After Neoantigen Prediction: Choosing the Next Validation Experiment
A high prediction score does not tell you which experiment to run next. Choose the assay that addresses the missing evidence: peptide–HLA binding, natural presentation, or recognition by a compatible T cell population.
Separate Three Validation Questions
Prediction ranks hypotheses. Experimental validation asks a narrower biological question, and different assays provide different kinds of evidence. This guide focuses on choosing that next experiment rather than repeating the sequencing, high-throughput screening, or manufacturing workflow.
| Question | Possible approach | What the result does not establish |
|---|---|---|
| Does the peptide bind the proposed HLA? | A suitable peptide–HLA binding or stability assay. | Binding does not establish that a tumor processes and displays the peptide. |
| Is the candidate naturally presented? | HLA-associated peptide analysis in a relevant source, where technically feasible. | Detection does not establish that a responsive T cell population exists or will kill the tumor. |
| Can T cells recognize the candidate? | A compatible antigen-stimulation or target-recognition assay. | Peptide-pulsed recognition does not necessarily demonstrate recognition of an unmodified tumor target. |
Choose the Experiment from the Uncertainty
- Uncertain HLA compatibility: verify typing, peptide identity, and the proposed restriction before committing limited cells to a functional assay.
- Strong prediction but weak biological context: review source-gene expression, mutation confirmation, and whether the intended target can present the relevant peptide.
- Binding evidence but no recognition data: assess availability of compatible antigen-presenting cells and responder cells. A generic PBMC preparation does not guarantee candidate-specific T cells.
- Peptide-positive but tumor-negative recognition: investigate processing, presentation, HLA compatibility, and target expression before assuming inadequate vaccine potency.
A project does not always need every assay in a fixed sequence. With scarce tumor material or cells, first identify the decision the data must support and the uncertainty most likely to change that decision.
Prepare a Candidate-Level Evidence Record
For each candidate, record the exact mutant and reference sequence, HLA allele, prediction method and version, available expression evidence, and any existing binding or functional result. Keep directly measured findings separate from computational annotations.
Also identify available tumor cells, tissue, or antigen-presenting cells, together with responder-cell source and preparation history. If a candidate is tested only as a synthetic peptide, label that condition explicitly. A tumor lysate, whole protein, and RNA construct require different assumptions about antigen handling.
Interpret Negative Results Within the Tested System
Failure to detect a peptide by mass spectrometry can reflect source material, abundance, HLA coverage, or analytical sensitivity. It is not universal proof that the epitope is never presented. Likewise, a negative T cell assay can reflect responder availability, sample quality, background, or the chosen stimulation format.
Useful handover: report “not detected under these conditions” together with controls and limitations, rather than reducing every candidate to an unexplained pass/fail label. Published lung-cancer immunopeptidome work and tumor-organoid recognition studies illustrate distinct evidence layers; neither supplies a universal validation threshold.
Frequently Asked Questions
Does a high predicted binding score prove immunogenicity?
No. It is a computational estimate related to a defined model. Natural presentation and functional recognition require their own evidence.
Is peptide stimulation enough to validate a vaccine construct?
It can assess peptide-associated recognition, but may bypass processing. Construct-level questions need an appropriate expression or antigen-presentation system.
References
1. Kraemer AI, et al. The immunopeptidome landscape associated with T cell infiltration, inflammation and immune editing in lung cancer. Nature Cancer. 2023;4:608–628. doi:10.1038/s43018-023-00548-5.
2. Li Z, et al. Identification and validation of tumor-specific T cell receptors from tumor infiltrating lymphocytes using tumor organoid co-cultures. Cancer Immunology, Immunotherapy. 2024;73:164. doi:10.1007/s00262-024-03749-8.
Discuss the Missing Evidence
Alfa Cytology’s cancer antigenic epitope studies and T cell response testing can be scoped around your candidate list, HLA context, and available materials.
Discuss Your Validation Plan