Table 3 In silico tools and pipelines for Neoantigen predictions

Identification of genome variant

 GATK

Genome analysis toolkit

Identify variants across genome using next generation sequencing data

[292]

 MuTect

Somatic point mutation identification in cancer genomes

[291]

HLA typing

 Polysolver

Three major MHC I genes alleles identification based on whole exome sequencing data

[294]

 OptiType

HLA genotyping algorithm that predicts all major and minor HLA class I alleles from next generation sequencing data

[293]

MHC-binding affinity

 netMHC/netMHCII/netMHCpan/netMHCpanII

Prediction of MHC binding affinity to Class I and Class II MHC molecules

[298,299,300, 327]

 SMM

Sequence specificity-based quantitative model to identify binding affinity to MHC I molecules

[328]

 SMMPMBEC

An amino acid similarity matrix derived based on experimental peptide-MHC binding interactions

Act as Bayesian prior for prediction of peptide-MHC class I complex interaction

[329]

 MHCflurry

Allele-specific neural networks trained on MHC ligands identified by mass spectrometry and binding affinity measurements to develop a model for prediction of MHC I complex proteins and their ligands

[330]

 EDGE

Deep learning approach of HLA prediction based on training data from 74 patients

[301]

Pipelines combining all steps of neoantigen prediction

 FRED 2

Prediction, selection, assembly and HLA typing of T-cell epitope

[302]

 NetTepi

Predicts peptide-MHC (pMHC) binding affinity based on integration pMHC stability and T-cell propensity predictions

[304]

 pVAC-Seq

Predicts tumor-specific neoantigen based on the integration of tumor mutation and expression data

[331]