NeoantigenID, a suite of advanced analytics, is available via the ImmunoID NeXT Platform for the accurate identification of tumor-specific somatic mutations (Figure 1). These mutations are a rich source of putative neoantigens. Current computational neoantigen prediction algorithms suffer from limited training datasets and poor performance. To overcome these challenges, Personalis has developed the Systematic HLA Epitope Ranking Pan Algorithm (SHERPA™). SHERPA leverages mass spectrometry-based immunopeptidomics training data to improve neoantigen presentation prediction based on MHC Class I binding potential, level of gene expression, and other key parameters (Figure 2). SHERPA is integrated into the NeoantigenID analytics engine for comprehensive characterization of putative neoantigens.
Figure 1: Workflow of the NeoantigenID Analytics Engine
Figure 2: Overview of SHERPA machine learning algorithm
The advanced computational pipeline configurations produce data-rich analytics:
- Accurate identification of SNVs, indels and fusions, which are an abundant source of potentially-immunogenic neoantigens
- Robust and accurate HLA typing of Class I and Class II MHC loci followed by somatic mutation and allele-specific loss of heterozygosity (LOH) detection in the HLA genes
- Comprehensive characterization of each putative neoantigen by combining tumor-specific small variants and fusions with the patient-specific HLA types and predicting neoantigens using our pan-allelic machine learning algorithm, SHERPA™
– SHERPA relies upon a proprietary, high quality and unambiguous training dataset generated by performing immunopeptidomics on a robust set of MHC Class I alleles using monoallelic cell lines (Figure 2).
– To comprehensively capture all aspects of epitope presentation, SHERPA incorporates both antigen processing machinery and gene expression information along with peptide and binding pocket information to predict a presentation rank.
- Additional immunogenomics analytics to capture the full tumor biology and the complexity of the tumor microenvironment
- Determination of neoantigen burden as well as the generation of the Personalis Composite Neoantigen Presentation Score (NEOPS) which has the potential to improve the predictive and/or prognostic utility of neoantigen burden-based biomarkers for precision oncology and immunotherapy applications.