METHOD OF COMPACT PEPTIDE VACCINES USING RESIDUE OPTIMIZATION

Abstract

A system for selecting an immunogenic peptide composition comprising a processor and a memory storing processor-executable instructions that, when executed by the processor, cause the processor to create a first peptide set by selecting a plurality of base peptides, wherein at least one peptide of the plurality of base peptides is associated with a disease, create a second peptide set by adding to the first peptide set a modified peptide, wherein the modified peptide comprises a substitution of at least one residue of a base peptide selected from the plurality of base peptides, and create a third peptide set by selecting a subset of the second peptide set, wherein the selected subset of the second peptide set has a predicted vaccine performance, wherein the predicted vaccine performance has a population coverage above a predetermined threshold, and wherein the subset comprises at least one peptide of the second peptide set.

Description

INCORPORATION BY REFERENCE

All documents cited herein are incorporated herein by reference in their entirety.

SEQUENCE LISTING

The instant application contains a Sequence Listing which has been submitted electronically in XML format and is hereby incorporated by reference in its entirety. The XML copy, created on Sep. 13, 2022, is named 2215269_00124US4_SL.xml and is 1,337,250 bytes in size.

TECHNICAL FIELD

The present invention relates generally to compositions, systems, and methods of peptide vaccines. More particularly, the present invention relates to compositions, systems, and methods of designing peptide vaccines to treat or prevent disease optimized based on predicted population immunogenicity.

BACKGROUND

The goal of a peptide vaccine is to train the immune system to recognize and expand its capacity to engage cells that display target peptides to improve the immune response to cancerous cells or pathogens. A peptide vaccine can also be administered to someone who is already diseased to increase their immune response to a causal cancer, other diseases, or pathogen. Alternatively, a peptide vaccine can be administered to induce the immune system to have therapeutic tolerance to one or more peptides. There exists a need for compositions, systems, and methods of peptide vaccines based on prediction of the target peptides that will be displayed to protect a host from cancer, other disease, or pathogen infection.

SUMMARY OF THE INVENTION

In one aspect, the invention provides for a system for selecting an immunogenic peptide composition comprising a processor, and a memory storing processor-executable instructions that, when executed by the processor, cause the processor to create a first peptide set by selecting a plurality of base peptides, wherein at least one peptide of the plurality of base peptides is associated with a disease, create a second peptide set by adding to the first peptide set a modified peptide, wherein the modified peptide comprises a substitution of at least one residue of a base peptide selected from the plurality of base peptides, and create a third peptide set by selecting a subset of the second peptide set, wherein the selected subset of the second peptide set has a predicted vaccine performance, wherein the predicted vaccine performance has a population coverage above a predetermined threshold, and wherein the subset comprises at least one peptide of the second peptide set.

In some embodiments, the plurality of base peptides of the first peptide set is derived from a target protein, wherein the target protein is a tumor neoantigen or a pathogen proteome. In some embodiments, selecting the plurality of base peptides to create the first peptide set comprises sliding a window of size n across an amino acid sequence encoding the target protein, wherein n is between about 8 amino acids and about 25 amino acids in length, and wherein n is a length of each peptide of the plurality of base peptides of the first peptide set. In some embodiments, a peptide of the plurality of base peptides binds to an HLA class I molecule or an HLA class II molecule. In some embodiments, the substitution of the at least one residue comprises substituting an amino acid at an anchor residue position for a different amino acid at the anchor residue position. In some embodiments, the system further comprises filtering the first peptide set to exclude a peptide with a predicted binding core that contains a target residue in an anchor position. In some embodiments, the second peptide set comprises the first peptide set. In some embodiments, the prediction to be bound by the one or more HLA alleles is computed using a binding affinity of less than about 1000 nM. In some embodiments, the predicted vaccine performance is determined by computing a plurality of peptide-HLA immunogenicities of the third peptide set to at least one HLA allele. In some embodiments, each peptide-HLA immunogenicity of the plurality of peptide-HLA immunogenicities of the third peptide set is based on a predicted binding affinity of less than about 500 nM. In some embodiments, the predicted vaccine performance is based on a population coverage, wherein the population coverage is computed based on a frequency of an HLA haplotype in a human population. In some embodiments, the predicted vaccine performance is based on a population coverage, wherein the population coverage is computed based on a frequency of at least two HLA alleles in a human population. In some embodiments, the plurality of base peptides is present in a single subject. In some embodiments, the predicted vaccine performance is an expected number of peptide-HLA hits. In some embodiments, the disease is cancer, and wherein the cancer is selected from the group consisting of pancreas, colon, rectum, kidney, bronchus, lung, uterus, cervix, bladder, liver, and stomach. In some embodiments, the plurality of base peptides of the first peptide set comprises at least one self-peptide.

In another aspect, the invention provides for a non-transitory computer-readable storage medium comprising computer-readable instructions for determining an immunogenic peptide composition that, when executed by a processor cause the processor to create a first peptide set by selecting a plurality of base peptides, wherein a first base peptide and a second base peptide of the plurality of base peptides are each scored for binding by two or more HLA alleles, wherein the first base peptide and the second base peptide are each predicted to be bound by one or more HLA alleles, and wherein the first base peptide and the second base peptide are associated with a disease, create a second peptide set comprising the first base peptide, the second base peptide, a first modified peptide, and a second modified peptide, wherein the first modified peptide comprises a substitution of at least one residue of the first base peptide, and wherein the second modified peptide comprises a substitution of at least one residue of the second base peptide, and create a third peptide set by selecting a subset of the second peptide set, wherein the selected subset of the second peptide set has a predicted vaccine performance, and wherein the predicted vaccine performance is a function of a peptide-HLA immunogenicity of at least one peptide of the third peptide set with respect to the two or more HLA alleles.

In some embodiments, the plurality of base peptides of the first peptide set is derived from a target protein, wherein the target protein is a tumor neoantigen or a pathogen proteome. In some embodiments, selecting the plurality of base peptides to create the first peptide set comprises sliding a window of size n across an amino acid sequence encoding the target protein, wherein n is between about 8 amino acids and about 25 amino acids in length, and wherein n is a length of each peptide of the plurality of base peptides of the first peptide set. In some embodiments, a peptide of the plurality of base peptides binds to an HLA class I molecule or an HLA class II molecule. In some embodiments, the substitution of the at least one residue comprises substituting an amino acid at an anchor residue position for a different amino acid at the anchor residue position. In some embodiments, the non-transitory computer-readable storage medium of further comprises filtering the first peptide set to exclude a peptide with a predicted binding core that contains a target residue in an anchor position. In some embodiments, the second peptide set comprises the first peptide set. In some embodiments, the prediction to be bound by the two or more HLA alleles is computed using a binding affinity of less than about 1000 nM. In some embodiments, the plurality of base peptides of the first peptide set comprises at least one self-peptide.

In another aspect, the invention provides for a system for selecting an immunogenic peptide composition comprising a processor, and a memory storing processor-executable instructions that, when executed by the processor, cause the processor to create a first peptide set by selecting a plurality of base peptides, wherein a first base peptide of the plurality of base peptides is scored for binding by three or more HLA alleles, wherein the first base peptide is predicted to be bound by one or more HLA alleles, and wherein the first base peptide is associated with a disease, create a second peptide set comprising the first base peptide and a modified peptide, wherein the modified peptide comprises a substitution of at least one residue of the first base peptide, and create a third peptide set by selecting a subset of the second peptide set, wherein the selected subset of the second peptide set has a predicted vaccine performance, and wherein the predicted vaccine performance is a function of a peptide-HLA immunogenicity of at least one peptide of the third peptide set with respect to the three or more HLA alleles.

In some embodiments, the first base peptide is scored for binding based on data obtained from experimental assays. In some embodiments, the predicted vaccine performance includes a peptide-HLA immunogenicity of the modified peptide bound to the first HLA allele of the one or more HLA alleles if the first base peptide is predicted to be bound to the first HLA allele of the one or more HLA alleles with a first binding core, wherein the first binding core is a binding core of the first base peptide, wherein the first binding core is identical to a second binding core, and wherein the second binding core is a binding core of the modified peptide bound to the first HLA allele.

In another aspect, the invention provides for a non-transitory computer-readable storage medium comprising computer-readable instructions for determining an immunogenic peptide composition that, when executed by a processor cause the processor to create a first peptide set by selecting a plurality of base peptides, wherein at least one peptide of the plurality of base peptides is associated with a disease, create a second peptide set comprising a first base peptide selected from the first base peptide set and a modified peptide, wherein the modified peptide comprises a substitution of at least one residue of the first base peptide, and create a third peptide set by selecting a subset of the second peptide set, wherein the selected subset of the second peptide set has a predicted vaccine performance, wherein the predicted vaccine performance has an expected number of peptide-HLA hits above a predetermined threshold, and wherein the subset comprises at least one peptide of the second peptide set.

In some embodiments, the first base peptide binds to an HLA class I molecule or an HLA class II molecule.

In another aspect, the invention provides for a system for selecting an immunogenic peptide composition comprising a processor, and a memory storing processor-executable instructions that, when executed by the processor, cause the processor to create a first peptide set by selecting a first plurality of peptides, wherein the first plurality of peptides comprises a plurality of target peptides that are associated with a first disease, and wherein the first peptide set has a first predicted vaccine performance value, create a second peptide set by selecting a second plurality of peptides, wherein the second plurality of peptides comprises a plurality of target peptides that are associated with a second disease, and wherein the second peptide set has a second predicted vaccine performance value, create a first weighted peptide set by multiplying a first weight by the first predicted vaccine performance value, create a second weighted peptide set multiplying a second weight by the second predicted vaccine performance value, and create a third peptide set by combining the first weighted peptide set and the second weighted peptide set.

In some embodiments, the first predicted vaccine performance value and the second predicted vaccine performance value are computed based on a population coverage of a vaccine. In some embodiments, the first predicted vaccine performance value and the second predicted vaccine performance value are computed based on an expected number of peptide-HLA hits. In some embodiments, the first plurality of peptides is derived from a tumor neoantigen or a pathogen proteome. In some embodiments, the second plurality of peptides is derived from a tumor neoantigen or a pathogen proteome. In some embodiments, the first disease is cancer, and wherein the cancer is selected from the group consisting of pancreas, colon, rectum, kidney, bronchus, lung, uterus, cervix, bladder, liver, and stomach. In some embodiments, the second disease is cancer, and wherein the cancer is selected from the group consisting of pancreas, colon, rectum, kidney, bronchus, lung, uterus, cervix, bladder, liver, and stomach. In some embodiments, the first plurality of peptides comprises a peptide that binds to an HLA class I molecule or an HLA class II molecule. In some embodiments, the second plurality of peptides comprises a peptide that binds to an HLA class I molecule or an HLA class II molecule.

BRIEF DESCRIPTION OF THE DRAWINGS

The following figures depict illustrative embodiments of the invention.

FIG. 1 is a flow chart of a vaccine optimization method.

FIG. 2 is a flow chart of vaccine optimization method with seed set compression.

FIG. 3 shows predicted population coverage for single target MHC class I vaccines by vaccine size for KRAS G12D, KRAS G12V, KRAS G12R, KRAS G12C, and KRAS G13D targets.

FIG. 4 shows predicted population coverage for single target MHC class II vaccines by vaccine size for KRAS G12D, KRAS G12V, KRAS G12R, KRAS G12C, and KRAS G13D targets.

FIG. 5 shows probabilities of disease presentations for pancreas, colon/rectum, and bronchus/lung and respective probabilities of target presentations for KRAS G12D, KRAS G12V, and KRAS G12R targets.

FIG. 6 is a flow chart for multiple target (combined) vaccine optimization methods.

FIG. 7 shows predicted population coverage for pancreatic cancer multiple target (combined) MHC class I vaccines by vaccine size for KRAS G12D, KRAS G12V, and KRAS G12R targets.

FIG. 8 shows predicted population coverage for pancreatic cancer multiple target (combined) MHC class II vaccines by vaccine size for KRAS G12D, KRAS G12V, and KRAS G12R targets.

FIG. 9 shows an example Python implementation of the MERGEMULTI function for combined vaccine design procedures.

FIG. 10 shows predicated peptide-HLA hits by vaccine size for a KRAS G12V vaccine for the HLA diplotype HLA-A02:03, HLA-A11:01, HLA-B55:02, HLA-B58:01, HLA-C03:02, HLA-C03:03.

DETAILED DESCRIPTION

In some embodiments, the disclosure provides for peptide vaccines that incorporate peptide sequences that will be displayed by Major Histocompatibility Complex (MHC) molecules on cells and train the immune system to recognize cancer or pathogen diseased cells. In some embodiments, the disclosure provides for peptide vaccines that that incorporate peptide sequences that will be displayed by Major Histocompatibility Complex (MHC) molecules on cells to induce therapeutic tolerance in antigen-specific immunotherapy for autoimmune diseases (Alhadj Ali et al., 2017, Gibson, et al. 2015). In some embodiments, a peptide vaccine is a composition that consists of one or more peptides. In some embodiments, a peptide vaccine is an mRNA or DNA construct administered for expression in vivo that encodes for one or more peptides.

Peptide display by an MHC molecule is necessary, but not sufficient, for a peptide to be immunogenic and cause the recognition of the resulting peptide-MHC complex by an individual's T cells to trigger T cell activation, expansion, and immune memory. In some embodiments, experimental data from assays such as the ELISPOT (Slota et al., 2011) or the Multiplex Identification of Antigen-Specific T Cell Receptors Using a Combination of Immune Assays and Immune Receptor Sequencing (MIRA) assay (Klinger et al., 2015) is used for scoring peptide display (e.g., binding affinity) by an MHC molecule (e.g., HLA allele). In some embodiments, experimental data from assays such as the ELISPOT (Slota et al., 2011) or the Multiplex Identification of Antigen-Specific T Cell Receptors Using a Combination of Immune Assays and Immune Receptor Sequencing (MIRA) assay (Klinger et al., 2015) can be combined with machine learning based predictions for scoring peptide display (e.g., binding affinity) by an MHC molecule (e.g., HLA allele). In some embodiments, the MHCflurry or NetMHCpan (Reynisson et al., 2020) computational methods (as known in the art) are used to predict MHC class I display of a peptide by an HLA allele (see Table 1). In some embodiments, the NetMHCIIpan computational method (Reynisson et al., 2020) is used to predict MHC class II display of a peptide by an HLA allele (see Table 2).

A peptide is displayed by an MHC molecule when it binds within the groove of the MEW molecule and is transported to the cell surface where it can be recognized by a T cell receptor. A target peptide refers to a foreign peptide or a self-peptide. In some embodiments, a peptide that is part of the normal proteome in a healthy individual is a self-peptide, and a peptide that is not part of the normal proteome is a foreign peptide. Foreign peptides can be generated by mutations in normal self-proteins in tumor cells that create epitopes called neoantigens, or by pathogenic infections. In some embodiments, a neoantigen is any subsequence of a human protein, where the subsequence contains one or more altered amino acids or protein modifications that do not appear in a healthy individual. Therefore, in this disclosure, foreign peptide refers to an amino acid sequence encoding a fragment of a target protein/peptide (or a full-length protein/peptide), the target protein/peptide consisting of: a neoantigen protein, a pathogen proteome, or any other undesired protein that is non-self and is expected to be bound and displayed by an HLA allele.

For example, KRAS gene mutations are the most frequently mutated oncogenes in cancer, but they have been very difficult to treat with small molecule therapeutics. The KRAS protein is part of a signaling pathway that controls cellular growth, and point mutations in the protein can cause constitutive pathway activation and uncontrolled cell growth. Single amino acid KRAS mutations result in minor changes in protein structure, making it difficult to engineer small molecule drugs that recognize a mutant specific binding pocket and inactivate KRAS signaling. KRAS oncogenic mutations include the mutation of position 12 from glycine to aspartic acid (G12D), glycine to valine (G12V), glycine to arginine (G12R), or glycine to cystine (G12C); or the mutation of position 13 from glycine to aspartic acid (G13D). The corresponding foreign peptides contain these mutations.

A challenge for the design of peptide vaccines is the diversity of human MHC alleles (HLA alleles) that each have specific preferences for the peptide sequences they will display. The Human Leukocyte Antigen (HLA) loci, located within the MHC, encode the HLA class I and class II molecules. There are three classical class I loci (HLA-A, HLA-B, and HLA-C) and three loci that encode class II molecules (HLA-DR, HLA-DQ, and HLA-DP). An individual's HLA type describes the alleles they carry at each of these loci. Peptides of length of between about 8 and about 11 residues can bind to HLA class I (or MHC class I) molecules whereas those of length of between about 13 and about 25 bind to HLA class II (or MHC class II) molecules (Rist et al., 2013; Chicz et al., 1992). Human populations that originate from different geographies have differing frequencies of HLA alleles, and these populations exhibit linkage disequilibrium between HLA loci that result in population specific haplotype frequencies. In some embodiments, methods are disclosed for creating effective vaccines that includes consideration of the HLA allelic frequency in the target population, as well as linkage disequilibrium between HLA genes to achieve a set of peptides that is likely to be robustly displayed.

The present disclosure provides for compositions, systems, and methods of vaccine designs that produce immunity to single or multiple targets. In some embodiments, a target is a neoantigen protein sequence, a pathogen proteome, or any other undesired protein sequence that is non-self and is expected to be bound and displayed by an HLA molecule (also referred to herein as an HLA allele). When a target is present in an individual, it may result in multiple peptide sequences that are displayed by a variety of HLA alleles. In some embodiments, it may be desirable to create a vaccine that includes selected self-peptides, and thus these selected self-peptides are considered to be the target peptides for this purpose.

The term peptide-HLA binding is defined to be the binding of a peptide to an HLA allele, and can either be computationally predicted, experimentally observed, or computationally predicted using experimental observations. The metric of peptide-HLA binding can be expressed as affinity, percentile rank, binary at a predetermined threshold, probability, or other metrics as are known in the art. The term peptide-HLA immunogenicity is defined as the activation of T cells based upon their recognition of a peptide when bound by an HLA allele. Peptide-HLA immunogenicity can vary from individual to individual, and the metric for peptide-HLA immunogenicity can be expressed as a probability, a binary indicator, or other metric that relates to the likelihood that a peptide-HLA combination will be immunogenic. In some embodiments, peptide-HLA immunogenicity is defined as the induction of immune tolerance based upon the recognition of a peptide when bound by an HLA allele. Peptide-HLA immunogenicity can be computationally predicted, experimentally observed, or computationally predicted using experimental observations. In some embodiments, peptide-HLA immunogenicity is based only upon peptide-HLA binding, since peptide-HLA binding is necessary for peptide-HLA immunogenicity. In some embodiments, peptide-HLA immunogenicity data or computational predictions of peptide-HLA immunogenicity can be included and combined with scores for peptide display in the methods disclosed herein. One way of combining the scores is using immunogenicity data for peptides assayed for immunogenicity in diseased or vaccinated individuals, and assigning peptides to the HLA allele that displayed them in the individual by choosing the HLA allele that computational methods predict has the highest likelihood of display. For peptides that are not experimentally assayed, computational predictions of display can be used. In some embodiments, different computational methods of predicting peptide-HLA immunogenicity or peptide-HLA binding can be combined (Liu et al., 2020b). For a given set of peptides and a set of HLA alleles, the term peptide-HLA hits is the number of unique combinations of peptides and HLA alleles that exhibit peptide-HLA immunogenicity or binding at a predetermined threshold. For example, a peptide-HLA hit of 2 can mean that one peptide is predicted to be bound (or trigger T cell activation) by two different HLA alleles, two peptides are predicted to be bound (or trigger T cell activation) by two different HLA alleles, or two peptides are predicted to be bound (or trigger T cell activation) by the same HLA allele. For a given set of peptides and HLA frequencies, HLA haplotype frequencies, or HLA diplotype frequencies, the expected number of peptide-HLA hits is the average number of peptide-HLA hits in each set of HLAs that represent an individual, weighted by their frequency of occurrence.

Since immunogenicity may vary from individual to individual, one method to increase the probability of vaccine efficacy is to use a diverse set of target peptides (e.g., at least two peptides) to increase the chances that some subset of them will be immunogenic in a given individual. Prior research using mouse models has shown that most MHC displayed peptides are immunogenic, but immunogenicity varies from individual to individual as described in Croft et al. (2019). In some embodiments, experimental peptide-HLA immunogenicity data are used to determine which target peptides and their modifications will be effective immunogens in a vaccine.

Considerations for the design of peptide vaccines are outlined in Liu et al., Cell Systems 11, Issue 2, p. 131-146 (Liu et al., 2020) and (Liu et al., 2020b) which are incorporated by reference in their entireties herein.

Certain target peptides may not bind with high affinity to a wide range of HLA molecules. To increase the binding of target peptides to HLA molecules, their amino acid composition can be altered to change one or more anchor residues or other residues. Anchor residues are amino acids that interact with an HLA molecule and have the largest influence on the affinity of a peptide for an HLA molecule. Peptides with altered anchor residues are called heteroclitic peptides. In some embodiments, heteroclitic peptides include target peptides with residue modifications at non-anchor positions. In some embodiments, heteroclitic peptides include target peptides with residue modifications that include unnatural amino acids and amino acid derivatives. Modifications to create heteroclitic peptides can improve the binding of peptides to both MHC class I and MHC class II molecules, and the modifications required can be both peptide and MHC class specific. Since peptide anchor residues face the MHC molecule groove, they are less visible than other peptide residues to T cell receptors. Thus, heteroclitic peptides have been observed to induce a T cell response where the stimulated T cells also respond to unmodified peptides. It has been observed that the use of heteroclitic peptides in a vaccine can improve a vaccine's effectiveness (Zirlik et al., 2006). In some embodiments, the immunogenicity of heteroclitic peptides are experimentally determined and their ability to activate T cells that also recognize the corresponding base (also called seed) peptide of the heteroclitic peptide is determined, as is known in the art. In some embodiments, these assays of the immunogenicity and cross-reactivity of heteroclitic peptides are performed when the heteroclitic peptides are displayed by specific HLA alleles.

Peptide Vaccines to Induce Immunity to One or More Targets

In some embodiments, a method is provided for formulating peptide vaccines using a single vaccine design for one or more targets. In some embodiments, a single target is a foreign protein with a specific mutation (e.g., KRAS G12D). In some embodiments, a single target is a self-protein (e.g., a protein that is overexpressed in tumor cells such as cancer/testis antigens). In some embodiments, multiple targets can be used (e.g. both KRAS G12D and KRAS G13D).

In some embodiments, the method includes extracting peptides to construct a candidate set from all target proteome sequences (e.g., entire KRAS G12D protein) as described in Liu et al. (2020).

FIGS. 1 and 2 depict flow charts for example vaccine design methods that can be used for MHC class I or MHC class II vaccine design. In some embodiments, extracted target peptides are of amino acid length of between about 8 and about 10 (e.g., for MHC class I binding (Rist et al., 2013)). In some embodiments, the extracted target peptides presented by MHC class I molecules are longer than 10 amino acid residues, such as 11 residues (Trolle et al., 2016). In some embodiments, extracted target peptides are of length between about 13 and about 25 (e.g., for class II binding (Chicz et al., 1992)). In some embodiments, sliding windows of various size ranges described herein are used over the entire proteome. In some embodiments, other target peptide lengths for MHC class I and class II sliding windows can be utilized. In some embodiments, computational predictions of proteasomal cleavage are used to filter or select peptides in the candidate set. One computational method for predicting proteasomal cleavage is described by Nielsen et al. (2005). In some embodiments, peptide mutation rates, glycosylation, cleavage sites, or other criteria can be used to filter peptides as described in Liu et al. (2020). In some embodiments, peptides can be filtered based upon evolutionary sequence variation above a predetermined threshold. Evolutionary sequence variation can be computed with respect to other species, other pathogens, other pathogen strains, or other related organisms. In some embodiments, a first peptide set is the candidate set.

As shown in FIGS. 1-2, in some embodiments, the next step of the method includes scoring the target peptides in the candidate set for peptide-HLA binding to all considered HLA alleles as described in Liu et al. (2020) and Liu et al. (2020b). In some embodiments, a first peptide set is the candidate set after scoring the target peptides. Scoring can be accomplished for human HLA molecules, mouse H-2 molecules, swine SLA molecules, or MHC molecules of any species for which prediction algorithms are available or can be developed. Thus, vaccines targeted at non-human species can be designed with the method. Scoring metrics can include the affinity for a target peptide to an HLA allele in nanomolar, eluted ligand, presentation, and other scores that can be expressed as percentile rank or any other metric. The candidate set may be further filtered to exclude peptides whose predicted binding cores do not contain a particular pathogenic or neoantigen target residue of interest or whose predicted binding cores contain the target residue in an anchor position. The candidate set may also be filtered for target peptides of specific lengths, such as length 9 for MHC class I, for example. In some embodiments, scoring of target peptides is accomplished with experimental data or a combination of experimental data and computational prediction methods. When computational models are unavailable to make peptide-HLA binding predictions for particular (peptide, HLA) pairs, the binding value for such pairs can be defined by the mean, median, minimum, or maximum immunogenicity value taken over supported pairs, a fixed value (such as zero), or inferred using other techniques, including a function of the prediction of the most similar (peptide, HLA) pair available in the scoring model.

In some embodiments, a base set (also referred to as seed set herein) is constructed by selecting peptides from the scored candidate set using individual peptide-HLA binding or immunogenicity criteria (e.g., first peptide set) (FIG. 1). The criteria used for scoring peptide-HLA binding during the scoring procedure can accommodate different goals during the base set selection and vaccine design phases. For example, a target peptide with peptide-HLA binding affinities of 500 nM may be displayed by an individual that is diseased, but at a lower frequency than a target peptide with a 50 nM peptide-HLA binding affinity. In some embodiments, during the scoring of a candidate set to qualify peptides for membership in the base set as potential immune system targets, 1000 nM or other less constrained affinity criteria than 50 nM may be utilized. During the combinatorial design phase of a vaccine, a more constrained affinity criteria may be used (e.g., when selecting a third peptide set), such a 50 nM, to increase the probability that a vaccine peptide will be found and displayed by HLA molecules. In some embodiments, peptides are scored for third peptide set potential inclusion that have peptide-HLA binding affinities less than about 500 nM. In some embodiments, peptides are selected for the base set that have peptide-HLA binding affinities less than about 1000 nM. Alternatively, predictions of peptide-HLA immunogenicity can be used to qualify target peptides for base set inclusion. In some embodiments, experimental observations of the immunogenicity of peptides in the context of their display by HLA alleles or experimental observation of the binding of peptides to HLA alleles can be used to score peptides for binding to HLA alleles or peptide-HLA immunogenicity. In some embodiments, computational predictions of the immunogenicity of a peptide in the context of display by HLA alleles can used for scoring such as the methods of Ogishi et al. (2019).

In some embodiments, the method further includes running the OptiVax-Robust algorithm as described in Liu et al. (2020) using the HLA haplotype frequencies of a population on the scored candidate set to construct a base set (also referred to as seed set herein) of target peptides (FIG. 2). In some embodiments, HLA diplotype frequencies can be provided to OptiVax. OptiVax-Robust includes algorithms to eliminate peptide redundancy that arises from the sliding window approach with varying window sizes, but other redundancy elimination measures can be used to enforce minimum edit distance constraints between target peptides in the candidate set. The size of the seed set is determined by a point of diminishing returns of population coverage as a function of the number of target peptides in the seed set. Other criteria can also be used, including a minimum number of vaccine target peptides, maximum number of vaccine target peptides, and desired predicted population coverage. In some embodiments, a predetermined population coverage is less than about 0.4, between about 0.4 and 0.5, between about 0.5 and 0.6, between about 0.6 and 0.7, between about 0.7 and 0.8, between about 0.8 and 0.9, or greater than about 0.9. Another possible criterion is a minimum number of expected peptide-HLA binding hits in each individual. In alternate embodiments, the method further includes running the OptiVax-Unlinked algorithm as described in Liu et al. (2020) instead of OptiVax-Robust.

The OptiVax-Robust method uses binary predictions of peptide-HLA immunogenicity, and these binary predictions can be generated as described in Liu et al. (2020b). The OptiVax-Unlinked method uses the probability of target peptide binding to HLA alleles and can be generated as described in Liu et al. (2020). In some embodiments, OptiVax-Unlinked and EvalVax-Unlinked are used with the probabilities of peptide-HLA immunogenicity. Either method can be used for the purposes described herein, and thus the term “OptiVax” refers to either the Robust or Unlinked method. In some embodiments, the HLA haplotype or HLA allele frequencies of a population provided to OptiVax for vaccine design describe the world's population. In alternative embodiments, the HLA haplotype or HLA allele frequencies of a population provided to OptiVax for vaccine design are specific to a geographic region. In alternative embodiments, the HLA haplotype or HLA allele frequencies of a population provided to OptiVax for vaccine design are specific to an ancestry. In alternative embodiments, the HLA haplotype or HLA allele frequencies of a population provided to OptiVax for vaccine design are specific to a race. In alternative embodiments, the HLA haplotype or HLA allele frequencies of a population provided to OptiVax for vaccine design are specific to individuals with risk factors such as genetic indicators of risk, age, exposure to chemicals, alcohol use, chronic inflammation, diet, hormones, immunosuppression, infectious agents, obesity, radiation, sunlight, or tobacco use. In alternative embodiments, the HLA haplotype or HLA allele frequencies of a population provided to OptiVax for vaccine design are specific to individuals that carry certain HLA alleles. In alternative embodiments, the HLA diplotypes provided to OptiVax for vaccine design describe a single individual, and are used to design an individualized vaccine.

In some embodiments, the base (or seed) set of target peptides (e.g., first peptide set) that results from OptiVax application to the candidate set of target peptides describes a set of unmodified target peptides that represent a possible compact vaccine design (Seed Set in FIG. 2). In some embodiments, the seed set (e.g., first peptide set) is based upon filtering candidate peptides by predicted or observed affinity or immunogenicity with respect to HLA molecules (Seed Set in FIG. 1). However, to improve the display of the target peptides in a wide range of HLA haplotypes as possible, some embodiments include modifications of the seed (or base) set. In some embodiments, experimental assays can be used to ensure that a modified seed (or base) peptide activates T cells that also recognize the base/seed peptide.

For a given target peptide, the optimal anchor residue selection may depend upon the HLA allele that is binding to and displaying the target peptide and the class of the HLA allele (MHC class I or class II). A seed peptide set (e.g., first peptide set) can become an expanded set by including anchor residue modified peptides of either MHC class I or II peptides (FIGS. 1-2). Thus, one aspect of vaccine design is considering how to select a limited set of heteroclitic peptides that derive from the same target peptide for vaccine inclusion given that different heteroclitic peptides will have different and potentially overlapping population coverages.

In some embodiments, all possible anchor modifications for each base set of target peptide are considered. There are typically two anchor residues in peptides bound by MHC class I molecules, typically at positions 2 and 9 for 9-mer peptides. At each anchor position, 20 possible amino acids are attempted in order to select the best heteroclitic peptides. Thus, for MHC class I binding, 400 (i.e., 20 amino acids by 2 positions=20²) minus 1 heteroclitic peptides are generated for each base target peptide. There are typically four anchor residues in peptides bound by MHC class II molecules, typically at positions 1, 4, 6, and 9 of the 9-mer binding core. Thus, for MHC class II binding there are 160,000 (i.e., 20 amino acids by 4 positions=20⁴) minus 1 heteroclitic peptides generated for each base target peptide. Other methods, including Bayesian optimization, can be used to select optimal anchor residues to create heteroclitic peptides from each seed (or base) set peptide. Other methods are presented in “Machine learning optimization of peptides for presentation by class II MHCs” by Dai et al. (2020), incorporated in its entirety herein. In some embodiments, the anchor positions are determined by the HLA allele that presents a peptide, and thus the set of heteroclitic peptides includes for each set of HLA specific anchor positions, all possible anchor modifications.

In some embodiments, for all of the target peptides in the base/seed set, new peptide sequences with all possible anchor residue modifications (e.g., MHC class I or class II) are created resulting in a new heteroclitic base set (Expanded set in FIGS. 1-2) that includes all of the modifications. In some embodiments, for all of the target peptides in the base/seed set, new peptide sequences with anchor residue modifications (e.g., MHC class I or class II) at selected anchor locations are created resulting in a new heteroclitic base set (Expanded set in FIGS. 1-2) that includes the selected modifications. In some embodiments, the anchor residue positions used for modifying peptides are selected from anchor residue positions determined by the HLA alleles considered during vaccine evaluation. In some embodiments, the heteroclitic base set (Expanded set in FIGS. 1-2) also includes the original seed (or base) set (Seed Peptide Set in FIGS. 1-2). In some embodiments, the heteroclitic base set includes amino acid substitutions at non-anchor residues. In some embodiments, modifications of base peptide residues is accomplished to alter binding to T cell receptors to improve therapeutic efficacy (Candia, et al. 2016). In some embodiments, the heteroclitic base set includes amino acid substitutions of non-natural amino acid analogs. The heteroclitic base set is scored for HLA affinity, peptide-HLA immunogenicity, or other metrics as described herein (another round of Peptide Filtering and Scoring as shown in FIGS. 1-2). The scoring predictions may be further updated for pairs of heteroclitic peptide and HLA allele, eliminating pairs where a heteroclitic peptide is predicted to be displayed by an allele but the seed (or base) peptide from which it was derived is not predicted to be displayed by the allele. The scoring predictions may also be filtered to ensure that predicted binding cores of the heteroclitic peptide displayed by a particular HLA allele align exactly in position with the binding cores of the respective seed (or base) set target peptide for that HLA allele. In some embodiments, the scoring predictions are filtered for an HLA allele to ensure that the heteroclitic peptides considered for that HLA allele are only modified at anchor positions determined by that HLA allele. Scoring produces a metric of peptide-HLA immunogenicity for peptides and HLA alleles that can be either binary, a probability of immunogenicity, or other metric of immunogenicity such as peptide-HLA affinity or percent rank, and can be based on computational predictions, experimental observations, or a combination of both computational predictions and experimental observations. In some embodiments, probabilities of peptide-HLA immunogenicity are utilized by OptiVax-Unlinked. In some embodiments, heteroclitic peptides are included in experimental assays such as MIRA (Klinger et al., 2015) to determine their immunogenicity with respect to specific HLA alleles. In some embodiments, the methods of Liu et al. (2020b), can be used to incorporate MIRA data for heteroclitic peptides into a model of peptide-HLA immunogenicity. In some embodiments, the immunogenicity of heteroclitic peptides are experimentally determined and their ability to activate T cells that also recognize the corresponding seed (or base) peptide of the heteroclitic peptide is performed as is known in the art to qualify the heteroclitic peptide for vaccine inclusion. In some embodiments, these assays of the immunogenicity and cross-reactivity of heteroclitic peptides are performed when the heteroclitic peptides are displayed by specific HLA alleles. In some embodiments, computational predictions of the immunogenicity of a heteroclitic peptide in the context of display by HLA alleles can used for scoring such as the methods of Ogishi et al. (2019).

In some embodiments, the next step involves inputting the heteroclitic base set (also referred to as Expanded set as shown in FIGS. 1-2) to OptiVax to select a compact set of vaccine peptides that maximizes predicted vaccine performance (Vaccine Performance Optimization; FIGS. 1-2). In some embodiments, predicted vaccine performance is a function of expected peptide-HLA binding affinity (e.g., a function of the distribution of peptide-HLA binding affinities across all peptide-HLA combinations for a given peptide set, or weighted by the occurrence of the HLA alleles in a population or individual). In some embodiments, predicted vaccine performance is the expected population coverage of a vaccine. In some embodiments, predicted vaccine performance is the expected number peptide-HLA hits produced by a vaccine in a population or individual. In some embodiments, predicted vaccine performance requires a minimum expected number of peptide-HLA hits (e.g., 1, 2, 3, 4, 5, 6, 7, 8, or more) produced by a vaccine. In some embodiments, predicted vaccine performance is a function of population coverage and expected number of peptide-HLA hits desired produced by a vaccine. In some embodiments, predicted vaccine performance is a metric that describes the overall immunogenic properties of a vaccine where all of the peptides in the vaccine are scored for peptide-HLA immunogenicity for two or more HLA alleles (e.g., three or more HLA alleles). In some embodiments, predicted vaccine performance excludes immunogenicity contributions by selected HLA alleles above a maximum number of peptide-HLA hits (e.g., 1, 2, 3, 4, 5, 6, 7, 8, or more). In some embodiments, predicted vaccine performance excludes immunogenicity contributions of individual HLA diplotypes above a maximum number of peptide-HLA hits (e.g., 1, 2, 3, 4, 5, 6, 7, 8, or more). In some embodiments, predicted vaccine performance is the fraction of covered HLA alleles, which is the expected fraction of HLA alleles in each individual that have a minimum number of peptides (e.g., 1, 2, 3, 4, 5, 6, 7, 8, or more) with predicted peptide-HLA immunogenicity produced by a vaccine. In some embodiments, predicted vaccine performance is the expected fraction of HLA alleles in a single individual that have a minimum number of peptides (e.g., 1, 2, 3, 4, 5, 6, 7, 8, or more) with predicted peptide-HLA immunogenicity produced by a vaccine.

Predicted vaccine performance refers to a metric. Predicted vaccine performance can be expressed as a single numerical value, a plurality of numerical values, any number of non-numerical values, and a combination thereof. The value or values can be expressed in any mathematical or symbolic term and on any scale (e.g., nominal scale, ordinal scale, interval scale, or ratio scale).

A seed (or base) peptide and all of the modified peptides that are derived from that seed (or base) peptide comprise a single peptide family. In some embodiments, in the component of vaccine performance that is based on peptide-HLA immunogenicity for a given HLA allele, a maximum number of peptides (e.g., 1, 2, 3, 4, 5, 6, 7, 8, or more) that are in the same peptide family are given computational immunogenicity credit for that HLA allele. This limit on peptide family immunogenicity limits the credit caused by many modified versions of the same base peptide. In some embodiments, the methods described herein are included for running OptiVax with an EvalVax objective function that corresponds to a desired metric of predicted vaccine performance. In some embodiments, population coverage means the proportion of a subject population that presents one or more immunogenic peptides that activate T cells responsive to a seed (or base) target peptide. The metric of population coverage is computed using the HLA haplotype frequency in a given population such as a representative human population. In some embodiments, the metric of population coverage is computed using marginal HLA frequencies in a population. Maximizing population coverage means selecting a peptide set (either a base peptide set, a modified peptide set, or a combination of base and modified peptides; e.g., a first peptide set, second peptide set, or third peptide set) that collectively results in the greatest fraction of the population that has at least a minimum number (e.g., 1, 2, 3, 4, 5, 6, 7, 8, or more) of immunogenic peptide-HLA bindings based on proportions of HLA haplotypes in a given population (e.g., representative human population). In some embodiments, this process includes the OptiVax selection of heteroclitic peptides (as described in this disclosure) that activate T cells that respond to their corresponding seed (or base) peptide and the heteroclitic base peptides to improve population coverage. In some embodiments, the seed (or base) target peptides are always included in the final vaccine design. In some embodiments, peptides are only considered as candidates for a vaccine design (e.g., included in a first, second, and/or third peptide set) if they have been observed to be immunogenic in clinical data, animal models, or tissue culture models.

Although heteroclitic peptides are used as exemplary embodiments in this disclosure, any modified peptide could be used in place of a heteroclitic peptide. A modified peptide is a peptide that has one or more amino acid substitutions of a target base/seed peptide. The amino acid substitution could be located at an anchor position or any other non-anchor position.

In some embodiments, a candidate vaccine peptide (e.g., a base peptide or a modified peptide) is eliminated from vaccine inclusion if it activates T cells that recognize self-peptides (e.g., this can be achieved at the first and/or second round of Peptide Filtering and Sorting as shown in FIGS. 1-2). In some embodiments, a candidate vaccine peptide (e.g., a base peptide or a modified peptide) is computationally eliminated from vaccine inclusion if its outward facing amino acids when bound by an HLA allele are similar to outward facing self-peptide residues that are presented by the same HLA allele, where similarity can be defined by identity or defined similarity metrics such as BLOSUM matrices (BLOSUM matrices are known in the art). Testing a vaccine peptide for its ability to activate T cells that recognize self-peptides can be experimentally accomplished by the vaccination of animal models followed by ELISPOT or other immunogenicity assay or with human tissue protocols. In both cases, models with HLA alleles that present the vaccine peptide are used. In some embodiments, human primary blood mononuclear cells (PBMCs) are stimulated with a vaccine peptide, the T cells are allowed to grow, and then T cell activation with a self-peptide is assayed as described in Tapia-Calle et al. (2019) or other methods as known in the art. In some embodiments, the vaccine peptide is excluded from vaccine inclusion if the T cells are activated by the self-peptide. In some embodiments, computational predictions of the ability of a peptide to activate T cells that also recognize self-peptides can be utilized. These predictions can be based upon the modeling of the outward facing residues from the peptide-HLA complex and their interactions with other peptide residues. In some embodiments, a candidate vaccine peptide (e.g., a base peptide or a modified peptide) is eliminated from vaccine inclusion or experimentally tested for cross-reactivity if it is predicted to activate T cells that also recognize self-peptides based upon the structural similarity of the peptide-MHC complex of the candidate peptide (e.g., a base peptide or a modified peptide) and the peptide-MHC complex of a self-peptide. One method for the prediction of peptide-MHC structure is described by Park et al. (2013).

In some embodiments, a candidate heteroclitic vaccine peptide (e.g., a modified peptide) is eliminated from vaccine inclusion if it does not activate T cells that recognize its corresponding base/seed target peptide (second round of Peptide Filtering and Scoring, FIGS. 1-2). Testing a candidate heteroclitic peptide (e.g., a modified peptide) for its ability to activate T cells that recognize its corresponding seed (or base) target peptide with respect to the same HLA allele can be experimentally accomplished by the vaccination of animal models followed by ELISPOT or other immunogenicity assay or with human tissue protocols. In both cases, models with HLA alleles that present the heteroclitic peptide are used. In some embodiments, human PBMCs are stimulated with the heteroclitic peptide, the T cells are allowed to grow, and then T cell activation with the seed (or base) target peptide is assayed as described in Tapia-Calle et al. (2019) or using other methods known in the art. In some embodiments, computational predictions of the ability of a heteroclitic peptide to activate T cells that also recognize the corresponding seed (or base) target peptide can be utilized. These predictions can be based upon the modeling of the outward facing residues from the peptide-HLA complex and their interactions with other peptide residues. In some embodiments, the structural similarity of the peptide-HLA complex of a heteroclitic peptide and the peptide-HLA complex of the corresponding seed (or base) target is used to qualify heteroclitic peptides for vaccine inclusion or to require experimental immunogenicity testing before vaccine inclusion.

FIG. 3 (MHC class I) and FIG. 4 (MHC class II) show the predicted population coverage of OptiVax-Robust selected single target-specific vaccines with differing number of peptides designed for the KRAS mutations G12D, G12V, G12R, G12C, and G13D. FIGS. 4-5 show that as the number of peptides increases for a vaccine, its predicted population coverage increases. The population coverage shown in FIGS. 4-5 are of those individuals that have the specific mutation that the vaccine is designed to cover. An increase in peptide count will also typically cause the average number of peptide-HLA hits in each individual to increase within the population.

OptiVax can be used to design a vaccine to maximize the fraction/proportion of the population whose HLA molecules are predicted to bind to and display at least p peptides from the vaccine. In some embodiments, this prediction (e.g., scoring) includes experimental immunogenicity data to directly predict at least p peptides will be immunogenic. The number p is input to OptiVax, and OptiVax can be run multiple times with varying values for p to obtain a predicted optimal target peptide set for different peptide counts p. Larger values of p will increase the redundancy of a vaccine at the cost of more peptides to achieve a desired population coverage. In some embodiments, it may not be possible to achieve a given population coverage given a specific heteroclitic base set. In some embodiments, the number p is a function of the desired size of a vaccine.

The methods described herein can be used to design separate vaccine formulations for MHC class I and class II based immunity.

In some embodiments, this procedure is used to create a vaccine for an individual. In some embodiments, the target peptides present in the individual are determined by sequencing the individual's tumor RNA or DNA, and identifying mutations that produce foreign peptides. One embodiment of this method is described in U.S. Pat. No. 10,738,355, incorporated in its entirety herein. In some embodiments, peptide sequencing methods are used to identify target peptides in the individual. One embodiment of this is described in U.S. Publication No.: 2011/0257890. In some embodiments, the target peptides used for the individual's vaccine are selected when a self-peptide, foreign peptide, or RNA encoding a self-peptide or foreign peptide is observed in a specimen from the individual is present at a predetermined level. The target peptides in the individual are used to construct a vaccine as described in the disclosure herein. For vaccine design, OptiVax is provided a diplotype comprising the HLA type of the individual. In an alternative embodiment, the HLA type of an individual is separated into multiple diplotypes with frequencies that sum to one, where each diplotype comprises one or more HLA alleles from the individual and a notation that the other allele positions should not be evaluated. The use of multiple diplotypes will cause OptiVax's objective function to increase the chance that immunogenic peptides will be displayed by all of the constructed diplotypes. This achieves the objective of maximizing the number of distinct HLA alleles in the individual that exhibit peptide-HLA immunogenicity and thus improves the allelic coverage of the vaccine in the individual.

FIG. 10 shows the predicted vaccine performance (predicted number of peptide-HLA hits) of ten example G12V MHC class I vaccines for a single individual with the MHC class I HLA diplotype HLA-A02:03, HLA-A11:01, HLA-B55:02, HLA-B58:01, HLA-C03:02, HLA-C03:03. OptiVax was used to design ten G12V MHC class I vaccines for this HLA diplotype with peptide counts ranging from 1 to 10. For the results in FIG. 10, OptiVax was run with six synthetic diplotypes, each equally weighted, each with one HLA allele from the individual's HLA diplotype, and the other allele positions marked to not be evaluated. The 10 peptide vaccine in FIG. 10 comprises SEQ ID NO: 3 (GAVGVGKSL), SEQ ID NO: 4 (LMVVGAVGV), SEQ ID NO: 7 (VVGAVGVGK), SEQ ID NO: 14 (GPVGVGKSV), SEQ ID NO: 69 (LMVVGAVGI), SEQ ID NO: 72 (LMVVGAVGL), SEQ ID NO: 131 (GAVGVGKSM), SEQ ID NO: 138 (GPVGVGKSA), SEQ ID NO: 142 (VTGAVGVGK), and SEQ ID NO: 198 (VAGAVGVGM). Two peptides, SEQ ID NO: 3 (GAVGVGKSL) and SEQ ID NO: 131 (GAVGVGKSM), are predicted to bind two of the HLA alleles with an affinity of 50 nM or less.

MHC Class I Vaccine Design Procedure

In some embodiments, WIC class I vaccine design procedures consist of the following computations steps.

In some embodiments, the inputs for the computation are:

• • P1 . . . n: Peptide sequence (length n) containing the neoantigen or pathogenic target(s) of interest (e.g., KRAS G12D, KRAS G12V, KRAS G12R, KRAS G12C, KRAS G13D). P_i denotes the amino acid at position i.
  • t: Position of target mutation in P, t∈[1, . . . n] (e.g., t=12 for KRAS G12D).
  • τ₁: Threshold for potential presentation of peptides by MHC for peptide-WIC scoring (e.g., 500 nM binding affinity)
  • τ₂: Threshold for predicted display of peptides by WIC for peptide-MHC scoring (e.g., 50 nM binding affinity)
  • : Set of HLA alleles (for HLA-A, HLA-B, HLA-C loci)
  • F: ³→: Population haplotype frequencies (for OptiVax optimization and coverage evaluation).
  • N: Parameter for EvalVax and OptiVax objective function. Specifies minimum number of predicted per-individual hits for population coverage objective to consider the individual covered. Default=1 (computes P(n≥1) population coverage).

In some embodiments, Peptide-HLA Scoring Functions used are:

• • ScorePotential: P×→: Scoring function mapping a (peptide, HLA allele) pair to a prediction of peptide-HLA display. If predicted affinity ≤τ₁, then returns 1, else returns 0. Options include MHCflurry, NetMHCpan, PUFFIN, ensembles, or alternative metrics or software may be used, including models calibrated against immunogenicity data.
  • SCOREDISPLAY: P×→: Scoring function mapping a (peptide, HLA allele) pair to a prediction of peptide-HLA display. If predicted affinity ≤τ₂, then returns 1, else returns 0. Options include MHCflurry, NetMHCpan, PUFFIN, ensembles, or alternative metrics or software may be used, including models calibrated against immunogenicity data.

Next, from the seed protein sequence (P), a set of windowed native peptides spanning the protein sequence(s) is constructed. In some embodiments, 9-mers are produced, but this process can be performed with any desired window lengths and the resulting peptide sets combined.

={P_{j . . . j+8}|j∈[t−8, . . . ,t],j≠{t−8,t−1}}

The second condition j≠{t−8, t−1} excludes peptides where the mutation at t is in positions P2 or P9 of the windowed 9-mer peptide (i.e., the anchor positions).

Next, each peptide sequence in is scored against all HLA alleles in for potential presentation using SCOREPOTENTIAL (with threshold τ₁=500 nM) and store results in a ||×|| matrix S:

S[p,h]=SCOREPOTENTIAL(p,h)∀p∈,h∈

• • Note that S is a binary matrix where 1 indicates the HLA is predicted to potentially present the peptide, and 0 indicates no potential presentation.Define base set of peptides B⊆:

B={p∈ |∃h s.t. S[p,h]=1}

Thus, B contains the native peptides that are predicted to be potentially presented by at least 1 HLA.

Create a Set of all Heteroclitic Peptides B′ Stemming from Peptides in B:

$B^{\prime }=\bigcup _{b\in B}\mathrm{ANCHOR}-\mathrm{MODIFIED}\left(b\right)$

• • where ANCHOR-MODIFIED(b) returns a set of all 399 anchor-modified peptides stemming from b (with all possible modifications to the amino acids at P2 and P9).

Next, all heteroclitic candidate peptides (e.g., modified peptides) in B′ are scored against all HLA alleles in for predicted display using SCOREDISPLAY (with threshold τ₂=50 nM), and store results in binary |B′|×|| matrix S′₁:

S′ ₁ [b′,h]=S CORE D ISPLAY(b′,h)∀b′∈B′,h∈

Next, an updated scoring matrix S′₂ is computed for heteroclitic peptides conditioned on the potential presentation of the corresponding base peptides by each HLA:

$S_2^{\prime }\left[b^{\prime },h\right]=\{\begin{array}{cc}{S}_1^{\prime }\left[b^{\prime },h\right],& \mathrm{if}S\left[b,h\right]=1\\ 0,& \mathrm{otherwise}\end{array}\forall b^{\prime }\in B^{\prime },h\in \mathscr{H}$

• • where each heteroclitic peptide b′∈B′ is a mutation of base peptide b∈B. This condition enforces that if h was not predicted to potentially present b, then all heteroclitic peptides b′ derived from b will not be displayed by h (even if h would otherwise be predicted to display b′).

In some embodiments, OptiVax-Robust is used to design a final peptide set (e.g., third peptide set) from the union of base peptides and heteroclitic peptides B∪B′ (with corresponding scoring matrices S and S′₂ for B and B′, respectively). Let _k denote the compact set of vaccine peptides output by OptiVax containing k peptides. Note that _k+1 is not necessarily a superset of _k. (In alternate embodiments, OptiVax can be used to augment the base set B with peptides from B′ using scoring matrix S′₂ to return set _k, and the final vaccine set _k+|B| consists of peptides B∪_k.)

In some embodiments, this procedure is repeated independently for each target of interest, and the resulting independent vaccine sets can be merged into a combined vaccine as described below.

MHC Class II Vaccine Design Procedure

In some embodiments, MEW class II vaccine design procedures consist of the following computations steps.

In some embodiments, the inputs for the computation are:

• • P_{1 . . . n}: Peptide sequence(s) (length n) containing the neoantigen(s) of interest (e.g., KRAS G12D, KRAS G12V, KRAS G12R, KRAS G12C, KRAS G13D). P_i denotes the amino acid at position i.
  • t: Position of target mutation in P, t E [1, . . . , n] (e.g., t=12 for KRAS G12D).
  • τ₁: Threshold for potential presentation of peptides by MHC for peptide-MHC scoring (e.g., 500 nM binding affinity)
  • τ₂: Threshold for predicted display of peptides by MHC for peptide-MHC scoring (e.g., 50 nM binding affinity)
  • : Set of HLA alleles (for HLA-DR, HLA-DQ, HLA-DP loci)

F: ³→: Population haplotype frequencies (for OptiVax optimization and coverage evaluation).

• • N: Parameter for EvalVax and OptiVax objective function. Specifies minimum number of predicted per-individual hits for population coverage objective to consider the individual covered. Default=1 (computes P(n≥1) population coverage).

In some embodiments, Peptide-HLA Scoring Functions used are:

• • SCOREPOTENTIAL: P×→: Scoring function mapping a (peptide, HLA allele) pair to a prediction of display. If predicted affinity ≤τ₁, then returns 1, else returns 0. Options include NetMHCIIpan, PUFFIN, ensembles, or alternative metrics or software may be used, including models calibrated against immunogenicity data.
  • SCOREDISPLAY: P×→: Scoring function mapping a (peptide, HLA allele) pair to a prediction of peptide-HLA display. If predicted affinity ≤τ₂, then returns 1, else returns 0. Options include NetMHCIIpan, PUFFIN, ensembles, or alternative metrics or software may be used, including models calibrated against immunogenicity data.
  • FindCore: P×→[1, . . . , n]: Function mapping a (peptide, HLA allele) pair to a prediction of the 9-mer binding core. The core may be specified as the offset position (index) into the peptide where the core begins.

Next, from the seed protein sequence (P), a set of peptides spanning the protein sequence are constructed. Here, we extract all windowed peptides of length 13-25 spanning the target mutation, but this process can be performed using any desired window lengths (e.g., only 15-mers).

$𝒫=\bigcup _{k\in \left[13,\dots ,25\right]}{𝒫}_k$ ${𝒫}_k=\left\{P_{j\dots j+\left(k-1\right)}❘j\in \left[t-\left(k-1\right),\dots ,t\right]\right\}$

• • where _k contains all sliding windows of length k, which are combined to form . Note that here (unlike MHC class I), no peptides are excluded based on binding core or anchor residue positions (for MHC class II, filtering is performed as described in this disclosure).

Next, each peptide sequence in P is scored against all HLA alleles in for potential presentation using SCOREPOTENTIAL (with threshold τ₁=500 nM) and store results in a ||×|| matrix S₁:

S ₁ [p,h]=SCOREPOTENTIAL(p,h)∀p∈,h∈

• • Note that S₁ is a binary matrix where 1 indicates the HLA is predicted to potentially present the peptide, and 0 indicates no potential presentation.

For each (peptide, HLA allele) pair (p, h), identify/predict the 9-mer binding core using FINDCORE. The predicted binding core is recorded in a matrix C:

C[p,h]=FINDCORE(p,h)∀p∈,h∈

Next, an updated scoring matrix S₂ is computed for native peptides in :

$S_2\left[p,h\right]=\{\begin{array}{cc}{S}_1\left[p,h\right],& \mathrm{if}C\left[p,h\right]\mathrm{specifies}{P}_t\mathrm{at}a\mathrm{non}-\mathrm{anchor}\mathrm{position}\mathrm{inside}\mathrm{core}\\ 0,& \mathrm{otherwise}\end{array}$ $\forall p\in 𝒫,h\in \mathscr{H}$

• • where _t is the target residue of interest (e.g., the mutation site of KRAS G12D). This condition enforces the target residue to fall within the binding core at a non-anchor position for all (peptide, HLA allele) pairs with non-zero scores in S₂, and allows the binding core to vary by allele per peptide (as the binding cores of a particular peptide may differ based on the HLA allele presenting the peptide). Thus, for each pair (p, h), if the predicted binding core C[p, h] specifies the target residue _t at an anchor position (P1, P4, P6, or P9 of the 9-mer core), or if _t is not contained within the binding core, then S₂ [p, h]=0. In an alternate embodiment, _t can be located outside of the core or inside the core in a non-anchor position.

Next, OptiVax-Robust is run with peptides and scoring matrix S₂ to identify a non-redundant base set of peptides B⊆. (In alternate embodiments, B can be chosen as the entire set rather than identifying a non-redundant base set.)

Next, a set of all heteroclitic peptides B′ is created stemming from peptides in B:

$B^{\prime }=\bigcup _{b\in \bigcup B}\left\{\mathrm{ANCHOR}-\mathrm{MODIFIED}\left(b,c\right)\forall c❘\exists hs.t.S_2\left[b,h\right]=1\right\}$

• • where ANCHOR−MODIFIED(b,c) returns a set of all 20⁴-1 anchor-modified peptides stemming from b with all possible modifications to the amino acids at P1, P4, P6, and P9 of the 9-mer binding core c. Thus, for each base peptide b, the heteroclitic set B′ contains all anchor-modified peptides b′ with modifications to all unique cores of b identified for any HLA alleles that potentially present b with a valid core position as indicated by scoring matrix S₂.

Next, all heteroclitic candidate peptides (e.g., modified peptides) in B′ are scored against all HLA alleles in for predicted display using SCOREDISPLAY (with threshold τ₂=50 nM), and store results in binary |b′|×|| matrix S′₁:

S′ ₁ [b′,h]=ScoreDisplay(b′,h)∀b′∈,h∈

For each (heteroclitic peptide, HLA allele) pair (b′,h), identify/predict the 9-mer binding core using FINDCORE. The predicted binding core is recorded in a matrix C′:

C′[b′,h]=FINDCORE(b′,h)∀b′∈B′,h∈

An updated scoring matrix Sz is computed for heteroclitic peptides conditioned on the identified binding cores of a heteroclitic and base peptides occurring at the same offset by a particular HLA:

$S_2^{\prime }\left[b^{\prime },h\right]=\{\begin{array}{cc}{S}_1^{\prime }\left[b^{\prime },h\right],& \mathrm{if}{C}^{\prime }\left[b^{\prime },h\right]=C\left[b,h\right]\\ 0,& \mathrm{otherwise}\end{array}\forall b^{\prime }\in B^{\prime },h\in \mathscr{H}$

• • where each heteroclitic peptide b′∈B′ is a mutation of base peptide b∈B. This condition enforces the binding core of the heteroclitic peptide b′ to be at the same relative position as the base peptide b, and, implicitly, enforces that the target residue P_t still falls in a non-anchor position within the 9-mer binding core (Step 3).

An updated scoring matrix S′₃ is computed for heteroclitic peptides conditioned on the potential presentation of the corresponding base peptides by each HLA:

$S_3^{\prime }\left[b^{\prime },h\right]=\{\begin{array}{cc}{S}_2^{\prime }\left[b^{\prime },h\right],& \mathrm{if}S\left[b,h\right]=1\\ 0,& \mathrm{otherwise}\end{array}\forall b^{\prime }\in B^{\prime },h\in \mathscr{H}$

• • where each heteroclitic peptide b′∈B′ is a mutation of base peptide b∈B. This condition enforces that if h was not predicted to display b, then all heteroclitic peptides b′ derived from b will not be displayed by h (even if h would otherwise be predicted to display b′).

OptiVax-Robust is used to design a final peptide set (e.g., third peptide set) from the union of base peptides and heteroclitic peptides B∪B′ (with corresponding scoring matrices S₂ and S′₃ for B and B′, respectively). Let _k denote the compact set of vaccine peptides output by OptiVax containing k peptides. Note that _k+1 is not necessarily a superset of _k. (In alternate embodiments, OptiVax can be used to augment the base set B with peptides from B′ using scoring matrix S′₂ to return set _k, and the final vaccine set _k+|B| consists of peptides B∪_k.)

In some embodiments, this procedure is repeated independently for each single target of interest, and the resulting independent vaccine sets can be merged into a combined vaccine as described below.

Methods for Combining Multiple Vaccines

The above described methods will produce an optimized target peptide set (e.g., third peptide set) for one or more individual targets. In some embodiments, a method is provided for designing separate vaccines for MEW class I and class II based immunity for multiple targets (e.g., two or more targets such as KRAS G12D and KRAS G12V).

In some embodiments, a method is disclosed for producing a combined peptide vaccine for multiple targets by using a table of presentations for a disease that is based upon empirical data from sources such as the Cancer Genome Atlas (TCGA). FIG. 5 shows one embodiment for factoring disease presentation type probabilities (pancreatic cancer, colon/rectum cancer, and bronchus/lung cancer) by probability, for each disease presentation, of target presented for various KRAS mutation targets (KRAS G12D, KRAS G12V, and KRAS G12R). A presentation is a unique set of targets that are presented by one form of a disease (e.g., distinct type of cancer as shown in FIG. 5). For each presentation, FIG. 5 shows an example of the probability of that presentation, and the probability that a given target is observed. For a given presentation, there can be one or more targets, each having a probability. In some embodiments, the method for multi-target vaccine design will allocate peptide resources for inducing disease immunity based on the presentation and respective target probabilities as shown in FIG. 5, for example. In some embodiments, presentations correspond to the prevalence of targets in different human populations or different risk groups. The probability of a target in a population is computed by summing for each possible presentation the probability of that presentation times the probability of the target in that presentation.

Referring to FIG. 6, in some embodiments, the method first includes designing an individual peptide vaccine for each target to create a combined vaccine design for multiple targets. This initially results in sets of target-specific vaccine designs. In some embodiments, the marginal predicted vaccine performance of each target-specific vaccine at size k is defined by predicted vaccine performance at size k minus the predicted vaccine performance of the vaccine at size k minus one (see FIGS. 3-4). The composition of a vaccine may change as the number of peptides used in the vaccine increases, and thus for computing contributions to a combined vaccine the marginal predicted vaccine performance of each target-specific vaccine is used instead of a specific set of peptides.

In some embodiments, the weighted marginal predicted vaccine performance of a target-specific vaccine design for each target specific vaccine size is computed as shown in FIG. 6. For a given target specific vaccine size, its weighted predicted vaccine performance is computed by multiplying its predicted vaccine performance times the probability of the target in the population (e.g., by using values as shown in FIG. 5). The marginal weighted predicted vaccine performance for a target specific vaccine is its weighted coverage at size k minus its coverage at size k minus one (e.g., see FIGS. 3-4). The marginal weighted predicted vaccine performance of a target specific vaccine of size one is its weighted predicted vaccine performance. The marginal weighted predicted vaccine performances for all vaccines are combined into a single list, and the combined list is sorted from largest to least by the weighted marginal predicted vaccine performances of the target specific vaccines as shown in FIG. 6. The combined vaccine of size n is then determined by the first n elements of this list. The peptides for the combined vaccine are determined by the individual peptide target vaccines whose sizes add to n and whose weighted predicted vaccine performances sums to the same sum as the first n elements of the sorted list. This maximizes the predicted vaccine performance of the combined vaccine of size n.

In some embodiments, the combined multiple target vaccine can be designed on its overall predicted coverage for the disease described depending on the presentation table used (e.g., see FIG. 5), by its predicted coverage for a specific indication, and/or by its predicted coverage for a specific target by adjusting the weighting used for predicted vaccine performance accordingly. Once a desired level of coverage is selected, the peptides of the combined vaccine are determined by the contributions of target-specific designs. For example, if the combined vaccine includes a target-specific vaccine of size k, then the vaccine peptides for this target at size k are used in the combined vaccine.

As an example of one embodiment, FIG. 5 shows three mutations (KRAS G12D, G12V, and G12R) and their respective probabilities of occurring in an individual with pancreatic cancer. FIG. 3 (MHC class I) and FIG. 4 (MHC class II) show the population coverage of target-specific vaccines for the KRAS G12D, G12V, G12R, G12C, and G13D targets using the methods for vaccines described herein. The marginal population coverage of each target-specific vaccine at a given vaccine size is the improvement in coverage at that size and the size less one. The coverage with no peptides is zero. The marginal coverage of each target-specific vaccine is multiplied by the probability of the target in the population as determined by the proportions as shown in FIG. 5 for the pancreas (pancreatic cancer). These weighted marginal coverages of all target-specific vaccines are sorted to determine the best target-specific compositions, and the resulting list describes the composition of a combined vaccine at each size k by taking the first k elements of the list. As an example of one embodiment, FIG. 7 (MHC Class I) and FIG. 8 (MHC Class II) show the target specific contributions at each vaccine size for a combined KRAS vaccine for the three mutations KRAS G12D, G12V, and G12R. The methods for combined vaccine protocol described herein was used to compute the examples in FIGS. 7 and 8. At each combined vaccine size, different components of the target-specific vaccines are utilized. Table 1 (below) contains the peptides present in independent (single target) and combined (multiple target) MHC class I vaccine designs for the KRAS G12D, G12V, G12R, G12C, and G13D targets. Table 2 (below) contains the peptides present in independent (single target) MHC class II vaccine designs for the KRAS G12D, G12V, G12R, G12C, and G13D targets, and any subset of the individual/single target vaccines can be combined to create an MHC class II vaccine for two or more multiple targets. For alternate embodiments, the Sequence Listing provides heteroclitic peptides useful in MHC class I vaccines for the KRAS G12D, G12V, G12R, G12C, and G13D targets.

Combined Vaccine Design Procedure

In some embodiments, the procedure described herein is used to combine individual compact vaccines optimized for different targets into a single optimized combined vaccine.

In some embodiments, the computational inputs for the procedure are:

• • : Set of neoantigen or pathogenic targets of interest (e.g., KRAS G12D, KRAS G12V, KRAS G12R)
  • : Vaccine sets optimized individually for each target. Let _t,k denote the optimal vaccine set of exactly k peptides for target t∈ (e.g., as computed by the procedures describe above). Note that _t,k+1 may not necessarily be a superset of _t,k.
  • W: →[0,1]: Target weighting function mapping each target t∈ to a probability or weight of t in a particular presentation of interest (e.g., pancreatic cancer; see Exhibit A, Table 1 for example).
  • POPULATIONCOVERAGE: →[0,1]: Function mapping a peptide set into population coverage (e.g., EvalVax). This function may also take as input additional parameters, including HLA haplotype frequencies and a minimum per-individual number of peptide-HLA hits N (here, we compute coverage as P(n≥1) using EvalVax-Robust).

For each target t (individually) and vaccine size (peptide count) k, the unweighted population coverage c_t,k is computed:

c _t,k=PopulationCoverage(_t,k)∀t∈,k

• • Note that for each target t, c_t,k is generally monotonically increasing and concave down for increasing values of k (each additional peptide increases coverage but with decreasing returns).

For each target t (individually), the marginal coverage m_t,k is computed of the k-th peptide added to the vaccine set:

$m_{t,k}=\{\begin{array}{cc}{c}_{t,k}& \mathrm{if}k=1\\ c_{t,k}-c_{t,k-1},& \mathrm{otherwise}\end{array}\forall t\in 𝒯,k$

Note that for each target t, m_t,k should be a monotonically decreasing function in k (by Step 1 above).

The weighted marginal population coverage {tilde over (m)}_t,k is computed using weights of each target in W:

{tilde over (m)} _t,k =W(t)·m_t,k∀t∈,k

• • The weighted marginal population coverage gives the effective marginal coverage of the k-th peptide in the vaccine weighted by the prevalence of the target in the presentation (by multiplication with the probability/weight of the target in the presentation).

The individual vaccines are combined into a combined vaccine via the MERGEMULTI procedure called on the weighted marginal population coverage lists {tilde over (m)}_t=[{tilde over (m)}_t,k, k∈1, 2, . . . ]. FIG. 9 shows an example Python implementation of the MERGEMULTI function. This procedure takes as input multiple sorted (descending) lists and merges them into a single sorted (descending) list. Let M indicate the output of MERGEMULTI where each element M_k contains both the marginal weighted coverage and source (target) of the k-th peptide in the combined vaccine. The combined vaccine contains peptides from different targets. In particular, the combined vaccine with k peptides contains C_t,k=Σ_j≤k{M_k from t} peptides from target t. Note that C_t,k∈[0, . . . , k] and Σ_tC_t,k=k (C_t,k gives the distribution of the k peptides in the combined vaccine across the targets).

The optimal combined vaccine set _k is defined as:

${\hat{v}}_k=\bigcup _{t\in 𝒯}{v}_{t,C_{t,k}}$

Thus, the combined vaccine with k peptides is the combination of the optimal individual (C_t,k)-peptide vaccines. The marginal weighted coverage values of the combined vaccine M_k can be cumulatively summed over k to give the overall effective (target-weighted) population coverage of the combined vaccine containing k peptides as Σ_j≤KM_k (taking into account both the probabilities/weights of the targets in the presentation and the expected population coverage of peptides based on HLA display). The final vaccine size k can vary based upon the specific population coverage goals of the vaccine.

MHC Class I Peptide Sequences

In some embodiments, a peptide vaccine (single target or combined multiple target vaccine) comprises about five, ten, or twenty MEW class I peptides with each peptide consisting of 8 or more amino acids. In some embodiments, an MHC class I peptide vaccine is intended for one or more of the KRAS G12D, G12V, and G12R targets. In some embodiments, the amino acid sequence of a first peptide in a five-peptide combined vaccine comprises SEQ ID NO: 1. GADGVGKSM (SEQ ID NO: 1). In some embodiments, the amino acid sequence of a second peptide in a five-peptide combined vaccine comprises SEQ ID NO: 2. LMVVGADGV (SEQ ID NO: 2). In some embodiments, the amino acid sequence of a third peptide in a five-peptide combined vaccine comprises SEQ ID NO: 3. GAVGVGKSL (SEQ ID NO: 3). In some embodiments, the amino acid sequence of a fourth peptide in a five-peptide combined vaccine comprises SEQ ID NO: 4. LMVVGAVGV (SEQ ID NO: 4). In some embodiments, the amino acid sequence of a fifth peptide in a five-peptide combined vaccine comprises SEQ ID NO: 5. VTGARGVGK (SEQ ID NO: 5). An example combined vaccine for the KRAS G12D, G12V, and G12R targets with five peptides (SEQ ID NO: 1 to SEQ ID NO: 5) is predicted to have a weighted population coverage of 0.3620.

In some embodiments, any one of the peptides (peptides 1-5) in the five-peptide vaccine comprise an amino acid sequence 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98 or 99% identical to SEQ ID NO: 1, SEQ ID NO: 2, SEQ ID NO: 3, SEQ ID NO: 4, or SEQ ID NO: 5.

In some embodiments, the amino acid sequence of peptides 1 to 5 in a ten-peptide combined vaccine comprise SEQ ID NO: 1, SEQ ID NO: 2, SEQ ID NO: 3, SEQ ID NO: 4, and SEQ ID NO: 5. In some embodiments, the amino acid sequence of a sixth peptide in a ten-peptide combined vaccine comprises SEQ ID NO: 6. VMGAVGVGK (SEQ ID NO: 6). In some embodiments, the amino acid sequence of a seventh peptide in a ten-peptide combined vaccine comprises SEQ ID NO: 7. VVGAVGVGK (SEQ ID NO: 7). In some embodiments, the amino acid sequence of an eight peptide in a ten-peptide combined vaccine comprises SEQ ID NO: 8. GARGVGKSY (SEQ ID NO: 8). In some embodiments, the amino acid sequence of a ninth peptide in a ten-peptide combined vaccine comprises SEQ ID NO: 9. GPRGVGKSA (SEQ ID NO: 9). In some embodiments, the amino acid sequence of a tenth peptide in a ten-peptide combined vaccine comprises SEQ ID NO: 10. LMVVGARGV (SEQ ID NO: 10). An example combined vaccine for the KRAS G12D, G12V, and G12R targets with ten peptides (SEQ ID NO: 1 to SEQ ID NO: 10) is predicted to have a weighted population coverage of 0.4374.

In some embodiments, any one of the peptides (peptides 1-10) in the ten-peptide vaccine comprise an amino acid sequence 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98 or 99% identical to SEQ ID NO: 1, SEQ ID NO: 2, SEQ ID NO: 3, SEQ ID NO: 4, SEQ ID NO: 5, SEQ ID NO: 6, SEQ ID NO: 7, SEQ ID NO: 8, SEQ ID NO: 9, or SEQ ID NO: 10.

In some embodiments, the amino acid sequence of peptides 1 to 10 in a twenty-peptide combined vaccine comprise SEQ ID NO: 1, SEQ ID NO: 2, SEQ ID NO: 3, SEQ ID NO: 4, SEQ ID NO: 5, SEQ ID NO: 6, SEQ ID NO: 7, SEQ ID NO: 8, SEQ ID NO: 9, and SEQ ID NO: 10. In some embodiments, the amino acid sequence of an 11^th peptide in a twenty-peptide combined vaccine comprises SEQ ID NO: 11. GADGVGKSL (SEQ ID NO: 11). In some embodiments, the amino acid sequence of a 12^th peptide in a twenty-peptide combined vaccine comprises SEQ ID NO: 12. GADGVGKSY (SEQ ID NO: 12). In some embodiments, the amino acid sequence of a 13^th peptide in a twenty-peptide combined vaccine comprises SEQ ID NO: 13. GYDGVGKSM (SEQ ID NO: 13). In some embodiments, the amino acid sequence of a 14^th peptide in a twenty-peptide combined vaccine comprises SEQ ID NO: 14. GPVGVGKSV (SEQ ID NO: 14). In some embodiments, the amino acid sequence of a 15^th peptide in a twenty-peptide combined vaccine comprises SEQ ID NO: 15. LTVVGAVGV (SEQ ID NO: 15). In some embodiments, the amino acid sequence of a 16^th peptide in a twenty-peptide combined vaccine comprises SEQ ID NO: 16. VVGAVGVGR (SEQ ID NO: 16). In some embodiments, the amino acid sequence of a 17^th peptide in a twenty-peptide combined vaccine comprises SEQ ID NO: 17. GARGVGKSM (SEQ ID NO: 17). In some embodiments, the amino acid sequence of an 18^th peptide in a twenty-peptide combined vaccine comprises SEQ ID NO: 18. GPRGVGKSV (SEQ ID NO: 18). In some embodiments, the amino acid sequence of a 19^th peptide in a twenty-peptide combined vaccine comprises SEQ ID NO: 19. LLVVGARGV (SEQ ID NO: 19). In some embodiments, the amino acid sequence of a 20^th peptide in a twenty-peptide combined vaccine comprises SEQ ID NO: 20. VAGARGVGM (SEQ ID NO: 20). An example combined vaccine for the KRAS G12D, G12V, and G12R targets with twenty peptides (SEQ ID NO: 1 to SEQ ID NO: 20) is predicted to have a weighted population coverage of 0.4604.

In some embodiments, any one of the peptides (peptides 1-20) in the twenty-peptide vaccine comprise an amino acid sequence 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98 or 99% identical to SEQ ID NO: 1, SEQ ID NO: 2, SEQ ID NO: 3, SEQ ID NO: 4, SEQ ID NO: 5, SEQ ID NO: 6, SEQ ID NO: 7, SEQ ID NO: 8, SEQ ID NO: 9, and SEQ ID NO: 10, SEQ ID NO: 11, SEQ ID NO: 12, SEQ ID NO: 13, SEQ ID NO: 14, SEQ ID NO: 15, SEQ ID NO: 16, SEQ ID NO: 17, SEQ ID NO: 18, SEQ ID NO: 19, or SEQ ID NO: 20

Table 1 shows MHC class I peptide sequences described herein including the respective SEQ ID NO, amino acid sequence corresponding to the SEQ ID NO, KRAS protein target (with specific mutation), the seed amino acid sequence (i.e., the amino acid sequence of the wild type KRAS fragment), the amino acid substitution (if any) for heteroclitic peptides at positions 2 and 9, and notes detailing embodiments in which the peptide may be included in a 5, 10, or 20 combined peptide vaccine as described herein. Table 1 also includes additional peptide sequences comprising SEQ ID NOs: 21-41. In some embodiments, any combination of peptides listed in Table 1 (SEQ ID NOs: 1-41) may be used to create a combined peptide vaccine having between about 2 and about 40 peptides. In some embodiments, any one of the peptides (peptides 1-41; SEQ ID NOs: 1-41) in the combined vaccine comprises an amino acid sequence 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98 or 99% identical to any of SEQ ID NOs: 1-41.

TABLE 1
Example KRAS Vaccine Peptides (MHC class I)
	Sequence			Heteroclitic	Heteroclitic
SEQ ID	corresponding			Modification	Modification
NO	to SEQ ID	Target	Seed	P2	P9	Note
SEQ ID	GADGVGKSM	KRAS	GADGVGKSA	—	A9M	Individual KRAS
NO: 1		G12D				G12D (MHCflurry);
						Combined (5 peptide)
						(MHCflurry);
						Combined (10
						peptide) (MHCflurry);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	LMVVGADGV	KRAS	LVVVGADGV	V2M	—	Individual KRAS
NO: 2		G12D				G12D (MHCflurry);
						Individual KRAS
						G12D (NetMHCpan);
						Combined (5 peptide)
						(MHCflurry);
						Combined (10
						peptide) (MHCflurry);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	GAVGVGKSL	KRAS	GAVGVGKSA	—	A9L	Individual KRAS
NO: 3		G12V				G12V (MHCflurry);
						Combined (5 peptide)
						(MHCflurry);
						Combined (10
						peptide) (MHCflurry);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	LMVVGAVGV	KRAS	LVVVGAVGV	V2M	—	Individual KRAS
NO: 4		G12V				G12V (MHCflurry);
						Individual KRAS
						G12V (NetMHCpan);
						Combined (5 peptide)
						(MHCflurry);
						Combined (10
						peptide) (MHCflurry);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	VTGARGVGK	KRAS	VVGARGVGK	V2T	—	Individual KRAS
NO: 5		G12R				G12R (MHCflurry);
						Combined (5 peptide)
						(MHCflurry);
						Combined (10
						peptide) (MHCflurry);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	VMGAVGVGK	KRAS	VVGAVGVGK	V2M	—	Individual KRAS
NO: 6		G12V				G12V (MHCflurry);
						Combined (10
						peptide) (MHCflurry);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	VVGAVGVGK	KRAS	VVGAVGVGK	—	—	Individual KRAS
NO: 7		G12V				G12V (MHCflurry);
						Individual KRAS
						G12V (NetMHCpan);
						Combined (10
						peptide) (MHCflurry);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	GARGVGKSY	KRAS	GARGVGKSA	—	A9Y	Individual KRAS
NO: 8		G12R				G12R (MHCflurry);
						Combined (10
						peptide) (MHCflurry);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	GPRGVGKSA	KRAS	GARGVGKSA	A2P	—	Individual KRAS
NO: 9		G12R				G12R (MHCflurry);
						Combined (10
						peptide) (MHCflurry);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	LMVVGARGV	KRAS	LVVVGARGV	V2M	—	Individual KRAS
NO: 10		G12R				G12R (MHCflurry);
						Individual KRAS
						G12R (NetMHCpan);
						Combined (10
						peptide) (MHCflurry);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	GADGVGKSL	KRAS	GADGVGKSA	—	A9L	Individual KRAS
NO: 11		G12D				G12D (MHCflurry);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	GADGVGKSY	KRAS	GADGVGKSA	—	A9Y	Individual KRAS
NO: 12		G12D				G12D (MHCflurry);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	GYDGVGKSM	KRAS	GADGVGKSA	A2Y	A9M	Individual KRAS
NO: 13		G12D				G12D (MHCflurry);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	GPVGVGKSV	KRAS	GAVGVGKSA	A2P	A9V	Combined (20
NO: 14		G12V				peptide) (MHCflurry)
SEQ ID	LTVVGAVGV	KRAS	LVVVGAVGV	V2T	—	Individual KRAS
NO: 15		G12V				G12V (NetMHCpan);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	VVGAVGVGR	KRAS	VVGAVGVGK	—	K9R	Individual KRAS
NO: 16		G12V				G12V (MHCflurry);
						Individual KRAS
						G12V (NetMHCpan);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	GARGVGKSM	KRAS	GARGVGKSA	—	A9M	Combined (20
NO: 17		G12R				peptide) (MHCflurry)
SEQ ID	GPRGVGKSV	KRAS	GARGVGKSA	A2P	A9V	Combined (20
NO: 18		G12R				peptide) (MHCflurry)
SEQ ID	LLVVGARGV	KRAS	LVVVGARGV	V2L	—	Individual KRAS
NO: 19		G12R				G12R (NetMHCpan);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	VAGARGVGM	KRAS	VVGARGVGK	V2A	K9M	Individual KRAS
NO: 20		G12R				G12R (MHCflurry);
						Combined (20
						peptide) (MHCflurry)
SEQ ID	LTVVGADGV	KRAS	LVVVGADGV	V2T	—	Individual KRAS
NO: 21		G12D				G12D (NetMHCpan)
SEQ ID	LLVVGADGV	KRAS	LVVVGADGV	V2L	—	Individual KRAS
NO: 22		G12D				G12D (NetMHCpan)
SEQ ID	LMVVGADGL	KRAS	LVVVGADGV	V2M	V9L	Individual KRAS
NO: 23		G12D				G12D (NetMHCpan)
SEQ ID	VMGAVGVGR	KRAS	VVGAVGVGK	V2M	K9R	Individual KRAS
NO: 24		G12V				G12V (NetMHCpan)
SEQ ID	VMGARGVGK	KRAS	VVGARGVGK	V2M	—	Individual KRAS
NO: 25		G12R				G12R (NetMHCpan)
SEQ ID	GACGVGKSL	KRAS	GACGVGKSA	—	A9L	Individual KRAS
NO: 26		G12C				G12C (MHCflurry)
SEQ ID	LMVVGACGV	KRAS	LVVVGACGV	V2M	—	Individual KRAS
NO: 27		G12C				G12C (MHCflurry);
						Individual KRAS
						G12C (NetMHCpan)
SEQ ID	LTVVGACGV	KRAS	LVVVGACGV	V2T	—	Individual KRAS
NO: 28		G12C				G12C (MHCflurry);
						Individual KRAS
						G12C (NetMHCpan)
SEQ ID	VTGACGVGK	KRAS	VVGACGVGK	V2T	—	Individual KRAS
NO: 29		G12C				G12C (MHCflurry)
SEQ ID	VVGACGVGR	KRAS	VVGACGVGK	—	K9R	Individual KRAS
NO: 30		G12C				G12C (MHCflurry)
SEQ ID	AADVGKSAM	KRAS	AGDVGKSAL	G2A	L9M	Individual KRAS
NO: 31		G13D				G13D (MHCflurry);
						Individual KRAS
						G13D (NetMHCpan)
SEQ ID	AEDVGKSAM	KRAS	AGDVGKSAL	G2E	L9M	Individual KRAS
NO: 32		G13D				G13D (MHCflurry)
SEQ ID	AYDVGKSAM	KRAS	AGDVGKSAL	G2Y	L9M	Individual KRAS
NO: 33		G13D				G13D (MHCflurry)
SEQ ID	DAGKSALTV	KRAS	DVGKSALTI	V2A	I9V	Individual KRAS
NO: 34		G13D				G13D (MHCflurry)
SEQ ID	GAGDVGKSM	KRAS	GAGDVGKSA	—	A9M	Individual KRAS
NO: 35		G13D				G13D (MHCflurry)
SEQ ID	LQVVGACGV	KRAS	LVVVGACGV	V2Q	—	Individual KRAS
NO: 36		G12C				G12C (NetMHCpan)
SEQ ID	VMGACGVGK	KRAS	VVGACGVGK	V2M	—	Individual KRAS
NO: 37		G12C				G12C (NetMHCpan)
SEQ ID	VMGACGVGR	KRAS	VVGACGVGK	V2M	K9R	Individual KRAS
NO: 38		G12C				G12C (NetMHCpan)
SEQ ID	AADVGKSAL	KRAS	AGDVGKSAL	G2A	—	Individual KRAS
NO: 39		G13D				G13D (NetMHCpan)
SEQ ID	ASDVGKSAL	KRAS	AGDVGKSAL	G2S	—	Individual KRAS
NO: 40		G13D				G13D (NetMHCpan)
SEQ ID	ASDVGKSAM	KRAS	AGDVGKSAL	G2S	L9M	Individual KRAS
NO: 41		G13D				G13D (NetMHCpan)

Additional amino acid sequences of MHC class I heteroclitic peptides are provided in Sequence Listings (SEQ ID NOs: 67-1522). In some embodiments, any combination of MHC class I peptides disclosed herein (SEQ ID NOs: 1-41 and 67-1522) may be used to create a combined peptide vaccine having between about 2 and about 40 peptides. In some embodiments, any one of the peptides (SEQ ID NOs: 1-41 and 67-1522) in the combined vaccine comprises or contains an amino acid sequence 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98 or 99% identical to any of SEQ ID NOs: 1-41 or 67-1522.

MHC Class II Peptide Sequences

In some embodiments, a peptide vaccine (single target or combined multiple target vaccine) comprises about 2 to 40 MHC class II peptides with each peptide consisting of about 20 amino acids. I n some embodiments, an MHC class II peptide vaccine is intended for one or more of the KRAS G12D, G12V, G12R, G12C, and G13D targets.

Table 2 summarizes MHC class II peptide sequences described herein including the respective SEQ ID NO, amino acid sequence corresponding to the SEQ ID NO, the amino acid sequence corresponding to the peptide's binding core, the KRAS protein target (with specific mutation), the seed amino acid sequence (i.e., the amino acid sequence of the wild type KRAS fragment), the seed amino acid sequence of the binding core, and the amino acid substitution (if any) for heteroclitic peptides at positions 1, 4, 6, and 9. Table 2 includes peptide sequences comprising SEQ ID NOs: 42-66. SEQ ID NOs: 42-65 (Table 2) encode for recombinant peptides. In some embodiments, any combination of peptides listed in Table 2 (SEQ ID NOs: 42-66) may be used to create a single target (individual) or combined peptide vaccine having between about 2 and about 40 peptides. In some embodiments, any one of the peptides (peptides 42-66; SEQ ID NOs: 42-66) in the combined vaccine comprises an amino acid sequence 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98 or 99% identical to any of SEQ ID NOs: 42-66.

TABLE 2
Example KRAS Vaccine Peptides (MHC class II)
						Hetero-	Hetero-	Hetero-	Hetero-
						clitic	clitic	clitic	clitic
	Sequence					Modifi-	Modifi-	Modifi-	Modifi-
SEQ	corresponding				Seed	cation	cation	cation	cation
ID NO	to SEQ ID	Core	Target	Seed	Core	P1	P4	P6	P9	Note
SEQ	EYKFVVFGSDGAG	FVVF	KRAS	EYKLVVV	LVVV	L1F	V4F	A6S	V9A	Individual
ID	KS	GSDG	G12D	GADGVGK	GADG					KRAS
NO:		A		S	V					G12D
42										(NetMHCIIpan)
SEQ	EYKFVVIGNDGAG	FVVI	KRAS	EYKLVVV	LVVV	L1F	V4I	A6N	V9A	Individual
ID	KSALTIQLIQN	GNDG	G12D	GADGVGK	GADG					KRAS
NO:		A		SALTIQL	V					G12D
43				IQN						(NetMHCIIpan)
SEQ	EYKFVVLGADGAG	FVVL	KRAS	EYKLVVV	LVVV	L1F	V4L	—	V9A	Individual
ID	KS	GADG	G12D	GADGVGK	GADG					KRAS
NO:		A		S	V					G12D
44										(NetMHCIIpan)
SEQ	MTEYKFVVSGADG	FVVS	KRAS	MTEYKLV	LVVV	L1F	V4S	—	V9I	Individual
ID	IGKSALT	GADG	G12D	VVGADGV	GADG					KRAS
NO:		I		GKSALT	V					G12D
45										(NetMHCIIpan)
SEQ	MTEYKFVVYGSDG	FVVY	KRAS	MTEYKLV	LVVV	L1F	V4Y	A6S	V9I	Individual
ID	IGKSALT	GSDG	G12D	VVGADGV	GADG					KRAS
NO:		I		GKSALT	V					G12D
46										(NetMHCIIpan)
SEQ	EYKFVVIGRVGHG	FVVI	KRAS	EYKLVVV	LVVV	L1F	V4I	A6R	V9H	Individual
ID	KS	GRVG	G12V	GAVGVGK	GAVG					KRAS
NO:		H		S	V					G12V
47										(NetMHCIIpan)
SEQ	EYKFVVLGTVGHG	FVVL	KRAS	EYKLVVV	LVVV	L1F	V4L	A6T	V9H	Individual
ID	KS	GTVG	G12V	GAVGVGK	GAVG					KRAS
NO:		H		S	V					G12V
48										(NetMHCIIpan)
SEQ	EYKFVVYGNVGMG	FVVY	KRAS	EYKLVVV	LVVV	L1F	V4Y	A6N	V9M	Individual
ID	KS	GNVG	G12V	GAVGVGK	GAVG					KRAS
NO:		M		S	V					G12V
49										(NetMHCIIpan)
SEQ	EYKIVVAGNVGIG	IVVA	KRAS	EYKLVVV	LVVV	L1I	V4A	A6N	V9I	Individual
ID	KS	GNVG	G12V	GAVGVGK	GAVG					KRAS
NO:		I		S	V					G12V
50										(NetMHCIIpan)
SEQ	TEYKIVVMGNVGY	IVVM	KRAS	TEYKLVV	LVVV	L1I	V4M	A6N	V9Y	Individual
ID	GK	GNVG	G12V	VGAVGVG	GAVG					KRAS
NO:		Y		K	V					G12V
51										(NetMHCIIpan)
SEQ	MTEYKFVVFGSRG	FVVF	KRAS	MTEYKLV	LVVV	L1F	V4F	A6S	—	Individual
ID	VGKSALT	GSRG	G12R	VVGARGV	GARG					KRAS
NO:		V		GKSALT	V					G12R
52										(NetMHCIIpan)
SEQ	MTEYKFVVIGNRG	FVVI	KRAS	MTEYKLV	LVVV	L1F	V4I	A6N	—	Individual
ID	VGKSALT	GNRG	G12R	VVGARGV	GARG					KRAS
NO:		V		GKSALT	V					G12R
53										(NetMHCIIpan)
SEQ	MTEYKFVVIGVRG	FVVI	KRAS	MTEYKLV	LVVV	L1F	V4I	A6V	V9D	Individual
ID	DGKSALT	GVRG	G12R	VVGARGV	GARG					KRAS
NO:		D		GKSALT	V					G12R
54										(NetMHCIIpan)
SEQ	MTEYKFVVMGSRG	FVVM	KRAS	MTEYKLV	LVVV	L1F	V4M	A6S	V9A	Individual
ID	AGKSALT	GSRG	G12R	VVGARGV	GARG					KRAS
NO:		A		GKSALT	V					G12R
55										(NetMHCIIpan)
SEQ	VVVIARGVPKSLL	IARG	KRAS	VVVGARG	GARG	G1I	—	G6P	A9L	Individual
ID	TI	VPKS	G12R	VGKSALT	VGKS					KRAS
NO:		L		I	A					G12R
56										(NetMHCIIpan)
SEQ	EYKFVVFGNCGAG	FVVF	KRAS	EYKLVVV	LVVV	L1F	V4F	A6N	V9A	Individual
ID	KS	GNCG	G12C	GACGVGK	GACG					KRAS
NO:		A		S	V					G12C
57										(NetMHCIIpan)
SEQ	EYKFVVSGACGVG	FVVS	KRAS	EYKLVVV	LVVV	L1F	V4S	—	—	Individual
ID	KS	GACG	G12C	GACGVGK	GACG					KRAS
NO:		V		S	V					G12C
58										(NetMHCIIpan)
SEQ	EYKFVVSGNCGLG	FVVS	KRAS	EYKLVVV	LVVV	L1F	V4S	A6N	V9L	Individual
ID	KS	GNCG	G12C	GACGVGK	GACG					KRAS
NO:		L		S	V					G12C
59										(NetMHCIIpan)
SEQ	EYKLVVMGPCGAG	LVVM	KRAS	EYKLVVV	LVVV	—	V4M	A6P	V9A	Individual
ID	KS	GPCG	G12C	GACGVGK	GACG					KRAS
NO:		A		S	V					G12C
60										(NetMHCIIpan)
SEQ	KLVIVGICKVGHS	IVGI	KRAS	KLVVVGA	VVGA	V1I	A4I	G6K	K9H	Individual
ID	AL	CKVG	G12C	CGVGKSA	CGVG					KRAS
NO:		H		L	K					G12C
61										(NetMHCIIpan)
SEQ	EYKFVVFGNGDLG	FVVF	KRAS	EYKLVVV	LVVV	L1F	V4F	A6N	V9L	Individual
ID	KS	GNGD	G13D	GAGDVGK	GAGD					KRAS
NO:		L		S	V					G13D
62										(NetMHCIIpan)
SEQ	EYKFVVMGNGDSG	FVVM	KRAS	EYKLVVV	LVVV	L1F	V4M	A6N	V9S	Individual
ID	KS	GNGD	G13D	GAGDVGK	GAGD					KRAS
NO:		S		S	V					G13D
63										(NetMHCIIpan)
SEQ	EYKFVVSGSGDVG	FVVS	KRAS	EYKLVVV	LVVV	L1F	V4S	A6S	—	Individual
ID	KS	GSGD	G13D	GAGDVGK	GAGD					KRAS
NO:		V		S	V					G13D
64										(NetMHCIIpan)
SEQ	EYKIVVMGRGDMG	IVVM	KRAS	EYKLVVV	LVVV	L1I	V4M	A6R	V9M	Individual
ID	KS	GRGD	G13D	GAGDVGK	GAGD					KRAS
NO:		M		S	V					G13D
65										(NetMHCIIpan)
SEQ	YKLVVVGAGDVGK	—	KRAS	—	—	—	—	—	—	Individual
ID	SA		G13D							KRAS
NO:										G13D
66										(NetMHCIIpan)

In some embodiments, any combination of MEW class I and/or MEW class II peptides disclosed herein (SEQ ID NOs: 1-1522) may be used to create a single target (individual) or combined peptide vaccine having between about 2 and about 40 peptides. In some embodiments, any one of the peptides (peptides 1-1522; SEQ ID NOs: 1-1522) in the combined vaccine comprises an amino acid sequence 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98 or 99% identical to any of SEQ ID NOs: 1-1522.

mRNA and DNA Vaccines

In some embodiments, vaccine peptides are encoded as mRNA or DNA molecules and are administered for expression in vivo as is known in the art. One example of the delivery of vaccines by mRNA is found in Kranz et al. (2016), incorporated herein by reference. In one embodiment, a construct comprises 10 peptides, including a five-peptide MHC class I combined pancreatic cancer vaccine (targets: KRAS G12D, G12V, G12R) and a five-peptide MEW class II combined pancreatic cancer vaccine (targets: KRAS G12D, G12V, G12R), as optimized by the procedure described herein. Peptides are prepended with a secretion signal sequence at the N-terminus and followed by an MEW class I trafficking signal (MITD) (Kreiter et al., 2008; Sahin et al., 2017). The MITD has been shown to route antigens to pathways for HLA class I and class II presentation (Kreiter et al., 2008). Here we combine all peptides of each MEW class into a single construct using non-immunogenic glycine/serine linkers from Sahin et al. (2017), though it is also plausible to construct individual constructs containing single peptides with the same secretion and MITD signals as demonstrated by Kreiter et al. (2008).

In some embodiments, the amino acid sequence encoded by the mRNA vaccine comprises SEQ ID NO: 1523. Underlined amino acids correspond to the signal peptide (or leader) sequence. Bolded amino acids correspond to MHC class I (9 amino acids in length; 5 peptides) and MHC class II (13-25 amino acids in length; 5 peptides) peptide sequences. Italicized amino acids correspond to the trafficking signal.

(SEQ ID NO: 1523)
MRVTAPRTLILLLSGALALTETWAGSGGSGGGGSGGGADGVGKSMGGSG
GGGSGGLMVVGADGVGGSGGGGSGGGAVGVGKSLGGSGGGGSGGLMVVG
AVGVGGSGGGGSGGVTGARGVGKGGSGGGGSGGEYKFVVLGTVGHGKSG
GSGGGGSGGEYKIVVAGNVGIGKSGGSGGGGSGGEYKFVVFGSDGAGKS
GGSGGGGSGGMTEYKFVVSGADGIGKSALTGGSGGGGSGGMTEYKFVVI
GNRGVGKSALTGGSLGGGGSGIVGIVAGLAVLAVVVIGAVVATVMCRRK
SSGGKGGSYSQAASSDSAQGSDVSLTA.

In some embodiments, the vaccine is an mRNA vaccine comprising a nucleic acids sequence encoding the amino acid sequence consisting of SEQ ID NO: 1523. In some embodiments, the nucleic acid sequence of the mRNA vaccine encodes for an amino acid sequence 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98 or 99% identical to SEQ ID NO: 1523.

In some embodiments, the vaccine is a DNA vaccine comprising a nucleic acids sequence encoding the amino acid sequence consisting of SEQ ID NO: 1523. In some embodiments, the nucleic acid sequence of the DNA vaccine encodes for an amino acid sequence 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98 or 99% identical to SEQ ID NO: 1523.

In some embodiments, one or more MHC class I and/or MHC class II peptides disclosed herein (SEQ ID NO: 1-1522) can be encoded in one or more mRNA or DNA molecules and administered for expression in vivo. In some embodiments between about 2 and about 40 peptide sequences are encoded in one or more mRNA constructs. In some embodiments, between about 2 and about 40 peptide sequences are encoded in one or more DNA constructs (i.e., nucleic acids encoding the amino acids sequences comprising on or more of SEQ ID NOs: 1-1522). In some embodiments, the amino acid sequence of the mRNA vaccine or the nucleic acid sequence of the DNA vaccine encodes for an amino acid sequence 80, 81, 82, 83, 84, 85, 86, 87, 88, 89, 90, 91, 92, 93, 94, 95, 96, 97, 98 or 99% identical to any of SEQ ID NOs: 1-1522.

Non-Limiting Embodiments of the Subject Matter

In one aspect, the invention provides for a system for selecting an immunogenic peptide composition comprising a processor, and a memory storing processor-executable instructions that, when executed by the processor, cause the processor to create a first peptide set by selecting a plurality of base peptides, wherein at least one peptide of the plurality of base peptides is associated with a disease, create a second peptide set by adding to the first peptide set a modified peptide, wherein the modified peptide comprises a substitution of at least one residue of a base peptide selected from the plurality of base peptides, and create a third peptide set by selecting a subset of the second peptide set, wherein the selected subset of the second peptide set has a predicted vaccine performance, wherein the predicted vaccine performance has a population coverage above a predetermined threshold, and wherein the subset comprises at least one peptide of the second peptide set.

In some embodiments, the plurality of base peptides of the first peptide set is derived from a target protein, wherein the target protein is a tumor neoantigen or a pathogen proteome. In some embodiments, selecting the plurality of base peptides to create the first peptide set comprises sliding a window of size n across an amino acid sequence encoding the target protein, wherein n is between about 8 amino acids and about 25 amino acids in length, and wherein n is a length of each peptide of the plurality of base peptides of the first peptide set. In some embodiments, a peptide of the plurality of base peptides binds to an HLA class I molecule or an HLA class II molecule. In some embodiments, the substitution of the at least one residue comprises substituting an amino acid at an anchor residue position for a different amino acid at the anchor residue position. In some embodiments, the system further comprises filtering the first peptide set to exclude a peptide with a predicted binding core that contains a target residue in an anchor position. In some embodiments, the second peptide set comprises the first peptide set. In some embodiments, the prediction to be bound by the one or more HLA alleles is computed using a binding affinity less than about 1000 nM. In some embodiments, the predicted vaccine performance is determined by computing a plurality of peptide-HLA immunogenicities of the third peptide set to at least one HLA allele. In some embodiments, each peptide-HLA immunogenicity of the plurality of peptide-HLA immunogenicities of the third peptide set is based on a predicted binding affinity of less than about 500 nM. In some embodiments, the predicted vaccine performance is based on a population coverage, wherein the population coverage is computed based on a frequency of an HLA haplotype in a human population. In some embodiments, the predicted vaccine performance is based on a population coverage, wherein the population coverage is computed based on a frequency of at least two HLA alleles in a human population. In some embodiments, the plurality of base peptides is present in a single subject. In some embodiments, the predicted vaccine performance is an expected number of peptide-HLA hits. In some embodiments, the disease is cancer, and wherein the cancer is selected from the group consisting of pancreas, colon, rectum, kidney, bronchus, lung, uterus, cervix, bladder, liver, and stomach. In some embodiments, the plurality of base peptides of the first peptide set comprises at least one self-peptide.

In another aspect, the invention provides for a non-transitory computer-readable storage medium comprising computer-readable instructions for determining an immunogenic peptide composition that, when executed by a processor cause the processor to create a first peptide set by selecting a plurality of base peptides, wherein a first base peptide and a second base peptide of the plurality of base peptides are each scored for binding by two or more HLA alleles, wherein the first base peptide and the second base peptide are each predicted to be bound by one or more HLA alleles, and wherein the first base peptide and the second base peptide are associated with a disease, create a second peptide set comprising the first base peptide, the second base peptide, a first modified peptide, and a second modified peptide, wherein the first modified peptide comprises a substitution of at least one residue of the first base peptide, and wherein the second modified peptide comprises a substitution of at least one residue of the second base peptide, and create a third peptide set by selecting a subset of the second peptide set, wherein the selected subset of the second peptide set has a predicted vaccine performance, and wherein the predicted vaccine performance is a function of a peptide-HLA immunogenicity of at least one peptide of the third peptide set with respect to the two or more HLA alleles.

In some embodiments, the plurality of base peptides of the first peptide set is derived from a target protein, wherein the target protein is a tumor neoantigen or a pathogen proteome. I n some embodiments, selecting the plurality of base peptides to create the first peptide set comprises sliding a window of size n across an amino acid sequence encoding the target protein, wherein n is between about 8 amino acids and about 25 amino acids in length, and wherein n is a length of each peptide of the plurality of base peptides of the first peptide set. In some embodiments, a peptide of the plurality of base peptides binds to an HLA class I molecule or an HLA class II molecule. In some embodiments, the substitution of the at least one residue comprises substituting an amino acid at an anchor residue position for a different amino acid at the anchor residue position. In some embodiments, the non-transitory computer-readable storage medium of further comprises filtering the first peptide set to exclude a peptide with a predicted binding core that contains a target residue in an anchor position. In some embodiments, the second peptide set comprises the first peptide set. In some embodiments, the prediction to be bound by the two or more HLA alleles is computed using a binding affinity less than about 1000 nM. In some embodiments, the plurality of base peptides of the first peptide set comprises at least one self-peptide.

In another aspect, the invention provides for a system for selecting an immunogenic peptide composition comprising a processor, and a memory storing processor-executable instructions that, when executed by the processor, cause the processor to create a first peptide set by selecting a plurality of base peptides, wherein a first base peptide of the plurality of base peptides is scored for binding by three or more HLA alleles, wherein the first base peptide is predicted to be bound by one or more HLA alleles, and wherein the first base peptide is associated with a disease, create a second peptide set comprising the first base peptide and a modified peptide, wherein the modified peptide comprises a substitution of at least one residue of the first base peptide, and create a third peptide set by selecting a subset of the second peptide set, wherein the selected subset of the second peptide set has a predicted vaccine performance, and wherein the predicted vaccine performance is a function of a peptide-HLA immunogenicity of at least one peptide of the third peptide set with respect to the three or more HLA alleles.

In some embodiments, the first base peptide is scored for binding based on data obtained from experimental assays. In some embodiments, the predicted vaccine performance includes a peptide-HLA immunogenicity of the modified peptide bound to the first HLA allele of the one or more HLA alleles if the first base peptide is predicted to be bound to the first HLA allele of the one or more HLA alleles with a first binding core, wherein the first binding core is a binding core of the first base peptide, wherein the first binding core is identical to a second binding core, and wherein the second binding core is a binding core of the modified peptide bound to the first HLA allele.

In another aspect, the invention provides for a non-transitory computer-readable storage medium comprising computer-readable instructions for determining an immunogenic peptide composition that, when executed by a processor cause the processor to create a first peptide set by selecting a plurality of base peptides, wherein at least one peptide of the plurality of base peptides is associated with a disease, create a second peptide set comprising a first base peptide selected from the first base peptide set and a modified peptide, wherein the modified peptide comprises a substitution of at least one residue of the first base peptide, and create a third peptide set by selecting a subset of the second peptide set, wherein the selected subset of the second peptide set has a predicted vaccine performance, wherein the predicted vaccine performance has an expected number of peptide-HLA hits above a predetermined threshold, and wherein the subset comprises at least one peptide of the second peptide set.

In some embodiments, the first base peptide binds to an HLA class I molecule or an HLA class II molecule.

In another aspect, the invention provides for a system for selecting an immunogenic peptide composition comprising a processor, and a memory storing processor-executable instructions that, when executed by the processor, cause the processor to create a first peptide set by selecting a first plurality of peptides, wherein the first plurality of peptides comprises a plurality of target peptides that are associated with a first disease, and wherein the first peptide set has a first predicted vaccine performance value, create a second peptide set by selecting a second plurality of peptides, wherein the second plurality of peptides comprises a plurality of target peptides that are associated with a second disease, and wherein the second peptide set has a second predicted vaccine performance value, create a first weighted peptide set by multiplying a first weight by the first predicted vaccine performance value, create a second weighted peptide set multiplying a second weight by the second predicted vaccine performance value, and create a third peptide set by combining the first weighted peptide set and the second weighted peptide set.

In some embodiments, the first predicted vaccine performance value and the second predicted vaccine performance value are computed based on a population coverage of a vaccine. In some embodiments, the first predicted vaccine performance value and the second predicted vaccine performance value are computed based on an expected number of peptide-HLA hits. In some embodiments, the first plurality of peptides is derived from a tumor neoantigen or a pathogen proteome. In some embodiments, the second plurality of peptides is derived from a tumor neoantigen or a pathogen proteome. In some embodiments, the first disease is cancer, and wherein the cancer is selected from the group consisting of pancreas, colon, rectum, kidney, bronchus, lung, uterus, cervix, bladder, liver, and stomach. In some embodiments, the second disease is cancer, and wherein the cancer is selected from the group consisting of pancreas, colon, rectum, kidney, bronchus, lung, uterus, cervix, bladder, liver, and stomach. In some embodiments, the first plurality of peptides comprises a peptide that binds to an HLA class I molecule or an HLA class II molecule. In some embodiments, the second plurality of peptides comprises a peptide that binds to an HLA class I molecule or an HLA class II molecule.

Compositions

In some embodiments, a peptide vaccine comprises one or more peptides of this disclosure and is administered in a pharmaceutical composition that includes a pharmaceutically acceptable carrier. In some embodiments, the peptide vaccine is comprised of the third peptide set, as described in this disclosure. In some embodiments, the pharmaceutical composition is in the form of a spray, aerosol, gel, solution, emulsion, lipid nanoparticle, nanoparticle, or suspension.

The composition is preferably administered to a subject with a pharmaceutically acceptable carrier. Typically, in some embodiments, an appropriate amount of a pharmaceutically acceptable salt is used in the formulation, which in some embodiments can render the formulation isotonic.

In certain embodiments, the peptides are provided as an immunogenic composition comprising any one of the peptides described herein and a pharmaceutically acceptable carrier. In certain embodiments, the immunogenic composition further comprises an adjuvant. In certain embodiments, the peptides are conjugated with other molecules to increase their effectiveness as is known by those practiced in the art. For example, peptides can be coupled to antibodies that recognize cell surface proteins on antigen presenting cells to enhance vaccine effectiveness. One such method for increasing the effectiveness of peptide delivery is described in Woodham, et al. (2018). In certain embodiments for the treatment of autoimmune disorders, the peptides are delivered with a composition and protocol designed to induce tolerance as is known in the art. Example methods for using peptides for immune tolerization are described in Alhadj Ali, et al. (2017) and Gibson, et al. (2015).

In some embodiments, the pharmaceutically acceptable carrier is selected from the group consisting of saline, Ringer's solution, dextrose solution, and a combination thereof. Other suitable pharmaceutically acceptable carriers known in the art are contemplated. Suitable carriers and their formulations are described in Remington's Pharmaceutical Sciences, 2005, Mack Publishing Co. The pH of the solution is preferably from about 5 to about 8, and more preferably from about 7 to about 7.5. The formulation may also comprise a lyophilized powder. Further carriers include sustained release preparations such as semipermeable matrices of solid hydrophobic polymers, which matrices are in the form of shaped articles, e.g., films, liposomes or microparticles. It will be apparent to those persons skilled in the art that certain carriers may be more preferable depending upon, for instance, the route of administration and concentration of peptides being administered.

The phrase pharmaceutically acceptable carrier as used herein means a pharmaceutically acceptable material, composition or vehicle, such as a liquid or solid filler, diluent, excipient, solvent or encapsulating material, involved in carrying or transporting the subject pharmaceutical agent from one organ, or portion of the body, to another organ, or portion of the body. Each carrier is acceptable in the sense of being compatible with the other ingredients of the formulation and not injurious to the patient. Some examples of materials which can serve as pharmaceutically acceptable carriers include: sugars, such as lactose, glucose and sucrose; starches, such as corn starch and potato starch; cellulose, and its derivatives, such as sodium carboxymethyl cellulose, ethyl cellulose and cellulose acetate; powdered tragacanth; malt; gelatin; talc; excipients, such as cocoa butter and suppository waxes; oils, such as peanut oil, cottonseed oil, safflower oil, sesame oil, olive oil, corn oil and soybean oil; glycols, such as butylene glycol; polyols, such as glycerin, sorbitol, mannitol and polyethylene glycol; esters, such as ethyl oleate and ethyl laurate; agar; buffering agents, such as magnesium hydroxide and aluminum hydroxide; alginic acid; pyrogen-free water; isotonic saline; Ringer's solution; ethyl alcohol; phosphate buffer solutions; and other non-toxic compatible substances employed in pharmaceutical formulations. The term carrier denotes an organic or inorganic ingredient, natural or synthetic, with which the active ingredient is combined to facilitate the application. The components of the pharmaceutical compositions also are capable of being comingled with the compounds of the present invention, and with each other, in a manner such that there is no interaction which would substantially impair the desired pharmaceutical efficiency. The composition may also include additional agents such as an isotonicity agent, a preservative, a surfactant, and, a divalent cation, preferably, zinc.

The composition can also include an excipient, or an agent for stabilization of a peptide composition, such as a buffer, a reducing agent, a bulk protein, amino acids (such as e.g., glycine or praline) or a carbohydrate. Bulk proteins useful in formulating peptide compositions include albumin. Typical carbohydrates useful in formulating peptides include but are not limited to sucrose, mannitol, lactose, trehalose, or glucose.

Surfactants may also be used to prevent soluble and insoluble aggregation and/or precipitation of peptides or proteins included in the composition. Suitable surfactants include but are not limited to sorbitan trioleate, soya lecithin, and oleic acid. In certain cases, solution aerosols are preferred using solvents such as ethanol. Thus, formulations including peptides can also include a surfactant that can reduce or prevent surface-induced aggregation of peptides by atomization of the solution in forming an aerosol. Various conventional surfactants can be employed, such as polyoxyethylene fatty acid esters and alcohols, and polyoxyethylene sorbitol fatty acid esters. Amounts will generally range between 0.001% and 4% by weight of the formulation. In some embodiments, surfactants used with the present disclosure are polyoxyethylene sorbitan mono-oleate, polysorbate 80, polysorbate 20. Additional agents known in the art can also be included in the composition.

In some embodiments, the pharmaceutical compositions and dosage forms further comprise one or more compounds that reduce the rate by which an active ingredient will decay, or the composition will change in character. So called stabilizers or preservatives may include, but are not limited to, amino acids, antioxidants, pH buffers, or salt buffers. Nonlimiting examples of antioxidants include butylated hydroxy anisole (BHA), ascorbic acid and derivatives thereof, tocopherol and derivatives thereof, butylated hydroxy anisole and cysteine. Nonlimiting examples of preservatives include parabens, such as methyl or propyl p-hydroxybenzoate and benzalkonium chloride. Additional nonlimiting examples of amino acids include glycine or proline.

The present invention also teaches the stabilization (preventing or minimizing thermally or mechanically induced soluble or insoluble aggregation and/or precipitation of an inhibitor protein) of liquid solutions containing peptides at neutral pH or less than neutral pH by the use of amino acids including proline or glycine, with or without divalent cations resulting in clear or nearly clear solutions that are stable at room temperature or preferred for pharmaceutical administration.

In one embodiment, the composition is a pharmaceutical composition of single unit or multiple unit dosage forms. Pharmaceutical compositions of single unit or multiple unit dosage forms of the invention comprise a prophylactically or therapeutically effective amount of one or more compositions (e.g., a compound of the invention, or other prophylactic or therapeutic agent), typically, one or more vehicles, carriers, or excipients, stabilizing agents, and/or preservatives. Preferably, the vehicles, carriers, excipients, stabilizing agents and preservatives are pharmaceutically acceptable.

In some embodiments, the pharmaceutical compositions and dosage forms comprise anhydrous pharmaceutical compositions and dosage forms. Anhydrous pharmaceutical compositions and dosage forms of the invention can be prepared using anhydrous or low moisture containing ingredients and low moisture or low humidity conditions. Pharmaceutical compositions and dosage forms that comprise lactose and at least one active ingredient that comprise a primary or secondary amine are preferably anhydrous if substantial contact with moisture and/or humidity during manufacturing, packaging, and/or storage is expected. An anhydrous pharmaceutical composition should be prepared and stored such that its anhydrous nature is maintained. Accordingly, anhydrous compositions are preferably packaged using materials known to prevent exposure to water such that they can be included in suitable formulary kits. Examples of suitable packaging include, but are not limited to, hermetically sealed foils, plastics, unit dose containers (e.g., vials), blister packs, and strip packs.

Suitable vehicles are well known to those skilled in the art of pharmacy, and non-limiting examples of suitable vehicles include glucose, sucrose, starch, lactose, gelatin, rice, silica gel, glycerol, talc, sodium chloride, dried skim milk, propylene glycol, water, sodium stearate, ethanol, and similar substances well known in the art. Saline solutions and aqueous dextrose and glycerol solutions can also be employed as liquid vehicles. Whether a particular vehicle is suitable for incorporation into a pharmaceutical composition or dosage form depends on a variety of factors well known in the art including, but not limited to, the way in which the dosage form will be administered to a patient and the specific active ingredients in the dosage form. Pharmaceutical vehicles can be sterile liquids, such as water and oils, including those of petroleum, animal, vegetable or synthetic origin, such as peanut oil, soybean oil, mineral oil, sesame oil and the like.

The invention also provides that a pharmaceutical composition can be packaged in a hermetically sealed container such as an ampoule or sachette indicating the quantity. In one embodiment, the pharmaceutical composition can be supplied as a dry sterilized lyophilized powder in a delivery device suitable for administration to the lower airways of a patient. The pharmaceutical compositions can, if desired, be presented in a pack or dispenser device that can contain one or more unit dosage forms containing the active ingredient. The pack can for example comprise metal or plastic foil, such as a blister pack. The pack or dispenser device can be accompanied by instructions for administration.

Methods of preparing these formulations or compositions include the step of bringing into association a compound of the present invention with the carrier and, optionally, one or more accessory ingredients. In general, the formulations are prepared by uniformly and intimately bringing into association a compound of the present invention with liquid carriers, or finely divided solid carriers, or both, and then, if necessary, shaping the product.

Formulations of the invention suitable for administration may be in the form of powders, granules, or as a solution or a suspension in an aqueous or non-aqueous liquid, or as an oil-in-water or water-in-oil liquid emulsion, or as an elixir or syrup, or as pastilles (using an inert base, such as gelatin and glycerin, or sucrose and acacia) and/or as mouthwashes and the like, each containing a predetermined amount of a compound of the present invention (e.g., peptides) as an active ingredient.

A liquid composition herein can be used as such with a delivery device, or they can be used for the preparation of pharmaceutically acceptable formulations comprising peptides that are prepared for example by the method of spray drying. The methods of spray freeze-drying peptides/proteins for pharmaceutical administration disclosed in Maa et al., Curr. Pharm. Biotechnol., 2001, 1, 283-302, are incorporated herein. In another embodiment, the liquid solutions herein are freeze spray dried and the spray-dried product is collected as a dispersible peptide-containing powder that is therapeutically effective when administered to an individual.

The compounds and pharmaceutical compositions of the present invention can be employed in combination therapies, that is, the compounds and pharmaceutical compositions can be administered concurrently with, prior to, or subsequent to, one or more other desired therapeutics or medical procedures (e.g., peptide vaccine can be used in combination therapy with another treatment such as chemotherapy, radiation, pharmaceutical agents, and/or another treatment). The particular combination of therapies (therapeutics or procedures) to employ in a combination regimen will take into account compatibility of the desired therapeutics and/or procedures and the desired therapeutic effect to be achieved. It will also be appreciated that the therapies employed may achieve a desired effect for the same disorder (for example, the compound of the present invention may be administered concurrently with another therapeutic or prophylactic).

The invention also provides a pharmaceutical pack or kit comprising one or more containers filled with one or more of the ingredients of the pharmaceutical compositions of the invention. Optionally associated with such container(s) can be a notice in the form prescribed by a governmental agency regulating the manufacture, use or sale of pharmaceuticals or biological products, which notice reflects approval by the agency of manufacture, use or sale for human administration.

The current invention provides for dosage forms comprising peptides suitable for treating cancer or other diseases. The dosage forms can be formulated, e.g., as sprays, aerosols, nanoparticles, liposomes, or other forms known to one of skill in the art. See, e.g., Remington's Pharmaceutical Sciences; Remington: The Science and Practice of Pharmacy supra; Pharmaceutical Dosage Forms and Drug Delivery Systems by Howard C., Ansel et al., Lippincott Williams & Wilkins; 7th edition (Oct. 1, 1999).

Generally, a dosage form used in the acute treatment of a disease may contain larger amounts of one or more of the active ingredients it comprises than a dosage form used in the chronic treatment of the same disease. In addition, the prophylactically and therapeutically effective dosage form may vary among different conditions. For example, a therapeutically effective dosage form may contain peptides that has an appropriate immunogenic action when intending to treat cancer or other disease. On the other hand, a different effective dosage may contain peptides that has an appropriate immunogenic action when intending to use the peptides of the invention as a prophylactic (e.g., vaccine) against cancer or another disease/condition. These and other ways in which specific dosage forms encompassed by this invention will vary from one another and will be readily apparent to those skilled in the art. See, e.g., Remington's Pharmaceutical Sciences, 2005, Mack Publishing Co.; Remington: The Science and Practice of Pharmacy by Gennaro, Lippincott Williams & Wilkins; 20th edition (2003); Pharmaceutical Dosage Forms and Drug Delivery Systems by Howard C. Ansel et al., Lippincott Williams & Wilkins; 7th edition (Oct. 1, 1999); and Encyclopedia of Pharmaceutical Technology, edited by Swarbrick, J. & J. C. Boylan, Marcel Dekker, Inc., New York, 1988, which are incorporated herein by reference in their entirety.

The pH of a pharmaceutical composition or dosage form may also be adjusted to improve delivery and/or stability of one or more active ingredients. Similarly, the polarity of a solvent carrier, its ionic strength, or tonicity can be adjusted to improve delivery. Compounds such as stearates can also be added to pharmaceutical compositions or dosage forms to alter advantageously the hydrophilicity or lipophilicity of one or more active ingredients to improve delivery. In this regard, stearates can also serve as a lipid vehicle for the formulation, as an emulsifying agent or surfactant, and as a delivery enhancing or penetration-enhancing agent. Different salts, hydrates, or solvates of the active ingredients can be used to adjust further the properties of the resulting composition.

Compositions can be formulated with appropriate carriers and adjuvants using techniques to yield compositions suitable for immunization. The compositions can include an adjuvant, such as, for example but not limited to, alum, poly IC, MF-59, squalene-based adjuvants, or liposomal based adjuvants suitable for immunization.

In some embodiments, the compositions and methods comprise any suitable agent or immune modulation which could modulate mechanisms of host immune tolerance and release of the induced antibodies. In certain embodiments, an immunomodulatory agent is administered in at time and in an amount sufficient for transient modulation of the subject's immune response so as to induce an immune response which comprises antibodies against for example tumor neoantigens (i.e., tumor-specific antigens (TSA)).

Expression Systems

In certain aspects, the invention provides culturing a cell line that expresses any one of the peptides of the invention in a culture medium comprising any of the peptides described herein.

Various expression systems for producing recombinant proteins/peptides are known in the art, and include, prokaryotic (e.g., bacteria), plant, insect, yeast, and mammalian expression systems. Suitable cell lines, can be transformed, transduced, or transfected with nucleic acids containing coding sequences for the peptides of the invention in order to produce the molecule of interest. Expression vectors containing such a nucleic acid sequence, which can be linked to at least one regulatory sequence in a manner that allows expression of the nucleotide sequence in a host cell, can be introduced via methods known in the art. Practitioners in the art understand that designing an expression vector can depend on factors, such as the choice of host cell to be transfected and/or the type and/or amount of desired protein to be expressed. Enhancer regions, which are those sequences found upstream or downstream of the promoter region in non-coding DNA regions, are also known in the art to be important in optimizing expression. If needed, origins of replication from viral sources can be employed, such as if a prokaryotic host is utilized for introduction of plasmid DNA. However, in eukaryotic organisms, chromosome integration is a common mechanism for DNA replication. For stable transfection of mammalian cells, a small fraction of cells can integrate introduced DNA into their genomes. The expression vector and transfection method utilized can be factors that contribute to a successful integration event. For stable amplification and expression of a desired protein, a vector containing DNA encoding a protein of interest is stably integrated into the genome of eukaryotic cells (for example mammalian cells), resulting in the stable expression of transfected genes. A gene that encodes a selectable marker (for example, resistance to antibiotics or drugs) can be introduced into host cells along with the gene of interest in order to identify and select clones that stably express a gene encoding a protein of interest. Cells containing the gene of interest can be identified by drug selection wherein cells that have incorporated the selectable marker gene will survive in the presence of the drug. Cells that have not incorporated the gene for the selectable marker die. Surviving cells can then be screened for the production of the desired protein molecule.

A host cell strain, which modulates the expression of the inserted sequences, or modifies and processes the nucleic acid in a specific fashion desired also may be chosen. Such modifications (for example, glycosylation and other post-translational modifications) and processing (for example, cleavage) of peptide/protein products may be important for the function of the peptide/protein. Different host cell strains have characteristic and specific mechanisms for the post-translational processing and modification of proteins and gene products. As such, appropriate host systems or cell lines can be chosen to ensure the correct modification and processing of the target protein expressed. Thus, eukaryotic host cells possessing the cellular machinery for proper processing of the primary transcript, glycosylation, and phosphorylation of the gene product may be used.

Various culturing parameters can be used with respect to the host cell being cultured. Appropriate culture conditions for mammalian cells are well known in the art (Cleveland W L, et al., J Immunol Methods, 1983, 56(2): 221-234) or can be determined by the skilled artisan (see, for example, Animal Cell Culture: A Practical Approach 2nd Ed., Rickwood, D. and Hames, B. D., eds. (Oxford University Press: New York, 1992)). Cell culturing conditions can vary according to the type of host cell selected. Commercially available medium can be utilized.

Peptides of the invention can be purified from any human or non-human cell which expresses the polypeptide, including those which have been transfected with expression constructs that express peptides of the invention. For protein recovery, isolation and/or purification, the cell culture medium or cell lysate is centrifuged to remove particulate cells and cell debris. The desired polypeptide molecule is isolated or purified away from contaminating soluble proteins and polypeptides by suitable purification techniques. Non-limiting purification methods for proteins include: size exclusion chromatography; affinity chromatography; ion exchange chromatography; ethanol precipitation; reverse phase HPLC; chromatography on a resin, such as silica, or cation exchange resin, e.g., DEAE; chromatofocusing; SDS-PAGE; ammonium sulfate precipitation; gel filtration using, e.g., Sephadex G-75, Sepharose; protein A sepharose chromatography for removal of immunoglobulin contaminants; and the like. Other additives, such as protease inhibitors (e.g., PMSF or proteinase K) can be used to inhibit proteolytic degradation during purification. Purification procedures that can select for carbohydrates can also be used, e.g., ion-exchange soft gel chromatography, or HPLC using cation- or anionexchange resins, in which the more acidic fraction(s) is/are collected.

Methods of Treatment

In one embodiment, the subject matter disclosed herein relates to a preventive medical treatment started after following diagnosis of cancer in order to prevent the disease from worsening or curing the disease. In one embodiment, the subject matter disclosed herein relates to prophylaxis of subjects who are believed to be at risk for cancer or have previously been diagnosed with cancer (or another disease). In one embodiment, said subjects can be administered the peptide vaccine described herein or pharmaceutical compositions thereof. The invention contemplates using any of the peptides produced by the systems and methods described herein. In one embodiment, the peptide vaccines described herein can be administered subcutaneously via syringe or any other suitable method know in the art.

The compound(s) or combination of compounds disclosed herein, or pharmaceutical compositions may be administered to a cell, mammal, or human by any suitable means. Non-limiting examples of methods of administration include, among others, (a) administration though oral pathways, which includes administration in capsule, tablet, granule, spray, syrup, or other such forms; (b) administration through non-oral pathways such as intraocular, intranasal, intraauricular, rectal, vaginal, intraurethral, transmucosal, buccal, or transdermal, which includes administration as an aqueous suspension, an oily preparation or the like or as a drip, spray, suppository, salve, ointment or the like; (c) administration via injection, including subcutaneously, intraperitoneally, intravenously, intramuscularly, intradermally, intraorbitally, intracapsularly, intraspinally, intrasternally, or the like, including infusion pump delivery; (d) administration locally such as by injection directly in the renal or cardiac area, e.g., by depot implantation; (e) administration topically; as deemed appropriate by those of skill in the art for bringing the compound or combination of compounds disclosed herein into contact with living tissue; (f) administration via inhalation, including through aerosolized, nebulized, and powdered formulations; and (g) administration through implantation.

As will be readily apparent to one skilled in the art, the effective in vivo dose to be administered and the particular mode of administration will vary depending upon the age, weight and species treated, and the specific use for which the compound or combination of compounds disclosed herein are employed. The determination of effective dose levels, that is the dose levels necessary to achieve the desired result, can be accomplished by one skilled in the art using routine pharmacological methods. Typically, human clinical applications of products are commenced at lower dose levels, with dose level being increased until the desired effect is achieved. Alternatively, acceptable in vitro studies can be used to establish useful doses and routes of administration of the compositions identified by the present methods using established pharmacological methods. Effective animal doses from in vivo studies can be converted to appropriate human doses using conversion methods known in the art (e.g., see Nair A B, Jacob S. A simple practice guide for dose conversion between animals and human. Journal of basic and clinical pharmacy. 2016 March; 7(2):27.)

Methods of Prevention

In some embodiments, the peptides prepared using methods of the invention can be used as a vaccine to promote an immune response against cancer (e.g., against tumor neoantigens). In some embodiments, the invention provides compositions and methods for induction of immune response, for example induction of antibodies to tumor neoantigens. In some embodiments, the antibodies are broadly neutralizing antibodies. In some embodiments, the peptides prepared using methods of the invention can be used as a vaccine to promote an immune response against a pathogen. In some embodiments, the peptides prepared using methods of the invention can be used to promote immune tolerance as an autoimmune disease therapeutic.

The compositions, systems, and methods disclosed herein are not to be limited in scope to the specific embodiments described herein. Indeed, various modifications of the compositions, systems, and methods in addition to those described will become apparent to those of skill in the art from the foregoing description.

Claims

1. A method of forming an immunogenic peptide composition, the method comprising: using a processor to perform the steps of: creating a first peptide set by selecting one or more peptide sequences, wherein each peptide sequence of the one or more peptide sequences is derived from a first target, and wherein the first target is a protein sequence associated with a tumor neoantigen, a pathogen proteome, or a self-protein;creating a second peptide set by selecting two or more peptide sequences, wherein each peptide sequence of the two or more peptide sequences is derived from the first target, and wherein the second peptide set is larger than the first peptide set by one peptide sequence;computing, with respect to the first peptide set, a first predicted vaccine performance value;computing, with respect to the second peptide set, a second predicted vaccine performance value;computing a first marginal predicted vaccine performance value by obtaining a difference between the first predicted vaccine performance value and the second predicted vaccine performance value;computing a first weighted marginal predicted vaccine performance value based on the first marginal predicted vaccine performance value and a weight associated with the first target;creating a third peptide set by selecting one or more peptide sequences, wherein each peptide sequence of the one or more peptide sequences is derived from a second target, and wherein the second target is a protein sequence associated with a tumor neoantigen, a pathogen proteome, or a self-protein;creating a fourth peptide set by selecting two or more peptide sequences, wherein each peptide sequence of the two or more peptide sequences is derived from the second target, and wherein the fourth peptide set is larger than the third peptide set by one peptide sequence;computing, with respect to the third peptide set, a third predicted vaccine performance value;computing, with respect to the fourth peptide set, a fourth predicted vaccine performance value;computing a second marginal predicted vaccine performance value by obtaining a difference between the third predicted vaccine performance value and the fourth predicted vaccine performance value;computing a second weighted marginal predicted vaccine performance value based on the second marginal predicted vaccine performance value and a weight associated with the second target;creating a fifth peptide set comprising: the first peptide set or the second peptide set; andthe third peptide set or the fourth peptide set,wherein creating the fifth peptide set is based on the first weighted predicted marginal vaccine performance value and the second weighted predicted marginal performance value;performing an experimental assay to obtain a peptide-HLA immunogenicity metric for one or more peptide sequences of the fifth peptide set; andforming an immunogenic peptide composition comprising the one or more peptide sequences of the fifth peptide set for which the experimental assay was performed.

2. The method of claim 1, wherein the first peptide set comprises two or more peptide sequences and the second peptide set comprises three or more peptide sequences.

3. The method of claim 2, wherein the third peptide set comprises two or more peptide sequences and the fourth peptide set comprises three or more peptide sequences.

4. The method of claim 1, wherein computing the first marginal predicted vaccine performance value and the second marginal predicted vaccine performance value comprises computing a population coverage.

5. The method of claim 4, wherein the population coverage is computed based on a frequency of an HLA allele in a human population.

6. The method of claim 5, wherein the population coverage is computed based on a frequency of at least three HLA alleles in the human population.

7. The method of claim 1, wherein computing the first marginal predicted vaccine performance value and the second marginal predicted vaccine performance value comprises computing an expected number of peptide-HLA hits.

8. The method of claim 1, wherein the weight associated with the first target is a probability of a presentation of the first target in a subject diagnosed with a disease.

9. The method of claim 1, wherein the weight associated with the second target is a probability of a presentation of the second target in a subject diagnosed with a disease.

10. The method of claim 1, wherein the first target is the protein sequence associated with the tumor neoantigen, wherein the tumor neoantigen is expressed in a subject diagnosed with cancer, and wherein the cancer is selected from the group consisting of pancreatic cancer, colon cancer, rectal cancer, kidney cancer, bronchus cancer, lung cancer, uterine cancer, cervical cancer, bladder cancer, liver cancer, and stomach cancer.

11. The method of claim 1, wherein the second target is the protein sequence associated with the tumor neoantigen, wherein the tumor neoantigen is expressed in a subject diagnosed with cancer, and wherein the cancer is selected from the group consisting of pancreatic cancer, colon cancer, rectal cancer, kidney cancer, bronchus cancer, lung cancer, uterine cancer, cervical cancer, bladder cancer, liver cancer, and stomach cancer.

12. The method of claim 1, wherein computing the first marginal predicted vaccine performance value and the second marginal predicted vaccine performance value is based on an HLA type of a subject.

13. The method of claim 12, wherein computing the first marginal predicted vaccine performance value and the second marginal predicted vaccine performance value comprises computing an expected number peptide-HLA hits in the subject.

14. The method of claim 12, wherein computing the first marginal predicted vaccine performance value and second marginal predicted vaccine performance value comprises computing an expected fraction of HLA alleles in a subject that has a minimum number of peptide-HLA hits, wherein the minimum number of peptide-HLA hits is based on a threshold.

15. The method of claim 1, wherein each peptide sequence of the first peptide set, the second peptide set, the third peptide set, and the fourth peptide set is configured to bind to an HLA class I molecule or an HLA class II molecule expressed in a subject in vivo.