OpenNeoantigens Protocol

A Public Commons for Shared Cancer Vaccine Targets with Blockchain-Verified Attribution

Daniel Uribe, CEO GenoBank.io

GenoBank Research Team

January 2026

Version 1.0 | Technical Specification

Executive Summary

OpenNeoantigens is a proposed public database and protocol for shared cancer neoantigen sequences derived from common driver mutations. Unlike private/patient-specific neoantigens that require individual tumor sequencing, shared neoantigens from hotspot mutations (KRAS G12D, TP53 R175H, BRAF V600E, etc.) are identical across all patients harboring those mutations and can be pre-computed, validated, and made freely available.

This specification defines the data architecture, HLA-peptide binding matrices, mRNA sequence designs, and blockchain attribution layer that together enable a public commons for cancer vaccine development. By open-sourcing the ~30% of targetable neoantigens that derive from shared mutations, we can dramatically accelerate access to personalized cancer immunotherapy while preserving attribution for contributors through Story Protocol IP registration.

OpenNeoantigens is not a replacement for fully personalized vaccines—it is a complementary public good that provides immediate, validated targets for patients with common driver mutations, while preserving the economic framework for patient-owned private neoantigens through the BioNFT system.

1. Rationale: Why Open Source Neoantigens?
2. Neoantigen Taxonomy: Shared vs. Private
3. Database Architecture
4. HLA-Peptide Binding Matrix
5. Pre-Designed mRNA Sequences
6. Clinical Validation Framework
7. Blockchain Attribution Layer
8. API Specification
9. Governance Model
10. Integration with BioNFT Ecosystem
11. Implementation Roadmap
12. Appendix: Initial Mutation Set

1. Rationale: Why Open Source Neoantigens?

1.1 The Bottleneck Problem

Personalized neoantigen vaccines face a fundamental bottleneck: each patient requires individual tumor sequencing, neoantigen prediction, and custom mRNA synthesis. This process takes 6-8 weeks and costs $50,000-100,000+ per patient. For rapidly progressing cancers, this timeline can be fatal.

However, approximately 30% of actionable neoantigens derive from shared "hotspot" driver mutations that are identical across thousands or millions of patients:

93%

PDAC with KRAS mutations

50%

Melanoma with BRAF V600E

70%

Cancers with TP53 mutations

~200

Common hotspot mutations

These shared neoantigens can be pre-computed, pre-validated, and stockpiled—eliminating the 6-8 week delay for a significant fraction of targetable epitopes.

1.2 The Open Source Advantage

    What Open Sourcing Enables
    Immediate availability: No waiting for individual tumor sequencing for shared targets
Massive validation: Clinical response data pooled across thousands of patients
Manufacturing efficiency: CDMOs can stockpile common mRNA sequences
Global access: Low-resource settings can access validated vaccine designs
Combination strategies: Shared neoantigens + private neoantigens in single vaccine

1.3 What Remains Proprietary

OpenNeoantigens does NOT open-source:

Private/passenger neoantigens: Unique to each patient's tumor—these remain patient-owned via BioNFTs
Manufacturing processes: GMP production methods remain CDMO intellectual property
Delivery systems: LNP formulations, adjuvants, etc. remain proprietary
Clinical protocols: Dosing, scheduling, combination therapies require regulatory expertise

"OpenNeoantigens provides the blueprint. Manufacturing, delivery, and clinical implementation remain value-added services requiring specialized expertise."

2. Neoantigen Taxonomy: Shared vs. Private

2.1 Classification Framework

flowchart TB subgraph NEOANTIGENS["ALL NEOANTIGENS"] direction TB subgraph SHARED["SHARED (OpenNeoantigens)"] A["Driver Hotspots
KRAS G12D, BRAF V600E"] B["Recurrent Fusions
BCR-ABL, EML4-ALK"] C["Viral Oncoproteins
HPV E6/E7, EBV LMP"] end subgraph PRIVATE["PRIVATE (BioNFT-Protected)"] D["Passenger Mutations
Unique to each tumor"] E["Rare Variants
<1% population frequency"] F["Patient-Specific Fusions
Novel breakpoints"] end end SHARED --> G["OpenNeoantigens
Public Database"] PRIVATE --> H["Patient BioNFT
Sovereign Ownership"] G --> I["Free Access
Attribution Required"] H --> J["Consent-Gated
Economic Rights"]

Figure 1: Neoantigen classification determining open vs. proprietary status

2.2 Shared Neoantigen Criteria

A neoantigen qualifies for OpenNeoantigens inclusion if it meets ALL of the following criteria:

Criterion	Requirement	Rationale
Recurrence	Present in ≥1% of a cancer type	Sufficient patient population to benefit
Sequence Identity	100% amino acid identity across patients	Same peptide = same vaccine target
Expression	Consistently expressed when mutation present	Driver mutations typically constitutively expressed
Actionability	Predicted to bind ≥1 common HLA allele	Must be presentable to T cells
No IP Encumbrance	Sequence not covered by existing patents	Must be freely usable

2.3 Initial Target Categories

Category A: Oncogenic Driver Mutations

Gene	Hotspot Mutations	Cancer Types	Frequency
KRAS	G12D, G12V, G12C, G12R, G13D, Q61H	PDAC, CRC, NSCLC	25-95%
BRAF	V600E, V600K	Melanoma, CRC, Thyroid	40-60%
TP53	R175H, R248Q, R273H, R282W, G245S	Pan-cancer	50-70%
PIK3CA	E545K, H1047R, E542K	Breast, Endometrial, CRC	20-40%
NRAS	Q61R, Q61K, G12D	Melanoma, AML	15-30%
IDH1	R132H, R132C	Glioma, AML, Cholangiocarcinoma	70-90%
EGFR	L858R, T790M, Exon 19 del	NSCLC	10-50%

Category B: Oncogenic Fusion Proteins

Fusion	Breakpoint Peptide	Cancer Type	Frequency
BCR-ABL	Junction-spanning peptides	CML, ALL	95%+ of CML
EML4-ALK	Multiple variants (V1-V5)	NSCLC	3-7% of NSCLC
TMPRSS2-ERG	Junction peptides	Prostate	40-50%
PML-RARA	Junction peptides	APL	95%+ of APL

Category C: Viral Oncoproteins

Virus	Oncoproteins	Cancer Type	Association
HPV16/18	E6, E7	Cervical, Oropharyngeal, Anal	70%+ of cervical
EBV	LMP1, LMP2, EBNA1	Nasopharyngeal, Hodgkin, Gastric	Variable
HBV/HCV	HBx, Core, NS proteins	Hepatocellular carcinoma	70-80%
HTLV-1	Tax, HBZ	Adult T-cell leukemia	100%

3. Database Architecture

3.1 Data Model

erDiagram MUTATION ||--o{ PEPTIDE : generates MUTATION { string mutation_id PK string gene string protein_change string dna_change string cosmic_id string dbsnp_id float population_frequency string[] cancer_types } PEPTIDE ||--o{ HLA_BINDING : has PEPTIDE { string peptide_id PK string mutation_id FK string sequence int length int start_position string flanking_sequence boolean is_mutant_residue_central } HLA_ALLELE ||--o{ HLA_BINDING : participates HLA_ALLELE { string hla_id PK string gene string allele_group string protein float population_frequency string[] population_distribution } HLA_BINDING { string binding_id PK string peptide_id FK string hla_id FK float ic50_nm float percentile_rank string prediction_tool string binding_category boolean validated_experimentally } PEPTIDE ||--o{ MRNA_DESIGN : encodes MRNA_DESIGN { string mrna_id PK string peptide_id FK string codon_optimized_sequence string five_prime_utr string three_prime_utr int poly_a_length string cap_structure float predicted_expression } PEPTIDE ||--o{ CLINICAL_EVIDENCE : has CLINICAL_EVIDENCE { string evidence_id PK string peptide_id FK string hla_id FK string trial_id int num_patients int num_responders string response_type string publication_doi string data_source }

Figure 2: OpenNeoantigens database entity-relationship diagram

3.2 Core Tables Schema

{ "mutations": { "mutation_id": "KRAS_G12D", "gene": "KRAS", "transcript": "ENST00000311936", "protein_change": "p.G12D", "dna_change": "c.35G>A", "chromosome": "12", "position": 25398284, "ref": "C", "alt": "T", "cosmic_id": "COSM521", "clinvar_id": "12582", "population_frequency": 0.0, "somatic_frequency": { "PDAC": 0.41, "CRC": 0.13, "NSCLC": 0.04 }, "functional_impact": "oncogenic_driver", "druggability": "emerging", "references": ["PMID:33106528", "PMID:34534430"] } }

{ "peptides": { "peptide_id": "KRAS_G12D_9mer_pos7", "mutation_id": "KRAS_G12D", "sequence": "VVVGADGVG", "length": 9, "mutant_position": 4, "wildtype_sequence": "VVVGAGGVG", "flanking_n": "MTEYKLVVV", "flanking_c": "GVGKSALTI", "physicochemical": { "molecular_weight": 771.87, "isoelectric_point": 5.52, "hydrophobicity": 0.73, "instability_index": 32.4 } } }

{ "hla_binding": { "binding_id": "KRAS_G12D_9mer_HLA-A0201", "peptide_id": "KRAS_G12D_9mer_pos7", "hla_allele": "HLA-A*02:01", "predictions": { "netmhcpan_4.1": { "ic50_nm": 45.2, "percentile_rank": 0.12, "binding_level": "strong" }, "mhcflurry_2.0": { "ic50_nm": 52.8, "percentile_rank": 0.15, "binding_level": "strong" }, "consensus": { "ic50_nm": 49.0, "confidence": "high" } }, "experimental_validation": { "validated": true, "method": "competitive_binding_assay", "measured_ic50_nm": 38.5, "source": "IEDB:1234567" } } }

3.3 Storage Architecture

Distributed Storage Model

Primary Database: PostgreSQL with TimescaleDB for clinical time-series data
Search Index: Elasticsearch for peptide sequence similarity search
Blob Storage: IPFS for mRNA sequence files (pinned, with deletion capability via gateway)
Cache Layer: Redis for frequently accessed HLA-binding lookups
Blockchain Index: Story Protocol for attribution records (on-chain)

4. HLA-Peptide Binding Matrix

4.1 HLA Coverage Strategy

The human population expresses thousands of HLA alleles, but coverage can be achieved efficiently by targeting the most common alleles:

HLA Class	Priority Alleles	Cumulative Coverage
Class I (CD8+ T cells)	HLA-A*02:01	~40% of Caucasians
	HLA-A*01:01	+15% = 55%
	HLA-A*03:01	+12% = 67%
	HLA-A*11:01	+10% = 77%
	HLA-A*24:02	+8% = 85%
	+ Top 20 B alleles	>95% global coverage
Class II (CD4+ T cells)	HLA-DRB1*01:01	Common helper epitopes
	HLA-DRB1*03:01	Enhanced CD4+ response
	HLA-DRB1*04:01	Broader coverage

4.2 Binding Prediction Pipeline

flowchart LR A[Mutation] --> B[Generate All
8-11mer Peptides] B --> C{NetMHCpan 4.1} B --> D{MHCflurry 2.0} B --> E{NetMHCIIpan 4.0} C --> F[Class I
Predictions] D --> F E --> G[Class II
Predictions] F --> H{Consensus
Filter} G --> H H --> I["Strong Binders
(IC50 < 50nM)"] H --> J["Moderate Binders
(IC50 50-500nM)"] H --> K["Weak/Non-binders
(IC50 > 500nM)"] I --> L[Priority 1:
Validated Targets] J --> M[Priority 2:
Experimental Candidates]

Figure 3: Multi-tool consensus binding prediction pipeline

4.3 Example: KRAS G12D Binding Matrix

Peptide	Length	HLA-A*02:01	HLA-A*03:01	HLA-A*11:01	HLA-B*07:02
VVGADGVGK	9	2450 nM ❌	45 nM ✅	38 nM ✅	8920 nM ❌
VVVGADGVG	9	52 nM ✅	380 nM ⚠️	245 nM ⚠️	5670 nM ❌
KLVVVGADGV	10	28 nM ✅✅	1890 nM ❌	2340 nM ❌	7800 nM ❌
LVVVGADGVG	10	67 nM ✅	290 nM ⚠️	89 nM ✅	4500 nM ❌
VVVGADGVGK	10	3200 nM ❌	62 nM ✅	55 nM ✅	6700 nM ❌

Legend: ✅✅ Strong binder (<50 nM) | ✅ Good binder (50-150 nM) | ⚠️ Moderate (150-500 nM) | ❌ Weak/Non-binder (>500 nM)

Key Insight: HLA-Matched Peptide Selection

A patient with HLA-A*02:01 and KRAS G12D should receive the KLVVVGADGV 10-mer (IC50 = 28 nM).

A patient with HLA-A*11:01 and KRAS G12D should receive the VVGADGVGK 9-mer (IC50 = 38 nM).

Same mutation. Different optimal peptides. This is why HLA matching is critical.

5. Pre-Designed mRNA Sequences

5.1 mRNA Architecture

┌─────────────────────────────────────────────────────────────────────────────────┐
│                        OpenNeoantigens mRNA Construct                           │
├─────────────────────────────────────────────────────────────────────────────────┤
│                                                                                  │
│  5' Cap ─── 5' UTR ─── Signal ─── [Neoantigen Cassette] ─── 3' UTR ─── PolyA   │
│    │          │          │                  │                   │         │     │
│   Cap1    Optimized   Secretion      Multiple epitopes      Stabilizing  120nt │
│  (m7G)    Kozak       peptide        with linkers           elements           │
│                                                                                  │
│  ┌──────────────────────────────────────────────────────────────────────────┐  │
│  │                     NEOANTIGEN CASSETTE DETAIL                            │  │
│  │                                                                           │  │
│  │  [Epitope1]──AAY──[Epitope2]──AAY──[Epitope3]──AAY──...──[Epitope20]    │  │
│  │      │                                                        │          │  │
│  │   25-mer                                                   25-mer        │  │
│  │  (8aa flank + 9mer + 8aa flank)                                          │  │
│  │                                                                           │  │
│  │  AAY = Alanine-Alanine-Tyrosine linker (proteasomal cleavage site)      │  │
│  └──────────────────────────────────────────────────────────────────────────┘  │
│                                                                                  │
└─────────────────────────────────────────────────────────────────────────────────┘

Figure 4: mRNA vaccine construct architecture for multi-epitope delivery

5.2 Codon Optimization

Each peptide sequence is reverse-translated using optimized codons for human expression:

# Example: KRAS G12D epitope KLVVVGADGV

Wild-type AA:  K    L    V    V    V    G    A    D    G    V
Optimized:    AAG  CTG  GTG  GTG  GTG  GGC  GCC  GAC  GGC  GTG

Key optimizations:
- GC content: 55-65% (optimal for stability)
- Avoid rare codons (<10% usage)
- Eliminate mRNA secondary structures
- Remove cryptic splice sites
- Eliminate poly-U stretches (immunogenic)
- Substitute all U → N1-methylpseudouridine (Ψ)

5.3 mRNA Sequence Repository

{ "mrna_design": { "design_id": "OPEN-NEO-KRAS-G12D-A0201-v1", "target_mutation": "KRAS_G12D", "target_hla": "HLA-A*02:01", "peptide_sequence": "KLVVVGADGV", "construct": { "five_prime_cap": "CleanCap_AG", "five_prime_utr": "GGGAAAUAAGAGAGAAAAGAAGAGUAAGAAGAAAUAUAAGAGCCACC", "signal_peptide": "MFVFLVLLPLVSSQCVNLT", "epitope_with_flanks": "MTEYKLVVVKLVVVGADGVGKSALTI", "linker": "AAY", "three_prime_utr": "GCUGCCUUCUGCGGGGCUUGCCUUCUGGCCAUGCCCUUCUUCUCUCCC...", "poly_a_length": 120 }, "full_sequence_length": 847, "gc_content": 0.58, "predicted_half_life_hours": 18.5, "modifications": { "uridine": "N1-methylpseudouridine", "cap": "Cap1" }, "version": "1.0", "validation_status": "computationally_validated", "license": "CC-BY-SA-4.0", "attribution_ip_id": "0x1234...Story_Protocol_IP_ID" } }

5.4 Multi-Epitope Constructs

For patients with multiple targetable shared mutations, multi-epitope mRNA constructs can be pre-designed:

Example: PDAC Multi-Target Vaccine

Patient profile: KRAS G12D + TP53 R175H + CDKN2A loss + SMAD4 R361H

HLA type: A*02:01, A*03:01, B*07:02

OpenNeoantigens construct:

KRAS G12D epitopes for A*02:01 (KLVVVGADGV) and A*03:01 (VVGADGVGK)
TP53 R175H epitopes for A*02:01 and A*03:01
SMAD4 R361H epitopes if available

Construct ID: OPEN-NEO-PDAC-MULTI-A0201-A0301-v1

6. Clinical Validation Framework

6.1 Evidence Hierarchy

Level	Evidence Type	Requirements	Database Status
1	Clinical response	Documented T-cell response + clinical benefit in trial	VALIDATED
2	Immunogenic (human)	T-cell response in human study, no clinical data yet	VALIDATED
3	Immunogenic (preclinical)	T-cell response in mouse model with human HLA transgene	EXPERIMENTAL
4	Binding validated	Experimental peptide-MHC binding assay	EXPERIMENTAL
5	Predicted only	Computational prediction, no experimental validation	COMPUTATIONAL

6.2 Response Tracking Schema

{ "clinical_evidence": { "evidence_id": "CE-2024-00142", "peptide_id": "KRAS_G12D_10mer_pos5", "hla_allele": "HLA-A*02:01", "trial": { "trial_id": "NCT04117087", "phase": "1/2", "sponsor": "Elicio Therapeutics", "vaccine": "ELI-002" }, "cohort": { "n_patients": 25, "cancer_types": ["PDAC", "CRC"], "hla_matched": true }, "outcomes": { "t_cell_responders": 20, "response_rate": 0.80, "median_magnitude": "2.3-fold increase", "duration_months": 12, "clinical_benefit": { "ctdna_clearance": 0.45, "ca19_9_reduction": 0.84 } }, "publication": { "doi": "10.1038/s41591-024-XXXXX", "pubmed_id": "38XXXXXX" }, "data_contributed_by": "0xABC...contributor_wallet", "contribution_timestamp": "2024-06-15T14:30:00Z" } }

6.3 Community Validation Process

flowchart TD A[New Clinical Data
Submitted] --> B{Data Quality
Review} B -->|Pass| C[Curator Review] B -->|Fail| D[Request
Clarification] D --> A C --> E{Validation
Committee} E -->|Approved| F[Update Database] E -->|Questions| G[Request Additional
Documentation] G --> C F --> H[Attribution Recorded
on Story Protocol] H --> I[Contributor Receives
IP Credit] subgraph "Validation Committee" J[3+ Independent
Oncologists] K[Biostatistician] L[Immunologist] end

Figure 5: Clinical evidence validation and attribution workflow

7. Blockchain Attribution Layer

7.1 Story Protocol Integration

Every contribution to OpenNeoantigens is registered as an IP Asset on Story Protocol, ensuring permanent attribution:

    What Gets Attributed
    Sequence submissions: First contributor of a validated peptide-HLA binding pair
Clinical data: Researchers who contribute response data from trials
mRNA designs: Contributors of optimized mRNA constructs
Validation work: Labs that perform experimental binding assays
Curation: Reviewers who validate submissions

7.2 IP Asset Structure

{ "ip_asset": { "ip_id": "0x7890...OpenNeoantigens_KRAS_G12D", "ip_type": "OPEN_NEOANTIGEN", "metadata": { "name": "KRAS G12D Neoantigen Dataset", "description": "Complete HLA binding matrix and mRNA designs for KRAS G12D", "mutation": "KRAS_G12D", "content_hash": "QmXyz...IPFS_hash", "version": "2.3.0" }, "contributors": [ { "address": "0xAAA...", "role": "original_submitter", "contribution": "Initial peptide predictions", "share": 40 }, { "address": "0xBBB...", "role": "experimental_validator", "contribution": "Binding assay data", "share": 30 }, { "address": "0xCCC...", "role": "clinical_contributor", "contribution": "Phase 1 response data", "share": 30 } ], "license": { "type": "PIL_PUBLIC_GOOD", "terms": { "commercial_use": true, "attribution_required": true, "derivatives_allowed": true, "share_alike": true } } } }

7.3 License Framework

License Type	Commercial Use	Attribution	Derivatives	Use Case
PIL #4: Public Good	✅ Free	✅ Required	✅ Share-alike	Core database entries
PIL #1: Non-Commercial	❌ No	✅ Required	✅ Allowed	Research-only datasets
PIL #2: Commercial	✅ Royalty	✅ Required	✅ Negotiable	Pharma-contributed data

7.4 Attribution NFT

Contributors receive a non-transferable Attribution NFT recording their contribution:

// OpenNeoantigens Attribution NFT (Soulbound)
contract OpenNeoantigenAttribution is ERC721, Soulbound {

    struct Contribution {
        string mutationId;
        string contributionType;  // "sequence", "binding_data", "clinical", "curation"
        uint256 timestamp;
        string ipfsMetadataHash;
        uint256 citationCount;
    }

    mapping(uint256 => Contribution) public contributions;
    mapping(address => uint256[]) public contributorTokens;

    function mint(
        address contributor,
        string memory mutationId,
        string memory contributionType,
        string memory metadataHash
    ) external onlyValidator returns (uint256) {
        // Mint soulbound attribution NFT
        // Cannot be transferred, but proves permanent credit
    }

    function incrementCitation(uint256 tokenId) external {
        // Called when someone uses/cites this contribution
        contributions[tokenId].citationCount++;
    }
}

8. API Specification

8.1 REST API Endpoints

BASE URL: https://api.openneoantigens.org/v1

# Query neoantigens by mutation
GET /mutations/{mutation_id}
GET /mutations?gene=KRAS&change=G12D

# Get HLA binding predictions
GET /bindings?mutation=KRAS_G12D&hla=HLA-A*02:01
GET /bindings/matrix?mutation=KRAS_G12D

# Get pre-designed mRNA sequences
GET /mrna?mutation=KRAS_G12D&hla=HLA-A*02:01
GET /mrna/{design_id}

# Query by patient HLA profile
POST /match
{
  "mutations": ["KRAS_G12D", "TP53_R175H"],
  "hla_class_i": ["HLA-A*02:01", "HLA-A*03:01", "HLA-B*07:02"],
  "hla_class_ii": ["HLA-DRB1*01:01"]
}

# Get clinical evidence
GET /evidence?mutation=KRAS_G12D&hla=HLA-A*02:01
GET /evidence?trial_id=NCT04117087

# Submit new data (authenticated)
POST /contributions/binding
POST /contributions/clinical
POST /contributions/mrna

8.2 GraphQL Schema

type Query {
  mutation(id: ID!): Mutation
  mutations(gene: String, cancerType: String): [Mutation!]!

  peptide(id: ID!): Peptide
  peptides(mutationId: ID!, minLength: Int, maxLength: Int): [Peptide!]!

  binding(peptideId: ID!, hlaAllele: String!): HLABinding
  bindingMatrix(mutationId: ID!): BindingMatrix!

  matchPatient(input: PatientMatchInput!): PatientMatchResult!

  mrnaDesign(id: ID!): MRNADesign
  mrnaDesigns(mutationId: ID!, hlaAllele: String): [MRNADesign!]!

  clinicalEvidence(mutationId: ID, trialId: String): [ClinicalEvidence!]!
}

type Mutation {
  id: ID!
  gene: String!
  proteinChange: String!
  dnaChange: String!
  cosmicId: String
  somaticFrequency: [CancerFrequency!]!
  peptides: [Peptide!]!
  clinicalEvidence: [ClinicalEvidence!]!
}

type Peptide {
  id: ID!
  sequence: String!
  length: Int!
  mutantPosition: Int!
  bindings: [HLABinding!]!
  mrnaDesigns: [MRNADesign!]!
}

type HLABinding {
  id: ID!
  hlaAllele: String!
  ic50Nm: Float!
  percentileRank: Float!
  bindingLevel: BindingLevel!
  predictionTools: [PredictionResult!]!
  experimentalValidation: ExperimentalValidation
}

type MRNADesign {
  id: ID!
  fullSequence: String!
  gcContent: Float!
  predictedHalfLife: Float!
  license: String!
  attributionIpId: String!
}

input PatientMatchInput {
  mutations: [String!]!
  hlaClassI: [String!]!
  hlaClassII: [String!]
  cancerType: String
}

type PatientMatchResult {
  recommendedPeptides: [PeptideRecommendation!]!
  mrnaConstructs: [MRNADesign!]!
  clinicalEvidence: [ClinicalEvidence!]!
  coverageScore: Float!
}

8.3 SDK Examples

# Python SDK Example
from openneoantigens import OpenNeoClient

client = OpenNeoClient(api_key="your_api_key")

# Find optimal epitopes for a patient
result = client.match_patient(
    mutations=["KRAS_G12D", "TP53_R175H"],
    hla_class_i=["HLA-A*02:01", "HLA-A*11:01"],
    hla_class_ii=["HLA-DRB1*04:01"]
)

for peptide in result.recommended_peptides:
    print(f"{peptide.sequence} | {peptide.hla} | IC50: {peptide.ic50_nm}nM | Evidence: {peptide.evidence_level}")

# Output:
# KLVVVGADGV | HLA-A*02:01 | IC50: 28nM | Evidence: CLINICAL_RESPONSE
# VVGADGVGK  | HLA-A*11:01 | IC50: 38nM | Evidence: IMMUNOGENIC_HUMAN
# HMTEVVRHC  | HLA-A*02:01 | IC50: 156nM | Evidence: BINDING_VALIDATED

# Get ready-to-manufacture mRNA sequence
mrna = client.get_mrna_design(
    mutation="KRAS_G12D",
    hla="HLA-A*02:01"
)
print(f"mRNA construct: {mrna.design_id}")
print(f"Sequence length: {mrna.sequence_length} nt")
print(f"License: {mrna.license}")  # CC-BY-SA-4.0

9. Governance Model

9.1 OpenNeoantigens DAO

flowchart TB subgraph DAO["OpenNeoantigens DAO"] A[Token Holders
Governance Votes] B[Scientific Advisory
Board] C[Validation
Committee] D[Technical
Committee] end subgraph DECISIONS["Governance Scope"] E[Inclusion Criteria
Updates] F[Evidence Standards
Changes] G[API Access
Policies] H[Attribution Rules
Modifications] I[Treasury
Allocation] end A --> E A --> G A --> I B --> E B --> F C --> F D --> G D --> H

Figure 6: OpenNeoantigens DAO governance structure

9.2 Governance Token

ONEO Token Distribution

40% - Contributors: Earned through validated data contributions
25% - Foundation: Long-term development and maintenance
20% - Scientific Advisors: Vested over 4 years
10% - Early Supporters: Initial funding contributors
5% - Ecosystem Grants: Tool developers, integrators

Earning ONEO: 1 token per validated peptide-HLA submission, 10 tokens per clinical evidence contribution, 5 tokens per mRNA design, 0.5 tokens per curation review.

9.3 Scientific Advisory Board

The SAB provides scientific oversight and consists of:

3 Academic oncologists/immunologists
2 Computational biologists
1 Regulatory expert
1 Patient advocate

SAB members serve 3-year terms, renewable once. They review evidence standards, resolve disputes, and guide research priorities.

10. Integration with BioNFT Ecosystem

10.1 Hybrid Vaccine Strategy

OpenNeoantigens complements—not replaces—the patient-owned BioNFT system:

flowchart LR subgraph PATIENT["Patient with PDAC"] A[Tumor Biopsy] B[HLA Typing] end subgraph ANALYSIS["Neoantigen Analysis"] C[Shared Mutations
KRAS G12D, TP53 R175H] D[Private Mutations
Unique passengers] end A --> C A --> D B --> C B --> D subgraph OPEN["OpenNeoantigens"] E[Pre-validated
Peptides] F[Ready mRNA
Designs] end subgraph BIONFT["Patient BioNFT"] G[Private Neoantigen
Sequences] H[Custom mRNA
Designs] end C --> E E --> F D --> G G --> H F --> I[Combined
Vaccine] H --> I I --> J["Multi-epitope mRNA Vaccine
Shared (free) + Private (patient-owned)"]

Figure 7: Integration of OpenNeoantigens with patient-owned BioNFT system

10.2 API Integration

# GenoBank BioNFT API integration with OpenNeoantigens

# Patient's BioNFT contains their private neoantigens
patient_bionft = genobank.get_bionft(biosample_id="55052008714000")

# Query OpenNeoantigens for their shared mutations
shared_targets = openneo.match_patient(
    mutations=patient_bionft.shared_mutations,  # ["KRAS_G12D", "TP53_R175H"]
    hla=patient_bionft.hla_type
)

# Combine with patient's private neoantigens
private_targets = patient_bionft.private_neoantigens  # Patient-owned, consent-gated

# Generate combined vaccine design
vaccine_design = genobank.design_vaccine(
    shared_epitopes=shared_targets.peptides,      # From OpenNeoantigens (free)
    private_epitopes=private_targets.peptides,    # From BioNFT (patient-owned)
    patient_hla=patient_bionft.hla_type
)

# Attribution tracking
vaccine_design.attribution = {
    "shared_components": shared_targets.attribution_ip_ids,  # Story Protocol IPs
    "private_components": patient_bionft.ip_id,              # Patient's BioNFT IP
    "design_contributor": "GenoBank.io"
}

10.3 Economic Model

Component	Source	Cost to Patient	Attribution
Shared neoantigen sequences	OpenNeoantigens	Free	CC-BY-SA (cite database)
Pre-designed mRNA (shared)	OpenNeoantigens	Free	Story Protocol IP credit
Private neoantigen sequences	Patient BioNFT	Patient-owned	Patient retains 100% IP
Custom mRNA design	GenoBank/CDMO	Service fee	Design IP to patient
GMP Manufacturing	CDMO Partner	Manufacturing cost	Process IP to CDMO
Clinical administration	Physician	Medical fee	N/A

11. Implementation Roadmap

Phase 1: Foundation (Q1-Q2 2026)

✅ Database schema design and implementation
✅ Initial mutation set curation (50 hotspots)
⬜ HLA binding predictions for top 30 alleles
⬜ Story Protocol IP registration framework
⬜ REST API v1.0 launch
⬜ Documentation and SDK release

Phase 2: Expansion (Q3-Q4 2026)

⬜ Expand to 200+ mutations
⬜ Add fusion protein neoantigens
⬜ Add viral oncoprotein database
⬜ GraphQL API launch
⬜ Clinical evidence integration from published trials
⬜ mRNA design repository with 1000+ constructs

Phase 3: Validation (2027)

⬜ Partner with 3+ academic medical centers for validation studies
⬜ Experimental binding validation for top 500 peptide-HLA pairs
⬜ Integration with CDMO manufacturing partners
⬜ DAO governance launch
⬜ First patient treatments using OpenNeoantigens-derived vaccines

Phase 4: Scale (2028+)

⬜ 1000+ validated neoantigens
⬜ Pan-cancer coverage
⬜ Real-world response data integration
⬜ AI-driven epitope optimization
⬜ Global CDMO network for manufacturing

12. Appendix: Initial Mutation Set

12.1 Priority 1: High-Frequency Drivers (Launch Set)

#	Gene	Mutation	Cancer Types	Frequency	Existing Vaccine Data
1	KRAS	G12D	PDAC, CRC, NSCLC	13-41%	ELI-002, autogene cevumeran
2	KRAS	G12V	PDAC, CRC, NSCLC	8-30%	ELI-002
3	KRAS	G12C	NSCLC, CRC	3-13%	Multiple programs
4	KRAS	G12R	PDAC	16%	ELI-002
5	KRAS	G13D	CRC	7%	Limited
6	BRAF	V600E	Melanoma, CRC, Thyroid	40-60%	Multiple trials
7	TP53	R175H	Pan-cancer	6%	Limited
8	TP53	R248Q	Pan-cancer	4%	Limited
9	TP53	R273H	Pan-cancer	4%	Limited
10	PIK3CA	H1047R	Breast, Endometrial	15-30%	Limited
11	PIK3CA	E545K	Breast, Endometrial	10-15%	Limited
12	IDH1	R132H	Glioma, AML	70-90%	Peptide vaccine trials
13	NRAS	Q61R	Melanoma	15%	Limited
14	NRAS	Q61K	Melanoma, AML	10%	Limited
15	EGFR	L858R	NSCLC	10-15%	Limited

12.2 Priority 2: Fusion Proteins

Fusion	Cancer Type	Frequency	Junction Peptides
BCR-ABL (p210)	CML	95%	ATGFKQSSKALQRPVAS
BCR-ABL (p190)	ALL	25%	HSATGFKQSSKALQRPVAS
EML4-ALK V1	NSCLC	2%	To be characterized
EML4-ALK V3	NSCLC	2%	To be characterized
TMPRSS2-ERG	Prostate	50%	To be characterized

12.3 Priority 3: Viral Oncoproteins

Virus	Protein	Cancer Type	Known Epitopes
HPV16	E6	Cervical, Oropharyngeal	Multiple HLA-restricted epitopes characterized
HPV16	E7	Cervical, Oropharyngeal	YMLDLQPETT (HLA-A*02:01)
HPV18	E6, E7	Cervical	Multiple characterized
EBV	LMP1, LMP2	Nasopharyngeal, Hodgkin	Extensive database available

Summary

OpenNeoantigens represents a new paradigm in cancer vaccine development: a public commons for shared cancer targets that accelerates access while preserving attribution through blockchain technology. By open-sourcing the ~30% of neoantigens derived from common driver mutations, we can:

Eliminate weeks of delay for patients with targetable hotspot mutations
Pool clinical validation data across thousands of patients
Enable global access to validated vaccine designs
Complement patient-owned private neoantigens through the BioNFT system

The protocol launches with 15 high-priority mutations covering the most common oncogenic drivers, with plans to expand to 200+ mutations and comprehensive fusion/viral databases by 2027.

OpenNeoantigens + BioNFTs = Complete patient sovereignty over cancer immunotherapy.

OpenNeoantigens Protocol

A GenoBank.io Initiative

January 2026 | Version 1.0

Contact: [email protected]

genobank.io | GitHub | Discord