Tumor Neoantigen mRNA Vaccine
✓ 已验证8-step end-to-end · From VCF/BAM to mRNA sequence design · Education · Service · Technology
The most mature and competitive pipeline in DiVo Gen²AI's health system. 100% coverage of all dry-lab steps for tumor neoantigen vaccines. This page serves three audiences: patients and public wanting to understand neoantigen vaccines, pharma teams seeking computational service partners, and investors and peers evaluating technical capabilities.
Read "What is a Neoantigen Vaccine", "5-Step Workflow", "Milestones" below--change your understanding of immunotherapy.
Focus on "DiVo's Role", "8-Step Pipeline", "Verified Capabilities", "Sample Reports"--we have deliverables.
Focus on "Three Differentiators", "Benchmarks", "TESLA Baseline", "Glossary"--evaluate technical barriers.
To Patients and Families
Our tumor neoantigen dry-computing pipeline service is a business that works in coordination with medical institutions, not a service directly for individual users and patients. Please contact us through your treating physician at the medical institution where you are being treated. We will not bypass the treating physician to provide any medical opinions about the patient's tumor symptoms and therapies to patients and families. Scientific data and reports should also be obtained and interpreted through your treating physician.
Have a Whole Genome Sequencing Report (VCF)?
If you have your whole genome sequencing report VCF document, we can provide you with non-medical scientific interpretation and analysis reports. Please refer to our general service--Genome Interpretation Report.
What is a Tumor Neoantigen mRNA Vaccine
Every tumor cell carries gene mutations that normal cells don't have. Some of these mutations produce new protein fragments--neoantigens (Neoantigen). They are "fingerprints" left on the surface of tumor cells. The immune system can theoretically identify and eliminate tumors based on these, but in reality the signal is often too weak and escape mechanisms too many, so immune responses fail to initiate.
Neoantigen mRNA vaccines work by encoding patient-specific tumor mutation fragments into mRNA. After being delivered into the body, human cells translate these antigens, actively activating specific T cell responses--like giving the immune system a "wanted poster" to precisely hunt down tumor cells carrying these fingerprints.
This is a completely personalized treatment: each patient's tumor mutation profile is different, so the corresponding vaccine sequence is also different. This is why computational prediction is the lifeline of the entire pipeline--a single tumor sample may produce thousands of candidate peptides, and it's impossible to experimentally validate each one. Algorithms must precisely filter out the few most likely to trigger immune responses.
Only exist in tumor cells, not in normal tissue. Theoretically can be precisely recognized by the immune system without damaging healthy tissue.
Fast design, short production cycle, no cell culture needed. Naturally fits the "one person, one drug" personalized scenario; sequence is the drug.
Thousands of candidate peptides cannot be individually experimentally validated. Algorithm filtering precision directly determines whether the vaccine can truly activate T cells.
5 Steps of mRNA Vaccines
From tumor sampling to immune activation, a complete personalized vaccine workflow
Sample
Obtain patient tumor tissue samples and normal tissue samples, perform whole exome sequencing (WES)
Identify Targets
Find tumor-specific mutations from sequencing data, screen for neoantigen peptides most likely to activate T cells through MHC binding prediction
Encode mRNA
Encode screened neoantigen peptide sequences into mRNA, optimize codons and UTR sequences for efficient translation in humans
Deliver
Encapsulate mRNA in lipid nanoparticles (LNP) or dendritic cells (DC), inject into patient
Immune Activation
Human cells translate mRNA to produce neoantigen proteins, MHC molecules present them to T cells, initiating specific anti-tumor immune responses
The entire process takes about 4-8 weeks (mainly sequencing + computation + custom mRNA synthesis). Step 2 "Identify Targets" and Step 3 "Encode mRNA"--from sequencing data to mRNA sequence design--are the complete coverage scope of DiVo Gen²AI's computational service.
Neoantigen mRNA Vaccine Development Milestones
From 2017 concept validation to 2025 industrialization acceleration
Sahin team (BioNTech) Nature first reported individualized neoantigen mRNA vaccine clinical results, melanoma patients developed neoantigen-specific T cell responses
Concept validation
Wells team Cell published TESLA benchmark, establishing neoantigen prediction gold standard (608 peptide-MHC)
AUROC benchmark
Moderna mRNA-4157/V940 + Keytruda Phase II significantly prolonged recurrence-free survival, FDA granted breakthrough therapy designation
Combined immune checkpoint
Likon Life LK101 injection received NMPA clinical approval, first domestic AI+personalized neoantigen mRNA vaccine
China first
BioNTech Autogene cevumeran Phase III advancing; Moderna personalized neoantigen vaccine pipeline expanding to multiple cancer types
Industrialization acceleration
DiVo Gen²AI's Role
In the neoantigen mRNA vaccine workflow, we handle all dry-lab steps of "target identification + mRNA design"
In the neoantigen mRNA vaccine workflow, "finding the right targets" and "designing optimal mRNA sequences" are the steps most dependent on computational prediction--impossible to experimentally validate each candidate peptide, algorithms must precisely filter.
- ▸HLA Typing--determine patient MHC type, wrong typing means everything downstream is wrong
- ▸Variant Detection & Peptide Generation--extract tumor-specific mutations and candidate peptides from VCF/BAM
- ▸MHC Binding Prediction + pMHC Structure Validation--from IC50 to atomic-level 3D structure confirmation
- ▸5-Dimensional Immunogenicity Scoring--not just affinity, 5 dimensions cross-filtering
- ▸mRNA Sequence Design + TCR Recognition Validation--from amino acid sequence to deliverable mRNA
We do not produce vaccine entities. We deliver sequence design proposals that can directly enter wet-lab synthesis and clinical filing.
8-Step End-to-End Pipeline
100% coverage of all dry-lab steps · Click steps to view underlying validation
HLA Typing
✓ 已验证OptiType + Polysolver dual-tool cross-validation
4-digit resolution HLA-I typing results
Variant Detection & Annotation
✓ 已验证GATK Mutect2 + VEP + ANNOVAR
Somatic variant list + functional annotation
Peptide Generation
✓ 已验证pVACseq + pVACfuse + ScanNeo2 + NeoGuider
Candidate neoantigen peptides
MHC Binding Prediction
✓ 已验证MHCflurry(14,883 alleles) + MHCnuggets + IEDB API
IC50 min 10.1 nM, 22 high-affinity candidates
pMHC 3D Structure Prediction
✓ 已验证DiVoFold5/Protenix
Atomic-level pMHC structure, pLDDT=95.4, ipTM=0.977
5-Dimensional Immunogenicity Scoring
✓ 已验证MHC affinity 30%+presentation 20%+processing 10%+known immunogenicity 15%+structure score 25%
45 candidates Tier-graded, 17 Tier-1
mRNA Sequence Design
✓ 已验证GEMORNA + RNALens self-tuned + DNAChisel + ViennaRNA
CAI 0.7->0.95, MRL Spearman=0.92
TCR Recognition Validation
✓ 已验证DeepTCR(Recon Acc=0.972) + ProTCR
TCR recognition feasibility assessment
Three Key Differentiators
Capabilities that traditional pipelines lack
pMHC 3D Structure Validation
Traditional pipelines only use IC50 values for binding. DiVo builds atomic-level pMHC 3D structures for each candidate, pLDDT/ipTM validates spatial conformation credibility. 400+ complexes validated at scale.
5-Dimensional Immunogenicity Scoring
Traditional pipelines only use MHC affinity for screening. DiVo adds 5 scoring dimensions--affinity 30%+presentation 20%+processing 10%+known immunogenicity 15%+structure score 25%, higher screening precision.
RNALens Self-Tuned Model
Ribosome loading efficiency prediction for mRNA sequence design--Spearman=0.92, R²=0.87--based on DNABERT-Z 117M three-stage fine-tuning. Not a third-party API call, but part of 560K lines of self-developed code.
Verifiable Engineering Foundation
| Metric | Value | Note |
|---|---|---|
| TESLA benchmark AUROC | 0.698 | Wells Cell 2020, 608 peptide-MHC |
| MHC-I allele coverage | 65 | IC50 measured min 10.1 nM |
| MHC-II allele coverage | 25 | 13 DR + 5 DP + 7 DQ |
| pMHC structure pLDDT | 95.4 | Approaching X-ray crystallography precision |
| pMHC ipTM | 0.977 | Complex overall confidence |
| RNALens MRL Spearman | 0.92 | R²=0.87, self-tuned |
| mRNA CAI improvement | 0.7->0.95 | Codon adaptation index optimization |
| Large-scale validation | 400+ | Complex batch prediction validation |
| Candidate Tier-1 count | 17 | Highest tier among 45 candidates |
Top 5 Tier-1 Candidate Neoantigens
Actual validation data, not simulated values
| Peptide | HLA | Antigen Source | IC50 (nM) | Score | Tier |
|---|---|---|---|---|---|
| LLFGYPVYV | HLA-A*02:01 | HTLV-1 Tax | 10.1 | 100 | Tier-1 |
| KVAELVHFL | HLA-A*02:01 | MAGE-A1 | 14.4 | 100 | Tier-1 |
| RMFPNAPYL | HLA-A*02:01 | WT1 | 14.3 | 100 | Tier-1 |
| FLWGPRALV | HLA-A*02:01 | WT1 | 13.6 | 93 | Tier-1 |
| IMDQVPFSV | HLA-A*02:01 | NY-ESO-1 | 15.9 | 93 | Tier-1 |
Verified Technical Capabilities
Each underlying pipeline has independent validation data + sample reports
HLA Typing (Dual-Tool Cross-Validation)
✓ 已验证· Has sample reportOptiType + Polysolver dual-tool, 4-digit resolution HLA-I typing. First step of neoantigen pipeline--wrong typing means everything downstream is wrong.
pMHC 3D Structure Prediction + Docking Validation
✓ 已验证· Has sample reportProtenix + Protenix-Dock, pLDDT=95.4, ipTM=0.977, 400+ complexes validated. From IC50 values to atomic-level 3D structures.
mRNA Sequence Optimization (RNALens)
✓ 已验证· Has sample reportGEMORNA + RNALens self-tuned, Spearman=0.92, CAI 0.7->0.95. 560K lines of self-developed code, not third-party API.
Genome Interpretation Report
✓ 已验证· Has sample report10-step end-to-end, 3 genomic FM cross-validation, HCC1395 WGS verified. Upstream sequencing data for neoantigen pipeline.
CAR-T Cell Therapy Computing
✓ 已验证Downstream extension of neoantigen pipeline--target recommendation + de novo Binder design + structure validation + mRNA optimization, 5-step pipeline.
Honest Boundaries
What we can and cannot do, clearly stated
What We Can Do
What We Don't Do
Sample Reports Available
Complete 8-step pipeline simulation report, including HLA typing, MHC binding prediction, pMHC structure, immunogenicity scoring, mRNA sequence design full workflow output. 14-page PDF + MD bilingual version.
Industry Market Price Reference
Compiled from public information · Not DiVo pricing · For reference only
Boundary Note: The following are market public reference prices for complete courses of tumor neoantigen mRNA vaccines, covering tumor sequencing, neoantigen screening, mRNA synthesis, delivery systems, personalized production, and all hospital-side costs. DiVo Gen²AI handles the dry-lab computational steps (i.e., the 8-step pipeline on this page), and is not the producer or seller of vaccine entities.
Cost Composition and Factors
The above are reference data from public reports. Actual costs require consulting specific hospitals/companies and assessing patient indications, safety, and efficacy. Such treatments are mostly self-paid. Consider combining medical advice, clinical trial enrollment (may reduce costs), or financial capacity.
Glossary
10 most common terms in neoantigen mRNA vaccines
| Abbr. | Full Name | Translation | Explanation |
|---|---|---|---|
| Neoantigen | Neoantigen | Neoantigen | Antigen peptides produced by tumor-specific mutations, not expressed in normal tissue, targets for personalized vaccines |
| MHC | Major Histocompatibility Complex | Major Histocompatibility Complex | The "display board" on cell surfaces, presenting intracellular protein fragments to T cells |
| HLA | Human Leukocyte Antigen | Human Leukocyte Antigen | The name for MHC in humans, determining which peptides can be presented to T cells |
| IC50 | Half Maximal Inhibitory Concentration | Half Maximal Inhibitory Concentration | Quantitative metric of MHC-peptide binding affinity, lower is stronger (<50 nM is strong binding) |
| pMHC | peptide-MHC complex | peptide-MHC Complex | Antigen peptide complex presented by MHC molecules, the target of T cell recognition |
| pLDDT | predicted Local Distance Difference Test | predicted Local Distance Difference Test | Protein structure prediction confidence, >90 is high confidence, >70 is usable |
| ipTM | interface predicted TM-score | interface predicted TM-score | Protein complex interface interaction confidence, >0.75 is high confidence |
| CAI | Codon Adaptation Index | Codon Adaptation Index | mRNA sequence translation efficiency metric, closer to 1 is more efficient |
| MRL | Mean Ribosome Load | Mean Ribosome Load | Direct measure of mRNA translation efficiency, higher means more protein expression |
| LNP | Lipid Nanoparticle | Lipid Nanoparticle | mRNA delivery vehicle, protecting mRNA and releasing it after entering human cells for translation |
CAPACITY TRACE
能力回溯
这项服务由哪些能力支撑——从硅片到你的场景
硅片(L1) → 模型(L2) → Agent(L3) → 管线(L4) → 你的场景