tcr-pmhc-prediction-agent

Core Capabilities

1. Binding Prediction: Predict TCR-pMHC binding affinity/probability.

2. Structural Modeling: Generate TCR-pMHC complex structures with AlphaFold3.

3. Epitope Specificity: Determine which epitopes a TCR recognizes.

4. Cross-Reactivity Assessment: Predict off-target self-peptide binding.

5. Immunogenicity Scoring: Rank peptide immunogenicity.

6. Therapeutic TCR Screening: Screen TCRs for desired specificity.

Prediction Approaches

Approach	Method	Strengths
AlphaFold3	Structure prediction	High accuracy, interpretable
TCR-BERT	Sequence transformer	Fast, large-scale
ERGO-II	RNN-based	Established benchmark
pMTnet	Multi-task learning	Generalizable
NetTCR	CNN-based	HLA-specific
TITAN	Attention-based	State-of-art sequence

Workflow

1. Input: TCR sequence (alpha/beta CDR3), peptide, HLA allele.

2. Structure Prediction: Generate pMHC and TCR structures.

3. Docking: Model TCR-pMHC complex.

4. Scoring: Calculate binding probability/affinity.

5. Cross-Reactivity: Screen against self-peptide database.

6. Validation Features: Extract structural determinants.

7. Output: Binding predictions, structures, safety assessment.