OPEN-SOURCE IMMUNOTHERAPY

Personalized Cancer Vaccines: From Lab to Patient

A man built a cancer vaccine for his dog. Then he open-sourced the blueprint. Here's how personalized mRNA cancer vaccines work — from tumor biopsy to syringe — in 8 steps.

0
Pipeline Steps
~$5-15K
Outsourced Cost
~1mg
mRNA Output

A Dog, a Diagnosis, and a Lab Notebook

When a pet owner's dog was diagnosed with cancer, he didn't just accept the prognosis. He built a personalized cancer vaccine — tailored specifically to his dog's tumor mutations.

🧬

From Pet Project to Open-Source Guide

Phil Fung, a former lab startup founder, documented the entire process and published it as OpenVaxx — a free, open-source guide to producing personalized mRNA cancer vaccines. It covers every step from tumor sequencing to the final vial, using open-source software and commercially available benchtop lab equipment. The guide is released under the Apache 2.0 license.

📖 Source: OpenVaxx Interactive Guide by Phil Fung — GitHub Repository (Apache 2.0 License)

What Is a Personalized Cancer Vaccine?

Unlike traditional vaccines that prevent infection, personalized cancer vaccines are therapeutic — they train your immune system to attack cancer cells that are already in your body.

Every tumor is genetically unique. It accumulates mutations that produce abnormal proteins called neoantigens — molecular flags that don't exist in healthy tissue. A personalized cancer vaccine encodes these neoantigens into synthetic mRNA, teaching the immune system to recognize and destroy cells displaying those flags.

The process follows a logical chain:

🧬 Sequence Tumor 🔬 Find Mutations 🤖 AI Predicts Targets 💻 Design mRNA 🧪 Synthesize 💉 Inject

The remarkable part: the software to do most of the intellectual work is already open-source and freely available. The hardware is commercially available, and each step can be outsourced to contract labs. The barrier isn't knowledge — it's access, training, and regulation.

8 Steps from Tumor to Vaccine

The OpenVaxx pipeline splits into two halves: a digital pipeline that converts biological samples into a computer-designed mRNA blueprint, and a physical pipeline that turns that blueprint into a vial of vaccine.

Part 1: Digital Pipeline — Data to Blueprint
1
Genome Sequencing
Reading the Blueprint — Digitizing the Cells
Convert tumor biopsy and normal blood into raw sequencing data. The sequencer reads DNA/RNA, turning biological chemistry into digital text files (.FASTQ). Requires both tumor and healthy baseline samples.
Illumina NextSeq 2000 ~$2,500/pt outsourced .FASTQ files
2
Variant Calling
Spotting the Typos — Finding the Mutations
Compare healthy DNA against tumor DNA to isolate somatic mutations — the cancer-specific "typos" in the genetic code. This is pure software — no lab equipment needed.
GATK Mutect2 (open source) Free (compute only) .VCF files
3
Neoantigen Prediction
Picking the Targets — AI Selects the Best Mutations
Neural networks predict which mutations the immune system will recognize as foreign. Uses the patient's HLA profile to rank candidates by how well they'll trigger an immune response.
pVACseq + MHCflurry (open source) Free (compute only) Ranked neoantigen list
4
mRNA Design
Writing the New Code — Sequence Assembly
String the top neoantigen targets together, add structural instructions (5' Cap, Poly-A tail), and optimize codons for folding stability. Output is a complete, printable mRNA blueprint.
pVACvector + LinearDesign (open source) Free (compute only) .FASTA file
Part 2: Physical Pipeline — Blueprint to Vial
5
DNA Synthesis
Printing the Master Copy
Convert the digital .FASTA blueprint back into physical DNA. A benchtop synthesizer uses cell-free enzymatic assembly to build the linear DNA construct in 1-2 days.
BioXp 3250 ~$200-900 outsourced ~75µg linear DNA
6
In Vitro Transcription (IVT)
Mass Production — DNA → mRNA
Continuous-flow bioreactors read the DNA template and print the corresponding mRNA strand. Uses N1-methylpseudouridine for immune cloaking and CleanCap® for human cell recognition — the same tech behind the COVID vaccines.
NTxscribe System ~$1-3K/mg outsourced ~1mg purified mRNA
7
LNP Encapsulation
Packaging for Delivery — Lipid Nanoparticles
Wrap the fragile mRNA in a protective lipid nanoparticle shell for cell entry. Microfluidic collisions force mRNA and lipids to self-assemble into 60-100nm nanoparticles. Over 90% encapsulation efficiency.
NanoAssemblr Ignite ~$2-5K/batch outsourced ~0.9mg encapsulated mRNA
8
Quality Control + Filling
Validation and Final Vials
Dynamic Light Scattering verifies particle size (60-100nm). Tangential Flow Filtration washes out toxic ethanol. Final product: ~10 sterile glass vials at 100µg/mL, stored at -80°C.
Stunner DLS + TFF ~$1-3K/batch outsourced ~10 doses @ 100µg/mL

The Digital Pipeline (Steps 1-4)

The upstream pipeline is almost entirely software. Once you have sequencing data, the rest runs on open-source tools on a standard workstation. This is where the democratization is happening.

🧬

Step 1: Genome Sequencing

Converting biology to data

The sequencer reads three samples: normal blood DNA (baseline, 30-50X depth), tumor DNA (deep coverage at 100-500X to catch rare mutations), and tumor RNA (50-100M reads to verify mutations are actually expressed). A separate analysis derives the patient's HLA profile — their immune system's "lock configuration."

Equipment
Illumina NextSeq 2000 (~$300K)
Outsource Cost
~$2,500/patient
Inputs
Tumor biopsy + normal blood
Outputs
3× .FASTQ files + HLA profile
@Patient_001:Baseline_Normal:1:1101:1234:5678 GATTTGGGGTTCAAAGCAGTATCGATCAAATAGTAAATCC + !''*((((***+))%%%++)(%%%%).1***-+*''))**
🔍

Step 2: Variant Calling

Subtracting healthy from tumor

GATK Mutect2 aligns sequencing reads to the human reference genome, then mathematically subtracts normal DNA from tumor DNA to isolate somatic mutations — changes that exist only in the cancer. The output is a filtered list of high-confidence, tumor-only mutations.

Software
GATK Mutect2 (free, open source)
Cost
$0 (compute only)
Inputs
Normal + tumor .FASTQ + reference genome
Outputs
Filtered somatic variants (.VCF)
#CHROM POS ID REF ALT QUAL FILTER INFO chr7 14045313 Mut_01 A T . PASS SOMATIC;DP=152;AF=0.24
🤖

Step 3: Neoantigen Prediction

AI picks the immune targets

pVACseq feeds mutations through MHCflurry neural networks to predict which peptide fragments will bind most strongly to the patient's specific HLA receptors. Stronger binding = more likely the immune system will recognize and attack cells with that mutation. RNA expression data filters out mutations that aren't actually active.

Software
pVACseq + MHCflurry (free)
Cost
$0 (compute only)
Inputs
Filtered .VCF + HLA profile + RNA data
Outputs
Ranked neoantigen predictions (.TSV)
HLA_Allele Peptide IC50_Score MHCflurry_EL HLA-A*02:01 YLLPAIVHI 24.5 0.98 HLA-A*02:01 LLDVPTAAV 45.2 0.92 HLA-B*07:02 APRGVFLLS 112.4 0.85
💻

Step 4: mRNA Design

Compiling the vaccine blueprint

pVACvector strings the top neoantigen targets together with linker sequences, adds structural components (5' UTR, Kozak sequence, start codon, stop codon, 3' UTR, poly-A tail), then LinearDesign optimizes codon usage for maximum mRNA stability and translational efficiency.

Software
pVACvector + LinearDesign (free)
Cost
$0 (compute only)
Inputs
Ranked neoantigen predictions
Outputs
Complete mRNA blueprint (.FASTA)
>Patient_001_Custom_Vaccine_v1 | 5'UTR-Kozak-AUG-Epitopes-Linkers-Stop-3'UTR-PolyA GGGAAAUAAGAGAGAAAAGAAGAGUAAGAAGAAAUAUAAGAGCCACCAUGGCUACUUGCUGCCAGCGAU UGUCCAUAUCCUCCUCUUCUUGGGCAAAAUUUGGCCGCUGCUUAUAAAAAAAAAAAAAAAAAAAAAA

The Physical Pipeline (Steps 5-8)

This is where digital meets physical. The downstream pipeline requires specialized lab equipment, biosafety protocols, and quality control — the expensive half. Each step can be outsourced to contract research organizations (CROs).

🖨️

Step 5: DNA Synthesis

Printing the physical master copy

A benchtop DNA synthesizer converts the digital .FASTA blueprint back into physical DNA. Two routes are available: cell-free linear synthesis (faster, ~1-2 days) or traditional plasmid-based cloning via E. coli (24-48 hours additional). The output is purified, linearized DNA template ready for transcription.

Equipment
BioXp 3250 (~$100K)
Outsource Cost
~$200-900/gene
Output
~75µg purified linear DNA
Outsource Options
Twist Bioscience, IDT, GenScript
⚗️

Step 6: In Vitro Transcription (IVT)

DNA → functional mRNA

Continuous-flow IVT bioreactors read the DNA template and print the corresponding mRNA strand. Key modifications: N1-methylpseudouridine replaces standard uridine (the same immune-cloaking trick used in Pfizer/Moderna COVID vaccines), and CleanCap® AG adds a 5' cap for human cell recognition. Post-transcription cleanup includes DNase I digestion and mRNA purification.

Equipment
NTxscribe System (~$250K)
Outsource Cost
~$1,000-3,000/mg
Output
~1.0mg purified mRNA
Key Reagents
T7 RNA Pol, m1Ψ, CleanCap®
💊

Step 7: LNP Encapsulation

Wrapping mRNA for cell delivery

Naked mRNA is fragile and can't enter cells on its own. Microfluidic mixing forces negatively charged mRNA and positively charged lipids into self-assembling nanoparticles (60-100nm). The four-lipid cocktail includes an ionizable lipid (e.g., ALC-0315), PEG-lipid, DSPC helper lipid, and cholesterol — the same LNP formula proven in billions of COVID vaccine doses.

Equipment
NanoAssemblr Ignite (~$150K)
Outsource Cost
~$2,000-5,000/batch
Efficiency
>90% encapsulation
Outsource Options
VectorBuilder, Lonza, Vernal Bio

Step 8: Quality Control + Filling

Verification and final vials

Dynamic Light Scattering (DLS) measures particle diameter to verify the nanoparticles are in the therapeutic 60-100nm range. Tangential Flow Filtration (TFF) removes the toxic ethanol used during mixing and buffer-exchanges into a cryoprotectant. RiboGreen assay quantifies encapsulation efficiency. Final output: ~10 sterile glass vials at 100µg/mL, stored at -80°C.

Equipment
Stunner DLS + TFF (~$80K)
Outsource Cost
~$1,000-3,000/batch
Output
~10 vials @ 100µg/mL
Storage
-80°C in Tris-Sucrose buffer

Open-Source Tools That Make This Possible

The digital half of the pipeline runs entirely on free, open-source software maintained by major research institutions. These tools are used in real clinical programs — they aren't experimental toys.

Cost Breakdown: In-House vs. Outsourced

Building a full in-house lab costs roughly $800K-$1M in equipment — but that enables production for many patients. The outsourced route costs approximately $5,000-$15,000 per patient.

Step Equipment (In-House) Per-Patient (Outsourced)
1. Genome Sequencing ~$300K (Illumina NextSeq) ~$2,500
2. Variant Calling Standard workstation $0 (open source)
3. Neoantigen Prediction Standard workstation $0 (open source)
4. mRNA Design Standard workstation $0 (open source)
5. DNA Synthesis ~$100K (BioXp 3250) ~$200-900
6. IVT (mRNA Synthesis) ~$250K (NTxscribe) ~$1,000-3,000
7. LNP Encapsulation ~$150K (NanoAssemblr) ~$2,000-5,000
8. QC + Filling ~$80K (Stunner + TFF) ~$1,000-3,000
Total ~$880K ~$6,700-$14,400

* Outsourced costs are estimates based on publicly available pricing from CROs and academic facilities. Actual costs vary by provider, volume, and complexity. Steps 2-4 cost nothing because the software is open-source.

Cost Distribution (Outsourced Route)

Genome Sequencing ~$2,500 (37%)
DNA Synthesis ~$550 (8%)
IVT (mRNA Synthesis) ~$2,000 (30%)
LNP Encapsulation ~$3,500 (17%)
QC + Filling ~$2,000 (8%)

How This Relates to Peptides & Immunotherapy

If you've been following the peptide research space, personalized cancer vaccines are a natural extension of the same principles. Here's the connection:

Neoantigens are peptides. The targets identified in Step 3 are short peptide sequences (typically 8-11 amino acids for MHC Class I) that the immune system can recognize. The entire field of neoantigen prediction is fundamentally peptide science — predicting which peptide fragments will bind to immune receptors.

Thymosin Alpha-1 and immunotherapy. Peptides like Thymosin Alpha-1 (Tα1) are already used as immune modulators in cancer therapy — they enhance T-cell function, the same immune cells that personalized vaccines aim to activate. The vaccine provides the target; immunomodulatory peptides ensure the immune system is equipped to respond.

BPC-157 and recovery. Cancer treatments are brutal on the body. Peptides like BPC-157 and TB-500 are studied for their tissue-healing and anti-inflammatory properties — relevant for patients recovering from biopsies, surgeries, and the immune responses triggered by cancer vaccines.

The Regulatory Landscape

Having the technical capability to produce a personalized cancer vaccine is one thing. Having the legal authority to administer it is another entirely.

✅ What's Real

  • The software pipeline works and is used in clinical research programs worldwide
  • Multiple companies have active Phase II/III clinical trials with personalized mRNA cancer vaccines
  • BioNTech's autogene cevumeran showed promising results in pancreatic cancer (Phase I)
  • Moderna's mRNA-4157 (V940) + Keytruda showed 44% reduction in recurrence for melanoma (Phase IIb)
  • All equipment is commercially available and legally purchasable

⚠️ What's Missing

  • No personalized mRNA cancer vaccine has FDA approval yet (as of early 2026)
  • Administering an unapproved biologic is illegal in most jurisdictions
  • GMP (Good Manufacturing Practice) compliance requires certified facilities
  • Quality control for injectables has standards far beyond what a private lab can typically meet
  • Endotoxin testing, sterility assurance, and potency verification require specialized equipment and protocols

The gap between "technically possible" and "legally/safely administrable" is significant. This guide exists to explain the science and engineering — not to provide a recipe for DIY medicine.

Companies Building Personalized Cancer Vaccines

Several major biotech companies are racing to bring individualized mRNA cancer vaccines to market. These programs use the same fundamental pipeline described above, at pharmaceutical scale with GMP compliance.

BioNTech
Autogene cevumeran (BNT122) — Phase II
The COVID vaccine maker is applying its mRNA platform to personalized cancer vaccines. Their individualized neoantigen-specific immunotherapy (iNeST) targets up to 20 neoantigens per patient. Phase I results in pancreatic cancer were promising — T-cell responses in 8 of 16 patients correlated with no recurrence.
Moderna
mRNA-4157 (V940) + Keytruda — Phase III
Moderna's personalized cancer vaccine, combined with Merck's checkpoint inhibitor Keytruda, showed a 44% reduction in recurrence or death in resected high-risk melanoma (Phase IIb KEYNOTE-942). Now in Phase III trials — the furthest any personalized mRNA cancer vaccine has progressed.
Gritstone bio
GRANITE — Phase II/III
Uses a self-amplifying mRNA (samRNA) approach combined with adenoviral prime for personalized cancer vaccination. Targeting microsatellite-stable colorectal cancer and other solid tumors. The samRNA technology could enable lower doses.
Transgene / NEC
TG4050 — Phase I/II
Uses a viral vector (MVA) rather than mRNA, combined with NEC's AI-powered neoantigen prediction. Currently in Phase I/II for head and neck cancer and ovarian cancer. Different delivery vehicle, same fundamental concept.

Educational resources and lab safety equipment related to immunotherapy and mRNA science.

Affiliate links help support HighPeptides at no extra cost to you.

Explore More on HighPeptides

Frequently Asked Questions

What is a personalized mRNA cancer vaccine?
A personalized mRNA cancer vaccine is a custom immunotherapy designed for a single patient. It works by sequencing the patient's tumor to identify unique mutations (neoantigens), then encoding those targets into synthetic mRNA that trains the immune system to attack cancer cells specifically. Unlike off-the-shelf drugs, each vaccine is unique to that patient's tumor.
How much does it cost to make a personalized cancer vaccine?
Using outsourced services for each step, the estimated total cost is approximately $5,000-$15,000 per patient. Building a fully in-house lab with all equipment costs approximately $800K-$1M upfront, but this enables producing vaccines for many patients at lower per-unit costs. The software pipeline (Steps 2-4) is entirely free — all tools are open-source.
Can you make an mRNA vaccine at home?
No. While the computational pipeline uses open-source tools that can run on a standard computer, the physical production (Steps 5-8) requires specialized laboratory equipment, BSL-2+ facilities, biosafety training, and strict quality control. This guide is for educational purposes only — actual production requires qualified personnel, proper facilities, and regulatory oversight.
What open-source tools are used to design mRNA vaccines?
The key open-source tools are: GATK Mutect2 (Broad Institute) for variant calling, pVACseq and MHCflurry for neoantigen prediction, pVACvector for vaccine sequence assembly, and LinearDesign for mRNA optimization. All are freely available on GitHub and actively maintained by major research institutions.
How long does it take to produce a personalized mRNA vaccine?
The digital pipeline (sequencing through mRNA design) takes approximately 1-2 weeks. The physical pipeline (DNA synthesis through quality control) takes another 1-2 weeks. Total turnaround is roughly 2-4 weeks, though commercial programs like BioNTech report 4-6 week timelines including additional quality assurance steps.
Is the OpenVaxx guide safe to follow?
The OpenVaxx guide is strictly for research and educational purposes. mRNA vaccine production involves severe biological hazards — working with human biological samples, ethanol, and injectable products. It requires BSL-2+ laboratory facilities, trained personnel, IRB approval for human research, and regulatory oversight. Do not attempt any part of this workflow without proper qualifications and institutional backing.

⚠️ Critical Disclaimer — Read Carefully

This page is for educational and informational purposes only. It is not medical advice, and it is not a guide for producing vaccines or any other biologic product for human or animal use.

Severe biological hazards: Working with human biological samples, recombinant DNA, mRNA, lipid nanoparticles, and injectable products carries serious risks including infection, contamination, endotoxin exposure, and allergic reactions. These processes require BSL-2+ laboratory facilities, trained personnel, and institutional oversight.

Legal restrictions: Manufacturing and administering unapproved biological products is illegal in most jurisdictions. Clinical use of personalized cancer vaccines requires Investigational New Drug (IND) applications, Institutional Review Board (IRB) approval, and compliance with Good Manufacturing Practice (GMP) regulations.

Not a recipe: This article describes the scientific principles and engineering steps involved in personalized mRNA vaccine production. It deliberately omits critical details related to sterility assurance, endotoxin testing, potency assays, and GMP compliance that are required for any product intended for administration.

Consult professionals: If you or someone you know is interested in personalized cancer vaccine therapy, consult with qualified oncologists at major cancer centers where clinical trials are available. Do not attempt DIY production of any injectable biologic product.

Source material: Technical details in this article are based on the OpenVaxx guide by Phil Fung (Apache 2.0 License), combined with publicly available information from equipment manufacturers and research publications. HighPeptides is not affiliated with OpenVaxx.