Methods for optimizing tumor vaccine antigen coverage for heterogeneous malignancies - Patents.com

Abstract

Disclosed herein are methods for selecting tumor-specific neoantigens from a subject's tumor that are suitable for a subject-specific immunogenic composition.

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of U.S. Provisional Application No. 63 / 161,023, filed March 15, 2021, the entire contents of which are incorporated herein by reference.

[0002] Reference to sequence listing This application contains a sequence listing in computer readable form, which is incorporated herein by reference. The ASCII copy was created on March 14, 2022, is named 146401_091707_SL.txt, and is 14,044 bytes in size. [Background technology]

[0003] Cancer is the leading cause of death worldwide, accounting for one in four of all deaths. Siegel et al., CA: A Cancer Journal for Clinicians, 68:7-30 (2018). In 2018, there were 18.1 million new cancer cases and 9.6 million cancer-related deaths. Bray et al., CA: A Cancer Journal for Clinicians, 68(6):394-424. There are numerous existing standard-of-care cancer therapies, including ablative techniques (e.g., surgery and radiation) and chemical techniques (e.g., chemotherapy agents). Unfortunately, such therapies are frequently associated with adverse prognostic outcomes, toxic side effects, and very high costs, as well as uncertain efficacy.

[0004] Cancer immunotherapy (e.g., cancer vaccines) has emerged as a promising cancer treatment. The goal of cancer immunotherapy is to harness the immune system for selective destruction of 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. Several clinical trials targeting tumor-associated antigens have failed to demonstrate sustained beneficial effects compared to standard treatments. Li et al.,Ann Oncol.,28(Suppl 12):xii11-xii17(2017).

[0005] Neoantigens represent attractive targets for cancer immunotherapy. Neoantigens are non-self proteins with individual specificity. Neoantigens originate from random somatic mutations in tumor cell genomes and are not expressed on the surface of normal cells (ibid.). Because neoantigens are exclusively expressed on tumor cells and therefore do not induce central immune tolerance, cancer vaccines targeting cancer neoantigens have potential advantages including reduced central immune tolerance and improved safety profile (ibid.).

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

[0007] Thus, efficiently and accurately predicting, prioritizing, and selecting neoantigen candidates for immunogenic compositions remains a challenge. Thus, there is a great 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 to be suitable for an effective immunogenic composition.

[0008] Summary of the Invention The present disclosure relates to novel methods for selecting suitable tumor-specific peptides for personalized (i.e., subject-specific) immunogenic compositions that provide coverage for heterogeneous malignancies. The present disclosure also relates to methods for treating cancer in a subject in need thereof by administering an immunogenic composition comprising a tumor-specific peptide selected using a novel approach for selecting a tumor-specific peptide and formulating an immunogenic composition comprising the selected tumor-specific peptide for optimal coverage for heterogeneous malignancies.

[0009] Suitable tumor-specific peptides are those that are predicted to be expressed in sufficient amounts to elicit an immune response in subjects, and optionally represent sufficient diversity across tumors and have relatively high manufacturing feasibility. The method of the present invention takes an initial set of peptides determined from tumor sequence data and selects a set therefrom for inclusion in a personalized immunogenic composition, such that the immunogenic composition provides optimal coverage across different tumor subclones while also performing well in terms of other quality factors such as cell surface presentation, binding affinity and immunogenic response. Optimizing peptide selection is particularly important due to the constraint that only a certain number of peptides can be included in the final product.

[0010] The techniques of the present invention utilize a list of peptides present in the tumor, a list of subclones present in the tumor, and a mapping between peptides and subclones indicating the probability that a given peptide belongs to a given subclone. A set of peptides is selected from the list of peptides based on an objective function that aims to maximize a value corresponding to the sum or product of the subclone scores across all subclones in the list of subclones. The subclone scores of the individual subclones are based on the probability that at least one of the selected peptides belongs to an individual subclone. The subclone scores of the individual subclones are based on the probability that at least one of the selected peptides belongs to an individual subclone and can be used to estimate or predict how mutations may cluster within a tumor. In some embodiments, the subclone scores of the individual subclones are based at least in part on individual peptide-subclone scores across the set of selected peptides. The individual peptide-subclone scores are based at least in part on the probability that an individual peptide of the set of selected peptides belongs to an individual subclone. The individual peptide-subclone scores are based at least in part on the probability that an individual peptide of the set of selected peptides belongs to an individual subclone and can be associated with a cancer cell fraction or cell incidence. According to an embodiment, the cell fractions may represent the fraction of cancers that contain mutations (e.g., mutation A is present in about 50% of cancers, mutation B is present in about 25% of cancers, and mutation C is present in about 25% of cancers). Individual peptide-subclone scores may additionally be based at least in part on the quality scores of the individual peptides, including various other characteristics of the peptides, such as probability of presentation, binding affinity, and / or immunogenic response. Cellular frequencies or cell fractions may be reorganized into a hierarchy that indicates whether a mutation occurs in the presence of other mutations or whether the mutation occurs in isolation.

[0011] The immunogenic composition formulated at least in part based on the techniques 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 synthetic long peptide. The immunogenic composition may further comprise an adjuvant.

[0012] 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 may be suitable for treating any number of cancers. The tumor may be from 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 leukemia, chronic lymphocytic leukemia, T-cell lymphocytic leukemia, bladder cancer or lung cancer. Preferably, the cancer is melanoma, breast cancer, lung cancer and bladder cancer.

[0013] Various embodiments according to the present disclosure will now be described with reference to the drawings. [Brief description of the drawings]

[0014] [Figure 1] 1 illustrates an example 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 for implementing some or all of the techniques described herein, according to some embodiments. [Figure 4] 1 is an example of a method that can be used to implement aspects of various embodiments. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] The present disclosure relates to a novel approach for selecting tumor-specific peptides for optimal coverage of heterogeneous malignancies 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 a tumor-specific peptide formed using the novel approach for selecting a tumor-specific peptide and formulating an immunogenic composition comprising the selected tumor-specific peptide.

[0016] In creating personalized cancer immunogenic compositions that target unique mutations present in a subject's tumor, a subset of neoantigens present in the tumor is selected for inclusion in the immunogenic composition. Thus, the methods of the present invention allow for the selection of a set of peptides that will create a viable and effective immunogenic composition. Notably, not only is each tumor unique, but within each tumor there are distinct populations of cells that have common mutations that may or may not be shared between groups. This is known as "tumor heterogeneity." Generally, tumors grow from one (or a few) tumor cells. Over time, various somatic mutations accumulate in certain populations of cells, but not uniformly across all populations of cells. Each of these distinct populations may be referred to as a "subclone." One or more of the methods described herein can be utilized to estimate how mutations cluster within a tumor. The methods of the present invention provide for the selection of peptides that have broad coverage across many tumor subclones.

[0017] All publications and patents cited in this disclosure are incorporated by reference in their entirety. To the extent that any material incorporated by reference contradicts or is inconsistent with this specification, this specification supersedes any such material. The citation of any reference herein is not an admission that such reference is prior art to this disclosure. When a range of values ​​is expressed, it includes embodiments using any specific value within that range. Furthermore, reference to values ​​stated in a range includes each and every value within that range. All ranges are inclusive of their endpoints and are combinable. When values ​​are expressed as approximations, by use of the preceding "about," it will be understood that the particular value forms another embodiment. Reference to a particular numerical value includes at least that particular value, unless the context clearly dictates otherwise. The use of "or" will mean "and / or" unless the specific context of the use dictates otherwise.

[0018] Throughout the specification and claims, various terms relating to the aspects of the specification are used. Such terms should be given their ordinary meaning in the art unless otherwise indicated. Other specifically defined terms should be interpreted consistent with the definitions provided herein. The techniques and procedures described or referenced herein are generally well understood and commonly employed using conventional methodology by those skilled in the art, such as the widely used molecular cloning 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. Where necessary, unless otherwise noted, procedures involving the use of commercially available kits and reagents are generally carried out according to manufacturer-defined protocols and conditions.

[0019] As used herein, the singular forms "a," "an," and "the" include the plural forms unless the content clearly dictates otherwise. Terms such as "include," "such as," and the like are intended to convey inclusion but not limitation, unless specifically indicated otherwise.

[0020] Unless otherwise indicated, the terms "at least," "less than," and "about" or similar terms preceding a series of elements or ranges are to 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.

[0021] The term "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 and proliferation rate, and / or certain morphological features. In many cases, cancer may be in the form of a tumor or mass, but may exist alone in a subject or circulate in the bloodstream as independent cells, such as leukemia or lymphoma cells. The term cancer includes all types of cancer and metastasis, including hematological malignancies, solid tumors, sarcomas, carcinomas, and other solid 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 cancer, hepatocellular carcinoma, gastrointestinal cancer, pancreatic cancer, glioblastoma, cervical cancer, ovarian cancer, liver cancer, bladder cancer, hepatoma, 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 carcinoma, kidney cancer, liver cancer, prostate cancer, vulvar cancer, thyroid cancer, hepatoma and various types of head and neck cancer. Triple-negative breast cancer refers to breast cancer that is negative for the expression of genes for estrogen receptor (ER), progesterone receptor (PR) and Her2 / neu. Hormone receptor positive breast cancer refers to breast cancer that is positive for at least one of the following: ER or PR, and negative for Her2 / neu (HER2).

[0022] The term "neoantigen" as used herein refers to an antigen that has at least one change that differs from the corresponding parent antigen, for example, through a mutation in a tumor cell or a post-translational modification specific to a tumor cell. The mutation may include a frameshift, an indel, a missense or nonsense substitution, a splice site change, a genomic rearrangement or gene fusion, or any genomic expression change that gives rise to a neoantigen. The mutation may include a splice mutation. The post-translational modification specific to a tumor cell may include an aberrant phosphorylation. The post-translational modification specific to a tumor cell may also include a proteasome-generated splice antigen. See Lipe et al., Science, 354(6310):354:358 (2016). In general, point mutations account for about 95% of mutations in tumors and indels, with frameshift mutations accounting for the remainder. See Snyder et al., N Engl J Med., 371:2189-2199 (2014).

[0023] As used herein, the term "tumor-specific neoantigens" refers to neoantigens that are present in specific tumor cells or tissues.

[0024] As used herein, the term "germline sibling" refers to a germline antigen that represents the non-mutated peptide equivalent of the corresponding neoantigen.

[0025] As used herein, the term "next-generation sequencing" or "NGS" refers to a sequencing technology that has increased throughput compared to older methods (e.g., Sanger sequencing) and has the capacity to generate hundreds of thousands of sequence reads at a time.

[0026] As used herein, the term "neural network" refers to a machine learning model for classification or regression that consists of multiple layers of linear transformations followed by element-wise nonlinearities typically trained via stochastic gradient descent and backpropagation.

[0027] As used herein, the term "subject" refers to any animal, such as any mammal, including, but not limited to, humans, non-human primates, rodents, etc. In some embodiments, the mammal is a mouse. In some embodiments, the mammal is a human.

[0028] As used herein, the term "tumor cell" refers to any cell that is or is derived from a cancer cell. The term "tumor cell" can also refer to a cell that exhibits cancer-like properties, such as uncontrolled reproduction, resistance to anti-growth signals, the ability to metastasize, and loss of the ability to undergo programmed cell death.

[0029] The term "subclone" as used herein refers to a subpopulation of cells that are descended from another clone, but have differentiated by accumulating mutations.

[0030] Further description of the methods, and guidance for practicing the methods, is provided herein.

[0031] I. Methods for selecting tumor-specific peptides for subclone coverage Disclosed herein is a method for selecting tumor-specific peptides suitable for subject-specific immunogenic compositions from a subject's tumor. Figure 4 shows an exemplary method of the embodiment provided herein. Suitable tumor-specific peptides are peptides that provide broad coverage across many tumor subclones, are likely to be presented on the cell surface of the tumor, are likely to be immunogenic, are predicted to be expressed in sufficient amounts to elicit an immune response in the subject, and optionally represent sufficient diversity across tumors and / or have relatively high manufacturing feasibility. The method of the present invention provides a technique for selecting a group of peptides (e.g., about 19, 20, 30 or any specified number of groups) for that purpose.

[0032] The set of peptides can be selected from a list of initial peptides. The list of initial peptides can be determined based on genomic sequence data of the tumor and the subject. In general, sequence data representing the polypeptide sequences of one or more tumor-specific peptides is determined by subjecting the tumor sample to sequence analysis. In some embodiments, obtaining sequence data includes receiving or accessing stored data from a previously performed sequencing. The sequence data can be, for example, exome sequence data, transcriptome sequence data, whole genome nucleotide sequence data, nucleotide sequence data, or polypeptide sequence data. Various methods of obtaining sequence data for tumors and subjects can be used in the methods described herein. Some exemplary sequencing methods are described in more detail below.

[0033] Once sequence data representing the polypeptide sequences of one or more tumor-specific peptides is obtained, the sequence data can be analyzed in conjunction with the subject's MHC molecules to identify and select candidate peptides for inclusion in an immunogenic composition for the subject. In some embodiments, the initial list of peptides is identified using a sliding window that spans each somatic mutation. In some embodiments, sequencing and identification of peptides present in the tumor can be performed prior to the techniques of the present invention. The sequencing and / or determination of peptides present can be performed by the same party / entity that performs the selection technique or by a different party / entity. In some embodiments, the initial list of peptides is received from a client device (e.g., a third-party device).

[0034] Additionally, each identified peptide has a quality score that may be based on the presentation probability, binding affinity, immunogenic response of the peptide, or a combination thereof. In some embodiments, the quality score is based at least in part on the predicted presentation probability. In some embodiments, the quality score is based at least in part on the predicted binding affinity. In some embodiments, the predicted presentation probability, predicted binding affinity, and predicted presentation probability are determined by one or more machine learning models and the subject's HLA class I and / or HLA class II alleles. 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 binding affinity between Class II alleles and a given peptide. In some embodiments, the quality score is based at least in part on the predicted immunogenic response. In some embodiments, the quality score is based at least in part on a combination of the predicted presentation probability, predicted binding affinity, and predicted presentation probability. The MHC class I and class II machine learning models used to determine such scores are described in more detail below.

[0035] In addition to the list of peptides present in the tumor, the selection technique of the present invention also utilizes a list of subclones present in the tumor. The growth of the tumor originates from one (or a few) of the tumor cells. Over time, certain groups of these cells accumulate various somatic mutations, while others do not. Each of these separate groups is a subclone. The subclones present in the tumor can be determined through various methods. For example, probabilistic methods for detecting subclones from whole exome or whole genome sequencing can be performed using Pyclone (Roth et al., 2014). In general, external resources can be used to predict how many subclones will be present and which subclone(s) each mutation and associated peptide belongs to. In some embodiments, the initial list of subclones is received from a client device (e.g., a third-party device).

[0036] In some embodiments, identifying subclones is probabilistic, meaning there is a probability or possibility that a particular subclone is present in the tumor. Thus, a subclone is considered "identified" or "present" when the probability meets a threshold or other decision cutoff. Identified peptides are mapped to the identified subclone to which they belong. For example, a particular peptide may be considered to be part of a particular subclone. In some cases, a peptide may belong to multiple subclones. Some subclones may not have any member peptides. The mapping of which peptides belong to which subclone may also be probabilistic, meaning there is a certain probability that a peptide belongs to a particular subclone. Thus, the mapping of peptides to subclones includes a probability of membership (i.e., a membership probability) between any peptide and any subclone. The membership probability may be expressed as a value between 0 and 1. In some embodiments, the mapping of peptides and subclones is received from a client device (e.g., a third-party device).

[0037] FIG. 4 illustrates a method for selecting tumor-specific peptides for a subject-specific immunogenic composition from a tumor of a subject. First, a list of peptides determined to be present in the tumor is obtained 410. For example, "obtaining" may include performing genetic sequencing of the tumor to identify peptides, or simply accessing this stored information. Each of the peptides in the list has a quality score that may be based on the presentation probability of the peptide, the binding affinity of the peptide, the immunogenic response of the peptide, or a combination thereof, among other possible characteristics. The quality score may range from 0 to 1, inclusive. In addition, a list of subclones determined to be present in the tumor is also obtained 420. Similarly, "obtaining" may include accessing stored information or performing a process to identify subclones. A mapping of peptides to subclones is also obtained 430. The mapping indicates which subclone(s) the peptide belongs to and the membership probability between each subclone-peptide combination.

[0038] Using the list of peptides, the list of subclones, and the mapping between peptides and subclones (i.e., membership probabilities), a set of peptides is selected 440 from the list of peptides based on an objective function. The objective function aims to maximize a value corresponding to the sum or product of the subclone scores across all subclones in the list of subclones. More specifically, the subclone scores of the individual subclones are based on the probability that at least one of the selected peptides belongs to the individual subclone. In some embodiments, the subclone scores of the individual subclones are based on the probability that at least one of the selected peptides belongs to the individual subclone and may be utilized to estimate or predict how mutations may cluster within a tumor. In some embodiments, the subclone scores of the individual subclones are based at least in part on the individual peptide-subclone scores across the set of selected peptides, where the individual peptide-subclone scores are based at least in part on the probability that an individual peptide of the set of selected peptides belongs to an individual subclone. The subclone scores may be associated with a cancer cell fraction or cell incidence. According to an embodiment, the cell fraction may represent the fraction of cancers that contain mutations (e.g., mutation A is present in about 50% of cancers, mutation B is present in about 25% of cancers, and mutation C is present in about 25% of cancers). The individual peptide-subclone score may additionally be based at least in part on the quality score of the individual peptides. For example, the individual peptide-subclone may be the product of the quality score of the individual peptide and the probability that the individual peptide belongs to the individual subclone. The cell frequency or cell fraction may be reorganized into a hierarchy that indicates whether a mutation occurs in the presence of other mutations or whether the mutation occurs in isolation.

[0039] Each peptide may have an assigned weight, and selection of peptides is constrained by a maximum total weight. In some embodiments, each peptide is assigned the same weight, such as the value "1". In these cases, the maximum total weight constraint may also be expressed as the maximum number of peptides that may be selected for inclusion in the immunogenic composition or for further analysis.

[0040] In some embodiments, such as those in which the value of the objective function is the sum of the subclone scores across all subclones, the maximum value of the objective function is equal to the number of subclones in the list of subclones, indicating that every subclone is covered by the selected peptide. The value represents the possible number of subclones that may have the peptide presented, bound, immunogenic, etc. In other embodiments, such as those in which the value of the objective function is the product of the subclone scores across all subclones, the maximum value of the objective function is 1 and the minimum value is 0. This can be interpreted as the probability that every subclone has at least one peptide presented, bound, immunogenic, etc.

[0041] In some embodiments, the techniques of the present invention can also be applied to any type of epitope and are not limited to peptides, for example, this can include RNA and DNA equivalents, mRNA, and concatemers.

[0042] Objective function The above objective function represents the problem of selecting a set of peptides that strikes an optimal balance between subclone coverage and peptide validity (represented by a quality score). In some embodiments, the objective function can be expressed as:

number

number

[0043] This problem is motivated by assuming that the quality scores are probability estimates for each peptide, s i =P(x i ), and the optimization task is to choose a restricted set of peptides from the list of peptides that maximizes the number of subclones that are likely to contain one or more peptides with high quality scores, subject to constraints. However, the overall value of the objective function is not just a probability value or a sum of quality scores. Instead, the quadratic term -AB in the definition of AVB results in diminishing returns to value for adding additional peptides that belong to subclones already covered by other selected peptides.

[0044] In some embodiments, all of the peptides have the same weight, represented by "1". In such cases, the constraint is the maximum number of peptides that can be selected. In exemplary embodiments, the maximum number of peptides that can be selected can be 18, 19, or 20. In some embodiments, the maximum number of peptides can be 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 peptides. The maximum number of peptides that can be selected can be about 2-20 peptides, about 2-30 peptides, about 2-40 peptides, about 2-50 peptides, about 2-60 peptides, about 2-70 peptides, about 2-80 peptides, about 2-90 peptides, or about 2-100 peptides. The objective function can be solved using a variety of methods, some of which are described below.

[0045] The objective function as a Lagrange multiplier problem The refutation function in Equation 1 and the constraints in Equation 2 can be expressed as the following Lagrangian multiplier problem, where the solution to this Lagrangian multiplier problem represents the set of selected peptides:

number

[0046] Appropriate parameterization Π θ Given, the problem can be recomputed as follows:

number

[0047] Various parameterization techniques can be used to find a solution to the Lagrangian multiplier problem. In some embodiments, the set of peptides is selected based on a parameterization of the Lagrangian multiplier problem using a logistic technique. Using this technique, for each peptide x in the list of peptides, i For each real parameter θ i Then, the function Π θ i =Φ(θ i ) is evaluated, where Φ(y)=1 / 1+exp(-y) is the sigmoid function.

[0048] In some embodiments, the set of peptides is selected based on a parameterization of the Lagrangian multiplier problem using attention techniques. For each peptide in the list of peptides, the peptide quality score, weight and subclone membership probability are concatenated. This is then processed for each peptide logit using a single encoder layer for a Transformer with parameters θ. The logits are then scaled to Π via a sigmoid function Φ. θ i is converted to

[0049] In some embodiments, the set of peptides is selected based on a parameterization of the Lagrangian multiplier problem using deep set techniques. In some embodiments, the set of peptides is selected using an evolutionary algorithm as an optimization procedure for solving the Lagrangian multiplier problem. In some embodiments, the set of peptides is selected based on gradient descent techniques, such as stochastic techniques, step size techniques, among others. In some embodiments, the set of peptides is selected using combinatorial optimization techniques, which can directly optimize the objective function without expressing it as a Lagrangian multiplier problem.

[0050] Greedy Cluster Allocation In addition to the exemplary Lagrangian-based optimization technique described above, the set of peptides may be selected using other suitable techniques. For example, a greedy cluster allocation technique may be used. In this example, the initial list of peptides is sorted by quality score such that the peptides are ordered by descending quality score. (In some cases, the peptides may be sorted by ascending score, where a lower score indicates a better peptide.) Starting with an empty set (i.e., no peptides have been selected yet), the peptides in the sorted list are then iterated in order, and if the peptide belongs to a subclone to which no other peptide in the set of selected peptides belongs, this peptide is added to the set of selected peptides. Otherwise, the peptide is not selected. The list of sorted peptides is iterated one or more times, and peptides are selected in this manner until one or more conditions are met. In some embodiments, the process stops when the number of selected peptides reaches a maximum number of peptides.

[0051] Direct Sorting In some embodiments, the set of peptides can be selected using a direct sorting technique, which includes: for each peptide in the list of peptides, for each subclone in the list of subclones, obtaining a membership probability between the individual peptide and the individual subclone; determining an average membership probability of the individual peptides across all of the subclones in the list of subclones; and determining a peptide sorting score for each peptide. The peptide sorting score is the product of the average membership probability of the individual peptides and the quality score of the individual peptide. The list of peptides is then sorted by descending peptide sorting score. Finally, the maximum number of top peptides is selected from the list of sorted peptides.

[0052] Additional selection analysis In some embodiments, the peptides may be subjected to manufacturability analysis and filtered for manufacturability. One or more additional inclusion criteria may be applied in addition to or in conjunction with the selection method presented herein. This may be done before the selection method disclosed herein, as part of the method disclosed herein, or after the method disclosed herein. For example, additional filtering / selection criteria may include: 1) RNA abundance (measured in transcripts per million, TPM) for the gene to which the somatic mutation belongs (e.g., the threshold can be set at a minimum of about 1, about 35 or about 100 TPM); 2) whether the somatic mutation is an essential or driver gene. (i.e., a driver gene is a gene whose mutation causes tumor growth and an essential gene is a gene that is essential for the survival of the organism); 3) whether the peptide is predicted to pass quality control thresholds for synthesizability and solubility; 4) how heterologous (i.e., different) is the mutated peptide from the corresponding germline peptide. (e.g., a minimum number of mutated amino acids may be required for a peptide to be considered or included), 5) the confidence level that a particular peptide candidate is present in a particular subject (e.g., rare somatic mutations are given a lower confidence score than more frequently occurring mutations), or 6) whether a peptide candidate contains a certain amino acid, such as cysteine.

[0053] Sequencing methods Various sequencing methods are well known in the art, including but not limited to PCR-based methods, including real-time PC, 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 method described in, for example, Sambrook et al., Molecular Cloning: A Laboratory Manual 4th ed. (2012) Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.See also Austell et al., Current Protocols in Molecular Biology, ed., Greene Publishing and Wiley-Interscience New York (1992) (regularly updated).

[0054] Sequencing methods can also include, but are not limited to, high-throughput sequencing, single cell RNA sequence, 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 array (Solexa), shotgun sequencing, Maxam-Hilbery or Sanger sequencing, whole genome sequencing, whole exome sequencing, primer walking, sequencing using PacBio, SOLid, Ion Torrent or nanopore platform, and any other sequencing method known in the art.The sequencing method employed herein to obtain sequence data is preferably high-throughput sequencing. High-throughput sequencing techniques are capable of sequencing multiple nucleic acid molecules in parallel, allowing millions of nucleic acid molecules to be sequenced at once. See Churko et al., Circ. Res. 112(12):1613-1623 (2013).

[0055] In some cases, the high throughput sequencing can be 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, California)). Any of these platforms can be employed to sequence the genetic material disclosed herein. Next generation sequencing is based on the sequencing of many 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. These steps are then repeated sequentially to identify the incorporated bases and determine the linear order of the nucleotides in the template using a signal from the detectable label. Exemplary detectable labels include radioactive labels, fluorescent labels, enzymatic labels, and the like. A number of techniques are known for detecting sequences, such as the Illumina NextSeq platform by cycle end sequencing.

[0056] Machine learning models Once sequence data representing the polypeptide sequences of one or more tumor-specific neoantigens is obtained, the sequence data, along with the MHC molecules of the subject, are input into a machine learning platform (i.e., model(s)). The machine learning platform generates a numerical probability score that predicts whether the one or more tumor-specific neoantigens are immunogenic (e.g., will elicit an immune response in the subject).

[0057] MHC molecules transport and present peptides on the cell surface. MHC molecules are classified as 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 originating from endogenous proteins or pathogens present inside the cell, and then 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 primarily present peptides from external antigen sources, i.e., outside the cell, to be processed by T helper (Th) cells (i.e., CD4+). 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 may also be expressed on cancer cells.

[0058] MHC class I molecules and / or MHC class II molecules may be input into the machine learning platform. Typically, either MHC class I molecules or MHC class II molecules are input into the machine learning platform. In some embodiments, MHC class I molecules are input into the machine learning platform. In other embodiments, MHC class II molecules are input into the machine learning platform. In some embodiments, the MHC class I machine learning platform may be trained on MHC class I training data. In some embodiments, the MHC class II machine learning platform may be trained on MHC class II training data. In some embodiments, the same machine learning platform may be trained on both MHC class I training data and class II training data. In some embodiments, the machine learning platform may include an MHC class I model and an MHC class II mode.

[0059] MHC class I molecules bind to short peptides. MHC class I molecules can accommodate peptides that are generally about 8 to about 10 amino acids long. In an embodiment, the sequence data encoding one or more tumor-specific neoantigens is a short peptide that is about 8 to about 10 amino acids long. MHC class II molecules bind to long peptides. MHC class II can accommodate peptides that are generally about 13 to about 25 amino acids long. In an embodiment, the sequence data encoding one or more tumor-specific neoantigens is a long peptide that is about 13 to 25 amino acids long.

[0060] The sequence data encoding one or more tumor-specific neoantigens can 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.

[0061] The machine learning platform predicts the likelihood that one or more tumor-specific neoantigens are immunogenic (e.g., will elicit an immune response).

[0062] Immunogenic tumor-specific neoantigens are not expressed in normal tissues. These neoantigens can be presented by antigen-presenting cells to CD4+ and CD8+ T cells to generate 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 assumed to be a T cell-mediated response. The immune response in the subject elicited by one or more tumor-specific neoantigens may involve one or more tumor-specific neoantigens that allow presentation to T cells by antigen-presenting cells such as dendritic cells. Preferably, the one or more tumor-specific neoantigens are capable of activating CD8+ T cells and / or CD4+ T cells.

[0063] In embodiments, the machine learning platform can predict the likelihood that one or more tumor-specific neoantigens will activate CD8+ T cells. In 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 cases, the machine learning platform can predict the frequency of CD8+ activation by one or more tumor-specific neoantigens.

[0064] The machine learning platform may include a model trained with training data. The training data may be obtained from a set of individual subjects. The training data may include data from healthy subjects and subjects with cancer. The training data may include a variety of data that can be used to generate a probability score indicating whether one or more tumor-specific neoantigens will elicit 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 tissue, 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 sequencing data, RNA sequencing data, clinical data from healthy subjects and subjects with cancer, cytokine profiling data, T cell cytotoxicity assay data, peptide-MHC monomer or multimer data, and proteomic data for synthetic proteins, normal and tumor human cell lines, fresh and frozen primary samples, and monoallelic cell lines engineered to express a given MHC allele that is subsequently exposed to a T cell assay.

[0065] 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 sequence-based techniques to generate a numerical probability that one or more tumor-specific neoantigens can elicit an immune response (e.g., induce a high or low antibody response, or a CD8+ response). Sequence-based predictions can include supervised machine learning modules including artificial neural networks (e.g., 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. These platforms require a training dataset that includes known MHC-binding peptides.

[0066] A number of predictive programs are employed to predict whether tumor-specific neoantigens are presented on MHC molecules and capable of eliciting 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 employed as a machine learning platform to generate a numerical probability score indicating whether a neoantigen will elicit an immune response.

[0067] Depending on the machine learning platform employed, additional filters can be applied to prioritize tumor-specific neoantigen candidates, including excluding hypothetical (Riken) proteins, using antigen processing algorithms to exclude epitopes that are unlikely to be constitutively or proteolytically produced by the immune proteasome, and prioritizing neoantigens that have a higher predicted binding affinity than the corresponding wild-type sequence.

[0068] The numerical probability score may be a number between 0 and 1. In embodiments, 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 having a higher numerical probability score compared to a lower numerical probability score indicates that the tumor-specific neoantigen will elicit a higher 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 having a numerical probability score of 1 will be more likely to elicit a greater immune response in a subject than a tumor-specific neoantigen having a numerical probability score of 0.05. Similarly, a tumor-specific neoantigen having a numerical probability score of 0.5 will be more likely to elicit a greater immune response in a subject than a tumor-specific neoantigen having a numerical probability score of 0.1.

[0069] Higher numerical probability scores are preferred compared to lower numerical probability scores. Preferably, tumor-specific neoantigens having 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 indicate a high likelihood that an immune response will be elicited in the subject.

[0070] Although a higher numerical probability score is preferred, a lower numerical probability score may still indicate that the tumor-specific neoantigen is capable of eliciting a sufficient immune response and thus is likely to be a suitable candidate.

[0071] In some instances, 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.

[0072] The method for selecting one or more tumor-specific neoantigens may further comprise a step of computing the affinity of one or more tumor-specific neoantigens for binding 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, less than about 50 nM may indicate that the one or more tumor-specific neoantigens may be suitable for an immunogenic composition. The affinity of one or more tumor-specific neoantigens for binding to an MHC molecule in a subject may predict the immunogenicity of the tumor-specific neoantigen. Alternatively, the median affinity may be an effective method for predicting tumor-specific neoantigen immunogenicity. The median affinity may be calculated using epitope prediction algorithms such as NetMHCpan, ANN, SMM, and SMMPMBEC.

[0073] 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 elicit an immune response in a subject. There are various methods for measuring RNA expression. Known techniques that can measure RNA expression include RNA-seq, and in situ hybridization (e.g., FISH), Northern blot, DNA microarray, tiling array, and quantitative polymerase chain reaction (qPCR). Other known techniques in the art can be used to quantify RNA expression. The RNA can be 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 any other RNA. Preferably, mRNA expression is measured.

[0074] The techniques of the present invention can further reduce the possibility of selecting tumor-specific neoantigens that may induce an autoimmune response in normal tissues. It is envisioned that tumor-specific neoantigens with sequences similar to normal antigens may 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 may 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.

[0075] Finally, one or more tumor-specific neoantigens based on the tumor-specific score are selected for formulation of a subject-specific immunogenic composition. 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 cases, at least about 20 tumor-specific neoantigens are selected.

[0076] 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.

[0077] 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 appeared). Solid tumors may 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, and melanoma tumors. Hematological tumors may include, but are not limited to, tumors 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).

[0078] The methods disclosed herein can be used with any suitable cancerous tumor, including hematological malignancies, solid tumors, sarcomas, carcinomas, and other solid and non-solid tumors. Exemplary 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, carcinoma of unknown primary, cardiac tumor, cervical cancer, chordoma, colon cancer, colorectal cancer, craniopharyngioma, ductal carcinoma, embryonal tumor, endometrial cancer, ependymoma, esophageal cancer, olfactory neuroblastoma, fibrous tissue tumor, and the like. Histocytoma, 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, pancreatic 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 Breast cancer, Merkel cell carcinoma, mesothelioma, metastatic squamous cell carcinoma of the neck of unknown primary site, occult primary cancer, midline nerve bundle carcinoma with NUT gene, oral cancer, multiple endocrine neoplasia syndrome, multiple myeloma, mycosis fungoides, myelodysplastic syndrome, myelodysplastic / myeloproliferative neoplasms, nasal and paranasal sinus cancer, nasopharyngeal carcinoma, neuroblastoma, non-small cell lung cancer, oropharyngeal cancer, osteosarcoma, ovarian cancer, pancreatic cancer, papilloma, paraganglioma, parathyroid cancer, penile cancer, pharyngeal cancer, brown Included are chromocytoma, pituitary tumor, pleuropulmonary blastoma, primary malignant lymphoma of the central nervous system, prostate cancer, rectal cancer, renal cell carcinoma, renal pelvis and ureter carcinoma, 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, teratocarcinoma, testicular cancer, pharyngeal cancer, thymoma and thymic carcinoma, thyroid cancer, urethral cancer, uterine cancer, vaginal cancer, vulvar cancer and Wilms' tumor.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 leukemia, chronic lymphocytic leukemia, T-cell lymphocytic leukemia, bladder cancer or lung cancer.Melanoma is particularly interesting. Breast, lung and bladder cancers are also of particular interest.

[0079] 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. PD-L1 expression in tumor cells is upregulated when attacked by T cells. Thus, tumor vaccines can induce the production of specific T cells and at the same time upregulate the expression of PD-L1, which may limit the efficacy of the immunogenic composition. In addition, while the immune system is activated, the expression of the T cell surface reporter CTLA-4 increases correspondingly, which binds to the ligand B7-1 / B7-2 on antigen-presenting cells and plays an immunosuppressive effect. Thus, in some cases, the subject may further be administered an anti-immunosuppressive or immunostimulatory drug, such as a checkpoint inhibitor. Checkpoint inhibitors may include, but are not limited to, anti-CTL4-A antibodies, anti-PD-1 antibodies, and anti-PD-L1 antibodies. These checkpoint inhibitors bind to immune checkpoint proteins on T cells to remove the inhibition of T cell function by tumor cells. Blockade of CTLA-4 or PD-L1 with antibodies can enhance the patient's immune response against cancerous cells. CTLA-4 has been shown to be effective when following vaccination protocols.

[0080] The immunogenic composition comprising one or more tumor-specific neoantigens can be administered to a subject who has been diagnosed with cancer, already has cancer, has recurrent cancer (i.e., has recurred), or is 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, or radiation). The immunogenic composition comprising one or more tumor-specific neoantigens can be administered to a subject prior to other standard of care cancer therapies (e.g., chemotherapy, immunotherapy, or radiation). The immunogenic composition comprising one or more tumor-specific neoantigens can be administered to a subject simultaneously with, after, or in combination with other standard of care cancer therapies (e.g., chemotherapy, immunotherapy, or radiation).

[0081] The subject may be a human, dog, cat, horse or any animal in which a tumor-specific response is desired.

[0082] The immunogenic composition is administered to the subject in an amount sufficient to elicit an immune response to 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 an enhanced dose of the immunogenic composition to the subject. The immune response to the immunogenic composition can be extended by administering an enhanced dose to the subject. In embodiments, at least one, at least two, at least three or more enhanced doses can be administered to alleviate the cancer. The first enhanced 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 enhanced 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 boosting 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%.

[0083] An amount sufficient to elicit an immune response is defined as a "therapeutically effective dose". Amounts effective for this use will depend, for example, on the composition, the mode of administration, the stage and severity of the disease being treated, the weight and general health of the patient, and the judgment of the physician. It should be noted that immunogenic compositions may generally be employed in severe disease conditions, i.e., life-threatening or potentially life-threatening conditions, particularly when cancer has metastasized. In such cases, taking into account the minimization of foreign matter and the relatively non-toxic nature of neoantigens, it is possible and the treating physician may find it desirable to administer these immunogenic compositions in substantial excess.

[0084] 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, or immunotherapy. Any suitable therapeutic treatment for the particular cancer can be administered. Exemplary chemotherapeutic agents include aldesleukin, altretamine, amifostine, asparaginase, bleomycin, capecitabine, carboplatin, carmustine, cladribine, cisapride, cisplatin, cyclophosphamide, cytarabine, dacarbazine (DTIC), dactinomycin, docetaxel, doxorubicin, dronabinol, epoetin alpha, etoposide, filgrastim, fludarabine, fluorouracil, gemcitabine, granisetron, hydroxyurea, idarubicin, ifosfamicin, cyclophosphamide ... The therapeutic agents include, but are not limited to, cyclosporine, 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, and vinorelbine tartrate. The subject may be administered a small molecule, or a targeted therapy (e.g., a kinase inhibitor). The subject may further be administered an anti-CTLA antibody or an anti-PD-1 antibody or an anti-PD-L1 antibody. Blocking CTLA-4 or PD-L1 with an antibody can enhance the patient's immune response to cancerous cells.

[0085] 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 raise a specific immune response.

[0086] The immunogenic composition can be formulated such that the selection and number of tumor-specific neoantigens are tailored 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.

[0087] 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 contain 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 contains at least about 10 tumor-specific neoantigens. Also, preferably, the immunogenic composition contains at least about 20 tumor-specific neoantigens.

[0088] The immunogenic composition may further comprise a natural or synthetic antigen. The natural or synthetic antigen can increase the immune response. Exemplary natural or synthetic antigens include, but are not limited to, pan-DR epitopes (PADRE) and tetanus toxin antigens.

[0089] The immunogenic composition can 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.

[0090] Tumor-specific neoantigens can also be included in viral vector-based vaccine platforms, such as cowpox, fowlpox, self-replicating alphabim, Maraba virus, adenovirus (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-generation / third-generation lentiviruses and recombinant lentiviruses of any generation designed to target specific cell types or receptors. Depending on the packaging capacity of the viral vector-based vaccine platform, this approach can deliver one or more nucleotide sequences encoding one or more tumor-specific neoantigen peptides. The sequences may be flanked by 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 the 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. Vacillus 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, may also be used.

[0091] The immunogenic compositions can contain components that are personalized depending on the personal needs of a particular subject.

[0092] The immunogenic compositions described herein can further include an adjuvant. An adjuvant is any substance whose admixture to the immunogenic composition increases or otherwise strengthens and / or enhances the immune response to tumor-specific neoantigens, but does not generate an immune response to tumor-specific neoantigens when the substance is administered alone. The adjuvant preferably generates an immune response to neoantigens and does not produce an allergic or other adverse reaction. It is contemplated herein that the immunogenic composition can be administered before, together with, in combination with, or after administration of the immunogenic composition.

[0093] Adjuvants can enhance immune responses by several mechanisms, including, for example, lymphocyte recruitment, stimulation of B cells and / or T cells, and 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, particulate adjuvants, microparticulate adjuvants, mucosal adjuvants, and immune stimulating adjuvants. Exemplary adjuvants include aluminum salts (alum) (such as aluminum hydroxide, aluminum phosphate, and aluminum sulfate), 3De-O-acylated monophosphoryl lipid A (MPL) (see GB2220211), MF59 (Novartis), AS03 (Glaxo SmithKline), AS04 (Glaxo SmithKline), polysorbate 80 (Tween 80, ICL Americas, Inc.), 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), and saponins such as QS21 (Kensil et al. al, Vaccine Design: The Subunit and Adjuvant Approach (eds. Powell & Newman, Plenum Press, NY, 1995), see 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 oil or peanut oil), optionally in combination with an immunostimulant such as monophosphoryl lipid A (see Stoute et al, N. Engl. J. Med. 336, 86-91 (1997)).

[0094] CpG immunostimulatory oligonucleotides have also been reported to enhance the effect of adjuvants in a vaccine setting. Other TLR binding molecules, such as RNA-binding TLR7, TLR8 and / or TLR9, may also be used.

[0095] 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, as well as immunologically active small molecules and antibodies that may act therapeutically and / or as adjuvants, such as cyclophosphamide, sunitmib, bevacizumab, Celebrex (celecoxib), NCX-4016, sildenafil, tadalafil, vardenafil, sorafinib, XL-999, CP-547632, pazopamb, ZD2171, AZD2171, ipilimumab, tremelimumab, and SC58175. In an embodiment, polyICLC is a preferred adjuvant.

[0096] 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 the osmotic pressure and / or buffers, and is prepared in a manner known per se, for example by conventional dispersion and suspension processes. In certain embodiments, such dispersions or suspensions may contain viscosity modifiers. The suspensions or dispersions may be kept at a temperature of approximately 2° C. to 8° C., or preferentially frozen for longer storage and then thawed immediately before use. For injection, the vaccine or immunogenic preparation may be formulated in an aqueous solution, preferably in a physiologically compatible buffer such as Hank's solution, Ringer's solution or physiological saline buffer. The solution may contain formulatory agents such as suspending, stabilizing and / or dispersing agents.

[0097] In certain embodiments, the compositions described herein additionally comprise a preservative, such as the mercury derivative thimerosal. In certain embodiments, the pharmaceutical compositions described herein comprise 0.001% to 0.01% thimerosal. In other embodiments, the pharmaceutical compositions described herein do not comprise a preservative.

[0098] An excipient can be present independent of an adjuvant. The function of an excipient can be, 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. An 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). An excipient can be, but is not limited to, a carrier protein of keyhole limpet hemocyanin, a serum protein such as transferrin, bovine serum albumin, human serum albumin, thyroglobulin or ovalbumin, an immunoglobulin, or a hormone such as insulin or palmitic acid. For human immunization, the carrier is generally a physiologically acceptable carrier that is acceptable to humans and is safe. Alternatively, the carrier can be a dextran, such as sepharose.

[0099] Cytotoxic T cells recognize antigens in the form of peptides bound to MHC molecules, not intact heterologous antigens themselves. The MHC molecules themselves are located on the cell surface of antigen-presenting cells. Thus, when a trimeric complex of peptide antigens, MHC molecules and antigen-presenting cells (APCs) is present, activation of cytotoxic T cells is possible. If not only one or more tumor-specific antigens are used to activate cytotoxic T cells, but additional APCs carrying the respective MHC molecules are added, the immune response may be enhanced. Thus, in some embodiments, the immunogenic composition additionally contains at least one APC.

[0100] The immunogenic composition may include an acceptable carrier (e.g., an aqueous carrier). A variety of aqueous carriers may be used, such as water, buffered water, 0.9% saline, 0.3% glycine, hyaluronic acid, and the like. These compositions may be sterilized by conventional, well-known sterilization techniques or may be sterile filtered. The resulting aqueous solutions may be packaged for use as is or lyophilized, and the lyophilized preparations may be combined with a sterile solution prior to administration. The compositions may contain pharma- ceutically acceptable auxiliary substances necessary to approximate physiological conditions, such as pH adjusting and buffering agents, isotonicity agents, wetting agents, and the like, for example, sodium acetate, sodium lactate, sodium chloride, potassium chloride, calcium chloride, sorbitan monolaurate, triethanolamine oleate, and the like.

[0101] Neoantigens can also be administered via liposomes to target specific cellular tissues, such as lymphoid tissues. Liposomes are also useful for increasing half-life. Liposomes include emulsions, foams, micelles, insoluble monolayers, liquid crystals, phospholipid dispersions, lamellar layers, and the like. In these preparations, the neoantigen to be delivered is incorporated as part of the liposome, either alone or in combination with a molecule that binds to a receptor prevalent among lymphoid cells, such as a monoclonal antibody that binds to the CD45 antigen, or other therapeutic or immunogenic compositions. Thus, liposomes loaded with the desired neoantigen can be directed to the site of lymphoid cells, where the liposome then delivers the selected immunogenic composition. Liposomes can be formed from standard vesicle-forming lipids, which generally include neutral and negatively charged phospholipids, and sterols, such as cholesterol. The choice of lipid is generally guided by considerations of, for example, liposome size, acid lability, and stability of the liposomes in the bloodstream. A variety of methods are available for preparing liposomes, as described, for example, 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.

[0102] To target immune cells, the ligand incorporated into the liposomes can include, for example, an antibody or fragment thereof specific for a cell surface determinant of the desired immune system cell. The liposome suspension can be administered intravenously, locally, topically, etc., in doses that vary depending on, among other things, the mode of administration, the peptide being delivered, and the stage of the disease being treated.

[0103] As an alternative method for 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 poly(lactic-co-glycolic acid) microparticles, which can entrap components of the immunogenic composition as endosomal delivery devices.

[0104] Nucleic acids encoding tumor-specific neoantigens as described herein can also be administered to patients for therapeutic or immunization purposes. Numerous methods are suitably used 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), and in U.S. Pat. Nos. 5,580,859 and 5,589,466. The nucleic acid can also be administered using ballistic delivery, as described, for example, in U.S. Pat. No. 5,204,253. Particles composed solely of DNA can be administered. Alternatively, the DNA can be attached to particles, such as gold particles. Approaches for delivering nucleic acid sequences can include viral vectors, mRNA vectors, and DNA vectors, with or without electroporation. The nucleic acid can also be delivered complexed to cationic compounds, such as cationic lipids.

[0105] The immunogenic compositions provided herein can be administered to a subject by routes including, but not limited to, oral, intradermal, intratumoral, intramuscular, intraperitoneal, intravenous, topical, subcutaneous, transdermal, intranasal, and inhalation, as well as via scarring (e.g., using a bifurcated needle to scratch the top layer of the skin). The immunogenic compositions can be administered to the tumor site to induce a localized immune response against the tumor.

[0106] The dosage of one or more tumor-specific neoantigens may depend on the type of composition, as well as the subject's age, weight, body surface area, individual condition, individual pharmacokinetic data, and mode of administration.

[0107] Also disclosed herein are methods of producing an immunogenic composition comprising one or more tumor-specific neoantigens selected by performing the steps of the methods disclosed herein. Immunogenic compositions as described herein can be produced using methods known in the art. For example, a method of producing a tumor-specific neoantigen or vector disclosed herein (e.g., a vector comprising at least one sequence encoding one or more tumor-specific neoantigens) can include culturing a host cell under conditions suitable for expressing the neoantigen or vector, the host cell comprising 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.

[0108] 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 may be a cDNA.

[0109] IV. Sample The method disclosed herein includes selecting one or more tumor-specific neoantigens from a tumor. The method of selecting one or more tumor-specific neoantigens includes obtaining sequence data from the tumor. Such sequence data can be derived from a tumor sample of a subject. The tumor sample can be obtained from a tumor biopsy specimen.

[0110] The tumor sample can be obtained from a human or non-human subject. Preferentially, the tumor sample is obtained from a human. The tumor sample can be obtained from various biological sources, including cancerous tumors. The tumor can be from the tumor site or circulating tumor cells from blood. Exemplary samples can include, but are not limited to, bodily fluids, tissue biopsy specimens, blood samples, serum plasma, stool, skin samples, and the like. The source of the sample can be a solid tissue sample, such as a tumor tissue biopsy specimen. The tissue biopsy specimen can be, for example, a biopsy specimen from lung, prostate, colon, skin, breast tissue, or lymph node. The sample can also be a sample of bone marrow, including, for example, a bone marrow aspirate and a bone marrow biopsy specimen. The sample can also be a liquid biopsy specimen, such as circulating tumor cells, cell-free circulating tumor DNA, or exosomes. The blood sample can be whole blood, partially purified blood, or a fraction of all or partially purified blood, such as peripheral blood mononuclear cells (PBMCs).

[0111] The tumor sample described herein can be obtained directly from the subject, or derived from the subject, or from a sample obtained from the subject, such as cultured cells derived from a biological fluid or tissue sample. The tumor biopsy specimen can be a fresh sample. The fresh sample can be fixed after removal from the subject using any known fixative (e.g., formalin, Zenker fixative or B-5 fixative). The tumor biopsy specimen can also be an archived sample, such as a frozen sample, a cryopreserved sample, of cells obtained directly from the subject or cells derived from the subject. Preferably, the tumor sample obtained from the subject is a fresh tumor biopsy specimen.

[0112] Tumor samples can be obtained from subjects by any means, including but not limited to tumor biopsy, needle aspiration, scraping, surgical resection, surgical incision, venipuncture or other means known in the art.Tumor biopsy is the preferred method for obtaining tumor.Tumor biopsy can be obtained from any cancerous site, for example, primary tumor or secondary tumor.Tumor biopsy from primary tumor is generally preferred.Those skilled in the art will recognize other suitable techniques for obtaining tumor samples.

[0113] Tumor samples can be obtained from a subject in a single procedure. Tumor samples can be obtained repeatedly from a subject over a period of time. For example, tumor samples can be obtained once a day, once a week, once a month, once every two years, or once a year. Obtaining multiple samples over a period of time can be useful for identifying and selecting new tumor-specific neoantigens. Tumor samples can be obtained from the same tumor or different tumors.

[0114] Tumor samples may be obtained from the primary tumor, one or more metastases, and / or individual sites of tumor growth (e.g., bone marrow from different skeletal sites such as hips, bones, or vertebrae). Tumor samples may be obtained from the same site or different sites.

[0115] All or any portion of the above may be implemented on a computing environment such as those illustrated in Figures 1-3. Figure 1 illustrates an example of a 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 that allow customers to purchase, rent, or otherwise obtain instances 912 of virtualized resources, including, but not limited to, computational and storage resources implemented on devices within the provider network or a network within 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 may also offer 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 obtain from the provider 900.

[0116] Conventionally, the provider network 900, via the virtualization service 910, may enable a customer of the service provider (e.g., a customer operating one or more client networks 950A-950C including one or more customer device(s) 952) to dynamically associate at least some public IP addresses 914 assigned or allocated to the customer with a particular resource instance 912 assigned to the customer. The provider network 900 may also enable a customer to remap a public IP address 914 previously mapped to one virtualized computing resource instance 912 allocated to the customer to another virtualized computing resource instance 912 also allocated to the customer. Using the virtualized computing resource instances 912 and public IP addresses 914 provided by the service provider, the customer of the service provider, such as an operator of customer network(s) 950A-950C, may, for example, implement customer-specific applications and present the customer's applications over an intermediate network 940, such as the Internet. Other network entities 920 on the intermediate network 940 can then generate traffic to the destination public IP address 914 exposed by the customer network(s) 950A-950C, which is routed to the service provider data center, where it is routed over the network substrate to the local IP address 916 of the virtualized computing resource instance 912 that is currently mapped to the destination public IP address 914. Similarly, response traffic from the virtualized computing resource instance 912 can be routed back to the intermediate network 940, over the network substrate, to the source entity 920.

[0117] A local IP address as used herein refers to an internal or "private" network address of a resource instance, for example, within a provider network. The local IP address may be within an address block reserved by Internet Engineering Task Force (IETF) Request for Comments (RFC) 1918 and / or within an address block with an address format specified by IETF RFC 4193, and may be changeable within the provider network. Network traffic originating from outside the provider network is not routed directly to the 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 may include a networking 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.

[0118] 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.

[0119] 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, the mapping of standard IP addresses to the local IP addresses of a resource instance is the default launch configuration for all resource instance types.

[0120] At least some public IP addresses may be assigned or acquired by customers of provider network 900, and the customers may then assign those allocated public IP addresses to specific resource instances allocated to the customers. These public IP addresses may be referred to as customer public IP addresses or simply customer IP addresses. Instead of being assigned to resource instances by provider network 900, as is the case with standard IP addresses, customer IP addresses may be assigned to resource instances by customers, for example, via an API provided by the service provider. Unlike standard IP addresses, customer IP addresses are allocated to customer accounts and can be remapped to other resource instances by the respective customers as needed or desired. Customer IP addresses are associated with a customer's account, not a specific resource instance, and the customer controls that IP address until the customer chooses to release it. 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 allows a customer to troubleshoot problems with a customer's resource instance or software, for example, by remapping the customer IP address to an alternative resource instance.

[0121] 2 is a block diagram of an example 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 rented or leased, for example, to customers of the provider network 1000 (e.g., customers implementing customer network 1050). Each computing resource 1024 may be provided with one or more local IP addresses. The provider network 1000 may be 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.

[0122] The provider network 1000 may provide customer networks 1050, coupled 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 coupled 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 may access 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, in the provider network 1000, each virtual computing system 1092 in the customer networks 1050 may correspond to a computational resource 1024 leased, rented, or otherwise provided to the customer network 1050.

[0123] From an instance of the virtual computing system 1092 and / or another customer device 1090 (e.g., via a console 1094), a customer can access functionality of the storage service 1010, e.g., via one or more APIs 1002, to access data from 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) can be provided in the customer network 1050 that can locally cache at least some data, e.g., frequently accessed or important data, and can communicate with the storage service 1010 via one or more communication channels to upload new or modified data from the local cache such that a primary store of data (the virtualized data store 1016) is maintained. In some embodiments, a user can mount and access virtual data store 1016 volumes 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 may appear to the user as local (virtualized) storage 1098.

[0124] 2, the virtualization service(s) may also be accessed from resource instances within the provider network 1000 via API(s) 1002. For example, a customer, appliance service provider, or other entity may access the virtualization service from within a respective virtual network on the provider network 1000 via API 1002 to request allocation of one or more resource instances within the virtual network or another virtual network.

[0125] Exemplary System In some embodiments, a system implementing some or all of the techniques described herein may include a general-purpose computer system that includes or is configured to access one or more computer-accessible media, such as computer system 1100 shown in FIG. 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 coupled to I / O interface 1130. Although FIG. 3 illustrates computer system 1100 as a single computing device, in various embodiments computer system 1100 may include one computing device or any number of computing devices configured to operate together as single computer system 1100.

[0126] In various embodiments, computer system 1100 may be a uniprocessor system including one processor 1110 or a multiprocessor system including several processors 1110 (e.g., two, four, eight, or another suitable number). 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 may typically, but not necessarily, implement the same ISA.

[0127] The system memory 1120 may store instructions and data accessible by the processor(s) 1110. In various embodiments, the system memory 1120 may be implemented using any suitable memory technology, such as random access memory (RAM), static RAM (SRAM), synchronous dynamic RAM (SDRAM), non-volatile / flash type memory, or any other type of memory. In the illustrated embodiment, program instructions and data implementing one or more desired functions, such as those methods, techniques, and data described above, are shown stored within the system memory 1120 as enzyme-substrate predictor service code 1125 and data 1126.

[0128] In one embodiment, the I / O interface 1130 may be configured to coordinate I / O traffic between the processor 1110, the system memory 1120, and any peripheral devices within the device, including the network interface 1140 or other peripheral interfaces. In some embodiments, the I / O interface 1130 may perform any necessary protocol, timing, or other data conversion 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 may include support for devices attached through various types of peripheral buses, such as, for example, variants of 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 may be 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 an interface to the system memory 1120, may be incorporated directly into the processor 1110.

[0129] Network interface 1140 may be configured to allow data to be exchanged between computer system 1100 and other devices 1160 attached to network(s) 1150. In various embodiments, network interface 1140 may support communications over any suitable wired or wireless general data network, such as, for example, a type of Ethernet network. Additionally, network interface 1140 may support communications over a telecommunications / telephone network, such as an analog voice network or a digital fiber communications network, over a storage area network (SAN), such as a Fibre Channel SAN, or over any other suitable type of I / O network and / or protocol.

[0130] 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 can 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 can manage the compute instances executing on the host electronic device. As an example, in some embodiments, offload card(s) 1170 can perform compute instance management operations such as pausing and / or unpausing compute instances, launching and / or terminating compute instances, performing memory transfer / copy operations, etc. These management operations may, in some embodiments, be performed by offload card(s) 1170 in coordination with (e.g., upon request from) a hypervisor executed by other processors 1110A-1110N of computer system 1100. However, in some embodiments, the virtualization manager implemented by offload card(s) 1170 may accommodate requests from other entities (e.g., from the compute instance itself) and may not coordinate with (or service) any separate hypervisor.

[0131] In some embodiments, the system memory 1120 may be one embodiment of a computer-accessible medium configured to store program instructions and data as 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. Generally speaking, the computer-accessible medium may include non-transitory storage or memory media, such as magnetic or optical media, e.g., disks or DVDs / CDs coupled to the computer system 1100 via the I / O interface 1130. The non-transitory computer-accessible storage media may also include any volatile or non-volatile media, such as RAM (e.g., SDRAM, double data rate (DDR) SDRAM, SRAM, etc.), read-only memory (ROM), etc., that may be included in some embodiments of the computer system 1100 as the system memory 1120 or another type of memory. Additionally, the computer-accessible medium may include a transmission medium, or a signal, such as an electrical, electromagnetic, or digital signal, conveyed over a communication medium, such as a network and / or a wireless link, such as may be implemented via the network interface 1140.

[0132] Various embodiments discussed or suggested herein can be implemented in a variety of operating environments, which in some cases may 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 may 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 may also include a number of workstations running any of a variety of commercially available operating systems and other known applications for purposes such as development and database management. These devices may also include other electronic devices, such as dummy terminals, thin clients, gaming systems, and / or other devices capable of communicating over a network.

[0133] 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 any combination thereof.

[0134] 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 be capable of executing 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 any 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.

[0135] The environments disclosed herein may include various data stores and other memory and storage media, as discussed above. These may reside in a variety of locations, such as storage media local to (and / or residing 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 in a storage area network (SAN), which is well known to those skilled in the art. Similarly, any necessary files to perform functions attributed 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, the elements 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 may also include 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, etc.

[0136] Such devices may 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 may 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 may have numerous variations from those described 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. Furthermore, connections to other computing devices, such as network input / output devices, may be used.

[0137] Storage media and computer readable media for containing the code or portions of code may include any suitable medium known or used in the art, including storage media and communication media, such as, but not limited to, volatile and non-volatile, 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.

[0138] In the preceding description, various embodiments are described. For purposes of explanation, specific configurations and details are set forth in order to provide a thorough understanding of the embodiments. However, it will be apparent to one 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.

[0139] Bracketed text and blocks with dashed borders (e.g., thick dashed, thin dashed, dotted and dotted) are used herein to indicate optional operations that add additional features to some embodiments. However, such notations should not be taken to mean that these are the only options or optional operations and / or that blocks with solid borders are not optional in certain embodiments.

[0140] A reference numeral with a suffix letter may be used to indicate that, in various embodiments, there may be one or more instances of the referenced entity, and when there are multiple instances, each need not be identical, but may instead share some common characteristic or act in a common manner. Furthermore, the particular suffix used is not meant to imply that a particular quantity of the entity is present, unless specifically stated otherwise. Thus, in various embodiments, two entities using the same or different prefix letters may or may not have the same number of instances.

[0141] References to "one embodiment," "an embodiment," "an example embodiment," and the like indicate that the embodiment being described may include a particular feature, structure, or characteristic, but not every embodiment may necessarily include 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 affect such feature, structure, or characteristic in connection with other embodiments, whether or not explicitly stated.

[0142] Additionally, in the various embodiments described above, unless specifically stated otherwise, 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, and should not be understood to, imply that a given embodiment requires at least one of A, at least one of B, or at least one of C to be present, respectively.

[0143] 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.

[0144] Equivalent It will be readily apparent to those skilled in the art that other suitable modifications and adaptations of the methods of the present invention described herein will be apparent and may be made using suitable equivalents without departing from the scope of the present disclosure or the present embodiments. Having now described certain compositions and methods in detail, the same will be more clearly understood by reference to the following examples, which are not intended to be limiting and are introduced for illustrative purposes only. EXAMPLES

[0145] The following examples are provided for illustrative purposes only and are not intended to be limiting in any way.

[0146] Example 1: [Table 1] [Table 2] [Table 3] [Table 4]

[0147] Example 1 shows short MHC class I vaccine peptide candidates and predicted variant epitopes for example variants, according to an example embodiment. In this example, the boxed letter "H" represents a variant subsequence of the vaccine peptide sequence "FVLQHLVFL" (SEQ ID NO: 26). According to one or more of the methods described elsewhere herein, the "cluster assignment probability" and "cell occurrence rate" can be determined by an objective function, and the cluster assignment probability or cell occurrence rate can correspond to a subclone score or subclone weight. The subclone score or weight can be based on the probability that at least one of the selected epitopes belongs to an individual subclone. In some embodiments, the subclone score or weight can be utilized to determine the relative ordering of peptides in a list of peptides.

[0148] Example 2: [Table 5] [Table 6] [Table 7] [Table 8]

[0149] Example 2 shows short MHC class I vaccine peptide candidates and predicted variant epitopes for example variants according to an example embodiment. In this example, the boxed letter "C" represents a variant subsequence of the vaccine peptide sequence "KACHYHSYNGW" (SEQ ID NO: 7).

[0150] Example 3: [Table 9] [Table 10] [Table 11] [Table 12]

[0151] Example 3 shows short MHC class I vaccine peptide candidates and predicted variant epitopes for example variants. In this example, the boxed letter "H" represents a variant subsequence of the vaccine peptide sequence "REEENHSFL" (SEQ ID NO:50).

[0152] Example 4: [Table 13] [Table 14] [Table 15] [Table 16-1] [Table 16-2] [Table 16-3]

[0153] Example 5: [Table 17] [Table 18]

[0154] In Examples 4 and 5 above, the short sequence "FVLQHLVFL" (SEQ ID NO: 26) from Example 1 was used to generate sequences for the long MHC class I vaccine peptide in Example 4 and the long MHC class II vaccine peptide in Example 5, both of which contain the same short subsequence (e.g., the letter "H" in a box) in the center of the sequence. As described elsewhere herein, amino acids can be added to both sides of the short subsequence according to the longest neoantigen, such that there is a first maximum number of adjacent amino acids on each side of the mutated amino acid. Predicted mutated epitopes can be generated or determined for both MHC class I and MHC class II vaccine peptides, along with corresponding subclone scores or weights. The subclone scores or weights may be based on the probability that at least one of the selected epitopes belongs to an individual subclone or may be determined by an objective function. In some embodiments, the subclone scores or weights can be utilized to determine the relative ordering of peptides in a list of peptides.

[0155] Example 6: [Table 19] [Table 20] [Table 21] [Table 22-1] [Table 22-2] [Table 22-3]

[0156] Example 7: [Table 23] [Table 24]

[0157] In Examples 6 and 7 shown above, the short sequence "KACHYHSYNGW" (SEQ ID NO: 7) of Example 2 was used to generate sequences for the long MHC class I vaccine peptide of Example 6 and the long MHC class II vaccine peptide of Example 7, both of which contain the same short subsequence (e.g., the letter "C" in a box) in the center of the sequence as in Example 2. Predicted variant epitopes can be generated or determined for both the MHC class I and MHC class II vaccine peptides along with corresponding subclone scores or weights. The subclone scores or weights may be based on the probability that at least one of the selected epitopes belongs to an individual subclone or may be determined by an objective function. In some embodiments, the subclone scores or weights may be utilized to determine the relative ordering of peptides in a list of peptides.

[0158] Example 8: [Table 25] [Table 26] [Table 27] [Table 28-1] [Table 28-2]

[0159] Example 9: [Table 29] [Table 30-1] [Table 30-2]

[0160] In Examples 8 and 9 shown above, the short sequence "REEENHSFL" (SEQ ID NO:50) of Example 3 was used to generate sequences for the long MHC class I vaccine peptide of Example 8 and the long MHC class II vaccine peptide of Example 9, both of which contain the same short subsequence (e.g., the letter "H" in a box) in the center of the sequence as in Example 3. Predicted variant epitopes can be generated or determined for both the MHC class I and MHC class II vaccine peptides, along with corresponding subclone scores or weights. The subclone scores or weights may be based on the probability that at least one of the selected epitopes belongs to an individual subclone, or may be determined by an objective function. In some embodiments, the subclone scores or weights may be utilized to determine the relative ordering of peptides in a list of peptides.

Claims

1. 1. A method for selecting tumor-specific neoantigens for a subject-specific immunogenic composition from a tumor of a subject, comprising: a) providing a list of epitopes determined to be present in said tumor; b) providing a list of the subclones determined to be present in said tumor; and c) providing a mapping of epitopes to subclones, said mapping indicating to which subclone(s) of said list of subclones an epitope of said list of epitopes belongs; d) selecting a set of epitopes from the list of epitopes based on an objective function configured to determine a maximum value calculated by applying a summing or integrating operation of subclone scores over epitopes for each subclone, the subclone score for each subclone being based at least in part on the probability that at least one of the selected epitopes belongs to the individual subclone, a constraint of the objective function limits the number of epitopes in the set of epitopes, and the maximum value is determined by solving the objective function subject to the constraints using an optimization algorithm. said selecting.

2. the constraints of the objective function limit the number of epitopes in the set of epitopes to a predetermined maximum number of epitopes; and / or 2. The method of claim 1, wherein the mapping of epitopes to subclones comprises, for each combination of epitope and subclone, a probability that the epitope belongs to the subclone.

3. 3. The method of claim 1 or 2, wherein the subclone score for the individual subclone is based on an individual epitope-subclone score across the set of selected epitopes, and the individual epitope-subclone score is based on the probability that an individual epitope of the set of selected epitopes belongs to the individual subclone.

4. Each epitope in the list of epitopes has a quality score, preferably The method of any one of claims 1 to 3, wherein the individual epitope-subclone scores are based on the quality scores of the individual epitopes.

5. 5. The method of claim 4, wherein the quality score of the individual epitope ranges from 0 to 1, inclusive, and the individual epitope-subclone score is the product of the quality score of the individual epitope and the probability that the individual epitope belongs to the individual subclone.

6. the quality score of an epitope in the list of epitopes is based on one or more of the presentation probability, binding affinity or immunogenic response of the epitope; and / or 6. The method of claim 4 or 5, wherein the quality score is determined based on an MHC Class I machine learning model, an MHC Class II machine learning model, manufacturability, or one or more inclusion criteria.

7. the constraint determines a maximum total weight for the set of selected epitopes, each epitope in the list being assigned a weight; and / or each epitope in said list of epitopes is weighted equally; and / or each epitope in said list of epitopes is assigned a weight of 1; and / or the maximum value of the objective function is the number of subclones in the list of subclones; and / or The method according to any one of claims 1 to 6, wherein the maximum value of the objective function is 1.

8. The method of any one of claims 1 to 7, further comprising expressing the objective function and the constraints as a Lagrangian multiplier problem, and wherein the set of epitopes is selected based on a solution of the Lagrangian multiplier problem.

9. or further comprising using a logistic technique to select the set of epitopes based on a parameterization of the Lagrangian multiplier problem; or further comprising using attention techniques to select the set of epitopes based on a parameterization of the Lagrangian multiplier problem; or further comprising using deep set techniques to select the set of epitopes based on a parameterization of a Lagrangian multiplier problem; selecting the set of epitopes based on a gradient descent technique. The method of claim 8 , wherein the gradient descent technique preferably comprises a stochastic technique.

10. The method of claim 8 , wherein the gradient descent technique comprises an adaptive step size technique.

11. selecting said set of epitopes based on combinatorial optimization techniques; or The method of any one of claims 1 to 7, further comprising selecting the set of epitopes based on an evolutionary algorithm technique.

12. Sorting the list of epitopes by quality score; and adding said epitope to said set of selected epitopes if said epitope belongs to a subclone to which no other epitope in said set of selected epitopes belongs until a stopping condition is met, said method further comprising: iterating through said sorted list one or more times in descending order to add said epitope to said set of selected epitopes, preferably the stopping condition is met when the number of epitopes in the set of selected epitopes reaches the maximum number of epitopes, or The method according to any one of claims 1 to 7, wherein the stopping condition is met when the number of epitopes in the set of selected epitopes reaches the number of epitopes in the list of epitopes.

13. For each epitope in the list of epitopes: determining, for each subclone in said list of subclones, a membership probability between each epitope and each subclone; determining an average probability of said individual epitopes across all of said subclones in said list of subclones; determining an epitope sorting score for each of the epitopes, the epitope sorting score being the product of the average membership probability of each of the epitopes and the quality score of each of the epitopes; sorting the list of epitopes by descending epitope sorting score; The method of any one of claims 1 to 7, further comprising selecting a maximum number of top epitopes from the list of sorted epitopes.

14. A subclone from the list of subclones is determined to be present in the tumor based on the probability that the subclone is present in the tumor; and / or the maximum number of epitopes is 18, 19 or 20; and / or The method of any one of claims 1 to 13, wherein each epitope in the list of epitopes meets one or more inclusion criteria.

15. and / or forming a subject-specific immunogenic composition comprising said set of selected epitopes. and / or receiving the list of epitopes from a client device. and / or further comprising identifying an epitope present in said tumor. and / or receiving the list of subclones from a client device. and / or further comprising identifying subclones present in said tumor. and / or receiving the mapping from a client device. and / or determining to which subclone(s) of said list of subclones an epitope of said list of epitopes belongs; and / or transmitting the list of selected epitopes to the client device. and / or providing a list of said selected epitopes for the preparation of a subject-specific immunogenic composition. The method of any one of claims 1 to 14, further comprising forming a subject-specific immunogenic composition comprising one or more epitopes of said set of epitopes.

16. The subject-specific immunogenic composition of claim 15 for use as a medicine.

Images