Proteogenomic analysis reveals RNA as a source for tumor-agnostic neoantigen identification

The methodology combined whole exome/genome sequencing, RNA sequencing, and liquid chromatography-tandem mass spectrometry of HLA class I-bound peptides, enhanced by a machine learning tool (Prosit). A key innovation was the inclusion of RNA-seq data to capture variants from RNA processing events like splicing and editing, which substantially increased the number of identified neoantigen candidates compared to a DNA-only approach. The pipeline included in vitro immunogenicity assessments and validation against normal tissue expression data. The work demonstrates the importance of RNA as a source for neoantigen discovery and provides a framework for identifying shared targets across cancer types.