Neoantigen × AI
Daily research signal
Glossary
New to the field? Start with the field primer, or look up any term below. These definitions are hand-curated and evergreen — unlike the daily brief, they don't change from day to day.

Allele

One of the alternative versions of a gene a person can carry — the unit of human-to-human genetic variation.

Antigen

Any molecule the immune system can recognize and react to — the general category of which a neoantigen is the tumor-specific, mutation-derived special case.

Antigen presentation

The cellular process of chopping up proteins and displaying their fragments on HLA so T cells can inspect them.

Checkpoint inhibitor

A drug that releases the brakes on T cells; often combined with neoantigen vaccines so the activated T cells can do their job.

Clinical trial phases (I / II / III)

The staged path a vaccine takes from first-in-human safety to large efficacy trials — the timeline that gates whether a method ever reaches patients.

Eluted-ligand vs binding-affinity prediction

Two ways to train HLA predictors — on mass-spec peptides actually eluted from HLA (EL, capturing the whole presentation pathway) versus on lab-measured binding strengths (BA) — with EL generally the better signal for what gets presented.

Epitope

The specific small piece of a molecule that an immune receptor actually binds — the precise “handle” the immune system grabs.

Epitope / binding prediction

Using algorithms to forecast which tumor peptides will be presented on HLA and recognized by T cells — the AI core of neoantigen discovery.

FDA expedited designations

FDA programs — Breakthrough Therapy, RMAT, Fast Track, accelerated approval — that speed a promising drug's development. They accelerate review and interaction; none is an approval.

Frameshift neoantigen

A neoantigen produced when an insertion or deletion shifts a gene's reading frame, generating a long stretch of entirely novel, highly foreign protein sequence.

Gene expression (RNA-seq)

How much a gene is actually transcribed into RNA — a mutation can't become a neoantigen if the gene it sits in is silent.

HLA / MHC

The cell-surface molecules that display peptide fragments to T cells — the body's “showcase” for what's inside each cell. Highly variable between people.

HLA typing

Determining a patient's specific HLA alleles — a required first step, since HLA dictates which neoantigens can be presented.

HPV vaccine

A preventive vaccine (e.g. Gardasil) that blocks cancer-causing HPV infection before it starts — the opposite end of the spectrum from a personalized neoantigen vaccine, and the bulk of today's cancer-vaccine market.

Immune tolerance (self vs non-self)

The mechanisms that teach the immune system to ignore the body's own molecules — and the reason neoantigens are attractive while autoimmunity is the risk.

Immunogenicity

How strongly something provokes an immune response — the property that separates a neoantigen that works from one that's merely present.

Immunopeptidome

The complete set of peptides actually displayed on a cell's HLA molecules — measured directly by mass spectrometry.

MHC class I vs class II

Two display systems: class I shows peptides to killer (CD8+) T cells; class II shows them to helper (CD4+) T cells.

Molecular residual disease (MRD)

Cancer that remains after treatment at a level too low to see on a scan, detected by sequencing tumor DNA fragments shed into the blood (ctDNA).

mRNA vaccine platform

A vaccine that delivers genetic instructions (mRNA) telling the patient's own cells to make the target neoantigens.

N-of-1

Shorthand for a medicine individualized to a single patient. In cancer vaccines it describes an N-of-1 product tested in an ordinary cohort trial — not a single-patient trial design.

Neoantigen

A protein fragment unique to a tumor — created by the cancer's own mutations — that the immune system can recognize as foreign.

NetMHCpan

The canonical neural-net predictor of peptide–HLA binding — the tool you'll see cited as the baseline almost everywhere in this field.

Pan-allele model

An HLA predictor that generalizes across alleles — including rare ones never measured — by learning from the HLA molecule's own sequence rather than training a separate model per allele.

Pan-RAS inhibitor

A drug that blocks multiple mutant forms of the RAS oncoprotein at once — relevant to neoantigen work because the same KRAS mutations are also shared-neoantigen vaccine targets.

Peptide-MHC (pMHC)

A single displayed peptide locked into an HLA molecule — the actual unit a T cell sees.

Personalized cancer vaccine

A vaccine custom-built for one patient from their own tumor's mutations — the central product of this field.

Platform regulation

Regulating an individualized therapy by validating its fixed manufacturing process and platform, rather than the final molecule — which differs for every patient.

Proteasome / proteasomal cleavage

The cellular machine that shreds proteins into peptides — its cut sites decide which fragments are even available to be presented on MHC class I.

Protein language model (PLM)

A transformer pretrained on millions of protein sequences (e.g. ESM) whose learned embeddings — fine-tuned for immunology tasks — are increasingly the backbone of neoantigen prediction.

Protein structure prediction (AlphaFold)

Predicting a protein's 3-D shape from sequence (AlphaFold being the breakthrough) — applied here to model pMHC complexes and, hardest of all, TCR recognition.

Shared (public) neoantigen

A neoantigen produced by a mutation that recurs across many patients' tumors — so a single off-the-shelf product can target it, instead of a vaccine built per person.

Somatic vs germline mutation

Somatic mutations arise in body cells during life (and drive cancer); germline mutations are inherited and present in every cell.

T cells (CD8+ / CD4+)

Immune cells that recognize displayed peptides: CD8+ “killer” cells destroy target cells; CD4+ “helper” cells coordinate the response.

T-cell receptor (TCR)

The receptor on a T cell that recognizes a specific peptide-MHC complex — the immune system's targeting system.

TCR-T cell therapy

A cell therapy that engineers a patient's T cells to express a chosen T-cell receptor — including ones that recognize neoantigens — then infuses them to attack the tumor.

TCR–pMHC prediction

Predicting whether a given T-cell receptor will recognize a specific peptide–HLA complex — the step that decides whether a presented neoantigen actually triggers a T-cell response.

Tumor heterogeneity (clonal vs subclonal)

A tumor is a mosaic of genetically distinct cell populations; mutations present in every cell (clonal) make far better vaccine targets than those in only a subset (subclonal).

Tumor neoantigen burden (TNB)

The count of predicted or experimentally presented neoantigens in a tumor — a more biology-aware cousin of tumor mutational burden that estimates how many mutations actually yield immune-visible targets.

Variant calling

The computational step that turns raw sequencing reads into a confident list of a tumor's mutations.