GRATCR: Epitope-Specific T Cell Receptor Sequence Generation With Data-Efficient Pre-Trained Models.

Area of Science:

• Immunology
• Computational Biology
• Bioinformatics

Background:

• T cell receptors (TCRs) are critical for immunotherapies targeting cancer cells.
• Acquiring and optimizing TCRs is experimentally intensive and time-consuming.
• Deep generative models show promise for functional protein sequence generation.

Purpose of the Study:

• To develop a novel framework, GRATCR, for de novo generation of TCR sequences targeting specific epitopes.
• To leverage a unique
• grafting
• strategy combining two pre-trained modules.

Main Methods:

• Utilized a deep generative model framework named GRATCR.
• Employed a novel
• grafting
• strategy to integrate two pre-trained modules.
• Generated TCR sequences targeting specific epitopes.

Main Results:

• GRATCR-generated TCRs demonstrated superior specificity for target epitopes compared to state-of-the-art models.
• Generated TCRs exhibited enhanced biological functionality.
• The model achieved these results using significantly less training data.
• Novel TCR sequences were generated, distinct from natural sequences.
• Interpretability analysis confirmed the model's ability to capture key binding patterns.

Conclusions:

• GRATCR offers an efficient and effective approach for generating high-specificity, functional TCR sequences for immunotherapy.
• The framework reduces reliance on extensive wet lab experiments.
• GRATCR represents a significant advancement in computational approaches for TCR design.