NAP-CNB

NAP-CNB predicts MHC-I-restricted neoantigens from tumor RNA sequencing data using variant calling and recurrent neural networks to support neoantigen discovery for cancer immunotherapy research.

Key Features:

• Integrated Pipeline: Provides a fully automated, integrated pipeline that streamlines neoantigen prediction from tumor-derived sequence data.
• Recurrent Neural Networks (RNN): Uses recurrent neural networks trained to predict T-cell epitope immunogenicity by processing sequence-based features.
• Variant Calling and Selection: Employs variant calling on RNA sequencing data to identify and select candidate neoantigenic variants.
• High Predictive Accuracy: Demonstrates high accuracy in estimating H-2 peptide ligands, reporting an area under the curve (AUC) of 0.95.
• Application to Murine Models: Validated as a proof-of-concept on the B16 melanoma mouse model, reporting putative neoantigens.

Scientific Applications:

• Personalized cancer vaccine target identification: Facilitates selection of MHC-I-restricted neoantigens for personalized vaccine design.
• Adoptive T-cell therapy target discovery: Identifies candidate epitopes that can be prioritized for adoptive T-cell therapy development.
• Preclinical tumor immunology: Supports neoantigen discovery and immunogenicity studies in murine models such as B16 melanoma using RNA-Seq-derived predictions.

Methodology:

Performs variant calling on tumor RNA sequencing data and applies recurrent neural networks trained to predict T-cell epitope immunogenicity and H-2 MHC-I peptide ligands.