Table 2.

Methods for analyzing steps in the antigen-processing pathway and for HLA typing

Predictor/tool	Key method	Reference
HLA class I binding
Allele-specific
SYFPEITHI	PSSM	[14]
RANKPEP	PSSM	[27]
BIMAS	PSSM	[28]
SVMHC	SVM	[7]
netMHC	ANN	[29]
Pan-specific
MULTIPRED	HMM/ANN	[39]
netMHCpan	ANN	[40]
PickPocket	PSSM	[41]
TEPITOPEpan	PSSM	[42]
ADT	Threading	[43]
UniTope	SVM	[44]
KISS	SVM	[45]
HLA class II binding
Allele-specific
SYFPEITHI	PSSM	[14]
netMHCII/SM-align	PSSM/ANN	[48, 49]
ProPred	PSSM	[50]
RANKPED	PSSM	[27]
TEPITOPE	PSSM	[51]
SVRMHC	SVM	[8]
MHC2MIL	Multi-instance learning	[52]
MHC2pred	SVM	–
Pan-specific
MULTIPRED	HMM/ANN	[39]
MHCIIMulti	Multi-instance learning	[55]
TEPITOPEpan	PSSM	[42]
netMHCIIpan	ANN	[56, 90]
Consensus methods
CONSENSUS	–	[57]
netMHCcon	–	[56]
Binding stability
netMHCstab	ANN	[47]
Proteasomal cleavage
in vitro
netChop 20S	ANN	[60]
PCM	PSSM	[61]
FragPredict	PSSM	[62]
Pcleavage	SVM	[63]
PAProC	ANN	[64]
in vivo
netChop Cterm	ANN	[60]
ProteaSMM	PSSM	[65]
TAP transport
PredTAP	HMM/ANN	[39]
SVMTAP	SVM	[61]
Integrated processing
EpiJen	–	[70]
WAPP	–	[61]
NetCTL	–	[71]
NetCTLpan	–	[72]
T-cell reactivity
POPI	SVM	[74]
POPISK	SVM	[75]
B-cell epitope prediction
Continuous
COBEpro	SVM	[78]
BCPRed	SVM	[79]
FBCPred	SVM	[79]
Discontinuous
EPMeta	SVM	[82]
Discotope 2.0	Linear regression	[83]
NGS-based HLA typing
ATHLATES	Contig assembly	[25]
seq2HLA	Greedy algorithm	[26]
OptiType	Integer linear programming	[19]
Polysolver	Bayesian classification	[20]

Abbreviations: ANN artificial neural network, HLA human leukocyte antigen, HMM hidden Markov model, NGS next-generation sequencing, PSSM position-specific scoring matrix, SVM support vector machine, TAP transporter associated with antigen processing