NeoSplice: a bioinformatics method for prediction of splice variant neoantigens

Area of Science:

• Oncology
• Immunology
• Bioinformatics

Background:

• Splice variant neoantigens are tumor-specific antigens (TSAs) found in various cancers, including acute myeloid leukemia.
• Existing genomic prediction tools for splice variant neoantigens lack comprehensive validation, particularly through integrated simulated and experimental approaches.
• There is a need for validated methods to identify these neoantigens for cancer vaccine development.

Purpose of the Study:

• To introduce NeoSplice, a novel computational method for predicting splice variant neoantigens.
• To validate NeoSplice's predictions using both in silico simulations and targeted immunopeptidome mass spectrometry.
• To provide a reliable tool for identifying potential tumor-specific antigen vaccine targets.

Main Methods:

• NeoSplice predicts tumor-specific k-mers from RNA-seq data.
• It aligns differentially expressed k-mers to the splice graph and infers variant transcripts with MHC binding predictions.
• Validation involved in silico simulated RNA-seq data and mass spectrometry analysis of the K562.A2 cell line immunopeptidome.

Main Results:

• NeoSplice achieved high sensitivity and precision (>80%) on simulated RNA-seq data across various splice variant classes.
• Mass spectrometry validated 4 out of 37 predicted splice variant neoantigens, corresponding to 3 unique splice junctions.
• The study provides a comparative analysis of NeoSplice against existing prediction tools.

Conclusions:

• NeoSplice is a highly sensitive and precise computational method for predicting splice variant neoantigens.
• The study presents the first targeted immunopeptidome mass spectrometry approach for validating predicted splice variant neoantigens.
• NeoSplice offers a well-validated platform for identifying TSA vaccine targets to advance cancer immunotherapy.