods of the present invention described herein will be apparent and may be made using appropriate equivalents without departing from the scope of the present disclosure or embodiments, as will be readily apparent to those skilled in the art. Although certain compositions and methods have been described in detail above, the same will be more clearly understood by reference to the following examples, which are presented for illustrative purposes only and are not intended to be limiting. EXAMPLES

[0157] The following examples are for illustrative purposes and are not intended to limit the scope of the invention.

[0158] Example 1: [Table 1] [Table 2] [Table 3] [Table 4]

[0159] Example 1 illustrates predicted mutated epitopes for short MHC class I vaccine peptide candidates and exemplary variants according to an exemplary embodiment. In this example, the letter "H" in a box represents the mutated subsequence of the vaccine peptide sequence "FVLQHLVFL". One or more mutated epitopes can be predicted and an immunogenicity score can be generated by one or more methods described herein. According to this example, the immunogenicity score indicates the probability that at least one epitope in a long peptide sequence is immunogenic to at least one MHC class I allele of a subject. The MHC class I binding score may indicate the probability that the peptide binds to at least one MHC class I allele of a subject. Additionally, the length may indicate the number of amino acids in the sequence, which can be used to distinguish between short and long neoantigens. The MHC class I immunogenicity binding score can be determined by multiplying the MHC class I immunogenicity score by the MHC class I binding score. The RNA TPM indicates the number of normalized RNA reads per gene length and sequencing depth in TPM. Maximum coding sequence coverage may refer to the number of RNA reads that cover the vaccine peptide sequence.

[0160] Example 2: [Table 5] [Table 6] [Table 7] [Table 8]

[0161] Example 2 shows predicted mutated epitopes for another short MHC class I vaccine peptide candidate and exemplary variants according to an exemplary embodiment. In this example, the boxed letter "C" represents the mutated subsequence of the vaccine peptide sequence "KACHYHSYNGW".

[0162] Example 3: [Table 9] [Table 10] [Table 11] [Table 12]

[0163] Example 3 shows predicted mutated epitopes for additional short MHC class I vaccine peptide candidates and exemplary variants. In this example, the boxed letter "H" represents the mutated subsequence of the vaccine peptide sequence "REEENHSFL".

[0164] Example 4: [Table 13] [Table 14] [Table 15] [Table 16-1] [Table 16-2] [Table 16-3]

[0165] Example 5: [Table 17] [Table 18]

[0166] In Examples 4 and 5 above, the short sequence "FVLQHLVFL" from Example 1 was used to generate the sequences of the long MHC class I vaccine peptide in Example 4 and the long MHC class II vaccine peptide in Example 5. Both contain the same short subsequence in the center of the sequence (e.g., the letter "H"). As described herein, amino acids are added to both sides of the short subsequence according to the longest neoantigen. This ensures that there is the first greatest number of amino acids adjacent to each side of the mutated amino acid. Predicted mutated epitopes are generated or determined for both the MHC class I vaccine peptide and the MHC class II vaccine peptide, along with corresponding immunogenicity scores. The MHC class I immunogenicity score indicates the probability that at least one epitope in the long peptide sequence is immunogenic for at least one of the subject's MHC class I alleles. The MHC class II immunogenicity score indicates the probability that at least one epitope in the long peptide sequence is immunogenic for at least one of the subject's MHC class II alleles.

[0167] Example 6: [Table 19] [Table 20] [Table 21] [Table 22-1] [Table 22-2] [Table 22-3]

[0168] Example 7: [Table 23] [Table 24]

[0169] In Examples 6 and 7 above, the short sequence "KACHYHSYNGW" of Example 2 was used to generate the sequences of the long MHC class I vaccine peptide of Example 6 and the long MHC class II vaccine peptide of Example 7. Both contain the same short partial sequence as Example 2 in the center of the sequence (e.g., the letter "C"). Predicted variant epitopes are generated or determined for both the MHC class I vaccine peptide and the MHC class II vaccine peptide, along with corresponding immunogenicity scores. The MHC class I immunogenicity score indicates the probability that at least one epitope in the long peptide sequence is immunogenic for at least one of the subject's MHC class I alleles. The MHC class II immunogenicity score indicates the probability that at least one epitope in the long peptide sequence is immunogenic for at least one of the subject's MHC class II alleles.

[0170] Example 8: [Table 25] [Table 26] [Table 27] [Table 28-1] [Table 28-2]

[0171] Example 9: [Table 29] [Table 30-1] [Table 30-2]

[0172] In Examples 8 and 9 above, the short sequence "REEENHSFL" of Example 3 was used to generate the sequences of the long MHC class I vaccine peptide of Example 8 and the long MHC class II vaccine peptide of Example 9. Both contain the same short partial sequence as Example 3 in the center of the sequence (e.g., the letter "H"). Predicted variant epitopes are generated or determined for both the MHC class I vaccine peptide and the MHC class II vaccine peptide, along with corresponding immunogenicity scores. The MHC class I immunogenicity score indicates the probability that at least one epitope in the long peptide sequence is immunogenic for at least one of the subject's MHC class I alleles. The MHC class II immunogenicity score indicates the probability that at least one epitope in the long peptide sequence is immunogenic for at least one of the subject's MHC class II alleles.

Claims

1. 1. A method for ranking tumor-specific neoantigens from a tumor of a subject with respect to a subject-specific immunogenic composition, comprising: a) identifying a plurality of somatic mutations present in said tumor; b) for each individual somatic mutation among the plurality of somatic mutations, i) determining a best short neoantigen from the first plurality of short neoantigens based at least in part on an immunogenicity score of the best short neoantigen, wherein each individual neoantigen in the first plurality of short neoantigens has 8, 9, 10, or 11 amino acids, and the immunogenicity score represents a quality of the best short neoantigen; ii) determining a best long neoantigen from the first plurality of long neoantigens based at least in part on an immunogenicity score of the best long neoantigen, wherein each individual neoantigen in the first plurality of long neoantigens has a length in the range of 15 to 30 amino acids, and the immunogenicity score represents a quality of the best long neoantigen; iii) adding the best short neoantigen to a short neoantigen candidate list; iv) adding the best long neoantigen to a long neoantigen candidate list; c) performing b) on the plurality of somatic mutations, wherein the completed short neoantigen candidate list includes the best short neoantigen for each of the plurality of somatic mutations, and the completed long neoantigen candidate list includes the best long neoantigen for each of the plurality of somatic mutations; d) ranking the short neoantigen candidate list in descending order of immunogenicity score; e) ranking the long neoantigen candidate list in descending order of immunogenicity score, wherein the short neoantigen candidate list and the long neoantigen candidate list comprise at least one MHC class I epitope and at least one MHC class II epitope.

2. For each individual somatic mutation, identifying the longest neoantigen sequence that contains the mutated amino acid; and identifying the first plurality of short neoantigens from the longest neoantigen sequence, wherein each individual neoantigen in the first plurality of short neoantigens has the mutated amino acid and a minimum number to a maximum number of amino acids.

3. 3. The method of claim 2, further comprising determining a neoantigen-allele score for each of the first plurality of short neoantigens for each individual allele in a plurality of HLA class I alleles present in the subject.

4. 4. The method of claim 3, wherein the neoantigen-allele scores between individual neoantigens and the individual alleles in the first plurality of short neoantigens are based at least in part on a probability that the individual neoantigen is presented by the individual allele, but a germline sibling of the individual neoantigen is not presented by the individual allele, optionally wherein the probability is determined at least in part based on data from an MHC class I machine learning model trained to determine the probability that a given allele in the plurality of HLA class I alleles presents a particular antigen.

5. 4. The method of claim 3, wherein for any two neoantigens in the initial plurality of short neoantigens, one of the two neoantigens includes the other, and further comprising removing from the initial plurality of short neoantigens the neoantigen of the two neoantigens that has a lower neoantigen allele score for the individual allele.

6. 6. The method of claim 5, further comprising identifying a short subsequence, the short subsequence being the shortest subsequence of the longest neoantigen sequence that includes all neoantigens in the initial plurality of short neoantigens, and no neoantigen in the initial plurality of short neoantigens is included in another neoantigen in the initial plurality of short neoantigens.

7. 7. The method of claim 6, further comprising determining a probability that the individual allele represents at least one neoantigen in a short neoantigen group and does not represent a germline sibling of the at least one neoantigen.

8. determining an immunogenicity score for each individual neoantigen in the group of short neoantigens, the immunogenicity score being based at least in part on the probability that at least one allele in the plurality of HLA class I alleles of the subject presents the individual neoantigen and does not present a germline sibling of the individual neoantigen; or 3. The method of claim 2, further comprising determining a probability that at least one allele in the plurality of HLA class I alleles of the subject presents at least one neoantigen in the group of short neoantigens and does not present a germline sibling of the at least one neoantigen.

9. identifying an extended sequence, the extended sequence being a subsequence of the longest neoantigen that includes the short subsequence and a first largest number of amino acids on each side of the mutated amino acid; 7. The method of claim 6, further comprising identifying a long neoantigen from the extended sequence, the long neoantigen having a length between the length of the short subsequence and a length of a second maximum number of amino acids.

10. 10. The method of claim 9, wherein the first maximum number is 29, or the second maximum number is 30, or the method further comprises removing non-manufacturable neoantigens from the population of long line neoantigens, or the method further comprises determining an immunogenicity score for each individual neoantigen in the population of long line neoantigens, the immunogenicity score being based at least in part on a probability that at least one allele in the subject's plurality of HLA class II alleles will present the individual neoantigen and not a germline sibling of the individual neoantigen, and the probability being optionally determined based at least in part on data from an MHC class II machine learning model trained to determine the probability that a given allele in the plurality of HLA class II alleles will present a particular antigen.

11. 2. The method of claim 1, further comprising trimming the top ranked long neoantigen candidates from the long neoantigen candidate list to a predetermined number based on immunogenicity scores.

12. selecting a subgroup of long neoantigen candidates from said long neoantigen candidate list based at least in part on manufacturability; removing any neoantigens included in any of the long neoantigen candidate subgroups from the short neoantigen candidate list; and forming a subject-specific immunogenic composition comprising one or more neoantigens from the long-chain neoantigen candidate list, and optionally administering the subject-specific immunogenic composition to the subject.

13. 2. The method of claim 1, further comprising trimming the top ranked short neoantigen candidates from the short neoantigen candidate list to a predetermined number based on immunogenicity scores, and selecting a subgroup of short neoantigen candidates from the short neoantigen candidate list based at least in part on manufacturability.

14. The method of claim 1 or 13, further comprising forming a subject-specific immunogenic composition comprising one or more neoantigens from the short neoantigen candidate list, and optionally further comprising administering the subject-specific immunogenic composition to the subject.

15. the first plurality of short neoantigens comprises short polypeptides comprising at least one MHC class I epitope associated with the subject; or the first plurality of long neoantigens comprises a long polypeptide comprising at least one MHC class I epitope and at least one MHC class II epitope associated with the subject; or the first plurality of short neoantigens and the first plurality of long neoantigens are derived from the tumor and contain the individual somatic mutations; or the immunogenicity score is determined based at least in part on data from a machine learning model; or The best short neoantigen for the individual somatic mutation is the short neoantigen with the highest immunogenicity score among the first plurality of short neoantigens for the individual somatic mutation; or 2. The method of claim 1, wherein the best long neoantigen for the individual somatic mutation is the long neoantigen with the highest immunogenicity score among the first plurality of long neoantigens for the individual somatic mutation.

Images