Neoantigen Prediction

High-resolutionHLA typingwas performed computationally usingSOAP-HLA fromexome sequencing (exome-seq) data.Each nonsynonymousSNVwas translated into a17-mer peptide sequence, centered on the mutated amino acid. Adjacent SNVs(such as those induced by UV irradiation) were first corrected for using MAC.Subsequently, the 17-mer was then used to create 9-mers via a sliding windowapproach for determination of MHC class I binding. NetMHC v4.0 was used todetermine the binding strength of mutated peptides to patient-specific HLAalleles. All peptides with a rank < 2% were considered for further analysis.If one mutation generated multiple 9-mer peptides that bound topatient-specific HLA alleles, it was only counted as one neoantigen (i.e., onemutation could only generate a single predicted neoantigen). We consideredneopeptides as the total number of predicted 9-mers that bound topatient-specific HLA alleles (i.e., one mutation can generate multipleneopeptides).[1]

To predictneoepitopes, ‘‘wild-type’’ petide 17mers (for HLA-I) and 29mers (for HLA-II)with the affected amino acid in the middle for each missense mutation wereretrieved from theGRCh37.74 human reference proteome.To generate ‘‘mutant’’ peptides, the affected aminoacid was replaced in silico with the corresponding mutant amino acid.(SNV)

HLA class I epitope binding predictions

Mutant peptideswere used as input for the T Cell Epitope Prediction Tools included in theImmune Epitope Database and Analysis Resource (IEDB). The HLA class I epitopebinding predictions were performed using the HLA-I IEDB algorithms Consensus andthe artificial neural network (NetMHC) independently yielding same conclusions.For Consensus method – which combines NetMHC, the stabilized matrix method, andthe combinatorial peptide libraries method – 9 mers with a relative percentilerank % 1% for each HLA-I allele were considered binders to cover most of thepotential immune responses as previously suggested.(HLA-I

allele)

HLA class II epitope binding predictions

HLA class IIepitope binding predictions on 15mers were performed using the HLA-II IEDBalgorithms Consensus, NetMHCII version 2.2, and Sturniolo since these were theonly available methods for the patientHLA-II alleles.[2]

[1]    RIAZ N, HAVEL J J, MAKAROV V, et al. Tumorand Microenvironment Evolution during Immunotherapy with Nivolumab [J]. Cell,2017, 171(4): 934-49 e15.

[2]    JIMENEZ-SANCHEZ A, MEMON D, POURPE S, et al.Heterogeneous Tumor-Immune Microenvironments among Differentially GrowingMetastases in an Ovarian Cancer Patient [J]. Cell, 2017, 170(5): 927-38 e20.