Neodb: a comprehensive neoantigen database and discovery platform for cancer immunotherapy

Area of Science:

• Oncology
• Immunology
• Bioinformatics

Background:

• Somatic DNA alterations yield cancer-specific neoantigens, crucial for targeted immunotherapy.
• Existing neoantigen discovery methods lack integration, and validated immunogenic neoantigen collections are scarce.
• There is a pressing need for a unified platform to streamline neoantigen discovery and analysis.

Purpose of the Study:

• To develop an integrated web-based platform for neoantigen discovery and analysis.
• To create the most extensive database of experimentally validated neoantigens.
• To introduce a novel graph neural network (GNN) model for enhanced neoantigen immunogenicity prediction.

Main Methods:

• Integrated existing neoantigen discovery tools into a comprehensive web platform.
• Conducted literature searches to compile experimentally validated neoantigens.
• Developed Immuno-GNN, a GNN model with attention mechanisms for predicting neoantigen immunogenicity.
• Filtered potential neoantigens from recurrent driver mutations to create a public neoantigen collection.

Main Results:

• Launched Neodb, an R/Shiny web-based platform containing the largest collection of experimentally validated neoantigens.
• Neodb includes modules for neoantigen prediction tools, public neoantigens from driver mutations, and the Immuno-GNN prediction model.
• Immuno-GNN demonstrated superior performance in neoantigen immunogenicity prediction compared to existing methods.
• This marks the first application of GNNs for neoantigen immunogenicity prediction.

Conclusions:

• The Neodb platform facilitates neoantigen discovery and analysis, supporting cancer immunotherapy research.
• The integrated approach and novel Immuno-GNN model advance the field of neoantigen identification and validation.
• Neodb is poised to accelerate the clinical application of neoantigen-based cancer immunotherapies.