We use optional Google Analytics to understand how visitors use this beta platform and to improve the service. Analytics will only be activated if you accept.
Prioritize peptide candidates by predicted binding to selected human HLA or mouse MHC targets.
Compare percentile rank and predicted affinity, then use the shortlist for downstream presentation,
immunogenicity or experimental validation studies.
Human & mouseClass I & Class IIPercentile-rank prioritizationResearch use
Class I and Class II are predicted separately because their peptide-binding grooves and antigen-presentation pathways differ.
Prediction workspace
Rank peptide candidates
Predict MHC / HLA Binding
Enter a peptide or protein sequence, select species and MHC class, choose the allele or strain,
then return the top-ranked candidate peptides.
A longer protein sequence can be scanned to identify candidate binding peptides.
Select the biological system.
Class I and Class II are analyzed separately.
Choose the available HLA allele or mouse MHC target.
Number of ranked candidates to return.
How to read the result. Lower percentile rank indicates stronger predicted binding relative
to the model's reference distribution for the selected target. Predicted binding is a prioritization
signal, not proof that the peptide is naturally presented or recognized by T cells.
Enter a sequence and select species, MHC class, and allele/strain to begin.
MHC-I Antigen Processing
Estimate whether Class I peptide candidates are likely to be generated by proteasomal cleavage, transported by TAP, and presented through MHC-I. This analysis uses the same sequence and selected Class I allele above.
Use immunoproteasome when the biological context is uncertain.
What this adds. Processing combines proteasomal cleavage and TAP transport; the total score additionally incorporates MHC-I binding. Higher processing/total scores indicate higher predicted efficiency within this model. This is not proof of natural presentation.
Select Class I and an allele above, then run processing prediction.
pMHC immunogenicity
Class I prioritization
MHC-I pMHC Immunogenicity
Prioritize Class-I peptide-MHC candidates using the IEDB immunogenicity model. The model uses amino-acid properties and peptide position and is best validated for 9-mer peptides.
IEDB notes validation primarily on 9-mers.
Uses the Class-I allele selected in the binding workspace above.
Interpretation. Higher scores prioritize candidates within this model. The score is not a probability and does not prove T-cell recognition.
Select a Class-I target above and run immunogenicity prediction.
WT vs Mutant
Neoepitope comparison
WT vs Mutant / Neoepitope Comparison
Compare substitution-style WT and mutant protein sequences and prioritize changed Class-I peptide windows using the existing binding and immunogenicity predictors.
Uses the Class-I allele selected in the main MHC workspace above.
Scope. This first version is deliberately restricted to equal-length substitution variants. Indels/frameshifts require alignment-aware comparison and are not silently approximated.
Scientific diagrams
Prediction and validation
Percentile rank is best used as a prioritization measure. Lower-ranked candidates are selected for closer review, not declared confirmed epitopes.
Binding prediction is a hypothesis-generation step. Presentation and T-cell recognition require additional evidence when they are part of the research claim.
Research use: predicted affinity and percentile rank prioritize candidates; they do not establish natural presentation, T-cell recognition or immunogenicity.
