Table 3.

Computational tools for predicting antigen specificity.

Computational tool	Input	Output	Unique features	References
GLIPH2	CDR3 region of TCRα and β chain	Clustering of TCR sequences into groups of shared specificity	– Efficiently processes and clusters millions of TCR sequences	(37)
iSMART	CDR3 region of TCRβ chain	Clustering of TCR sequences into groups of shared specificity	– Built on pan-cancer multi-omics analysis – Utilizes TRUST algorithm to assemble hypervariable CDR3 regions	(38, 65)
ERGO-II	CDR3 region of TCRα and β chain	TCR-peptide binding prediction	– Includes MHC typing and V and J gene segment alleles in prediction model	(67)
TCRMatch	CDR3 region of TCRβ chain	TCR-peptide binding prediction	– Can be used in conjunction with other tools	(68)
NetTCR 2.0	Paired TCRα and β chain	TCR-peptide binding prediction	– Accounts for different TCR sequence lengths – Improved processing of paired TCRα and β chain	(69)
pMTnet	CDR3 region of TCRβ chain	TCR-peptide binding prediction	– Accurately predicts TCRs, epitopes, and HLAs previously not observed – Utilizes transfer learning for related TCR and pMHC data lacking pairing information	(35)
IGoR	TCRβ and BCR heavy chains	V(D)J recombination and somatic hypermutation prediction	– Performance in identifying the correct V(D)J recombination scenario is 2× better than current methods – Can assess the mutational landscape of BCRs	(70)
OLGA	CDR3 sequence of TCR and BCR	V(D)J recombination prediction	– Computes the probabilities of CDR3 sequences and motifs of four chain loci (human and mouse TCRβ, human TCRα and human IGH) – Can identify outlying sequences in repertoire sequencing datasets to study disease immune response	(74)
SONIA	CDR3 sequence of TCRβ chain	V(D)J recombination and selection pressure prediction	– Distinguishes between generation and selection to quantify contribution to the overall variability	(79)

Note: Many tools have been developed to improve and streamline the identification of TCR-peptide pairs. Input, output, and unique features are listed.