NeoSplice

NeoSplice predicts splice variant neoantigens from RNA-seq data to identify tumor-specific antigens (TSAs) for cancer immunogenomics studies, including acute myeloid leukemia.

Key Features:

• Tumor-specific k-mer prediction: Predicts tumor-specific k-mers derived from RNA-seq data.
• Alignment to splice graphs: Aligns differentially expressed k-mers to a splice graph to identify splice variants.
• Variant transcript inference and MHC binding prediction: Infers variant transcripts and predicts their binding affinity to Major Histocompatibility Complex (MHC) molecules.
• Mass spectrometry immunopeptidome validation: Validates predicted neoantigens by mass spectrometry of the immunopeptidome, including confirmation against synthetic peptide references in the K562.A2 cell line.

Scientific Applications:

• Simulated-data performance: Demonstrates over 80% sensitivity and precision across various splice variant classes on simulated RNA-seq data.
• Experimental validation: Confirmed four of 37 predicted antigens corresponding to three unique splice junctions by mass spectrometry of the K562.A2 immunopeptidome using synthetic peptide references.
• Vaccine target discovery: Supports identification of splice variant-derived tumor-specific antigen vaccine targets for cancers, including acute myeloid leukemia.

Methodology:

Predicts tumor-specific k-mers from RNA-seq, aligns differentially expressed k-mers to a splice graph, infers variant transcripts, and predicts MHC binding affinities.