Table 3 Acquisition strategies and main methods for neoantigens

Direct acquisition of neoantigens

Sampling and screening of an existing population of tumor tissue, usually from biopsy or surgical resection

Aim: to find available targets in their naturally occurring antigen pools

Immunoprecipitation-MS [102,103,104,105]

Patient tumor tissues

To capture primary tumor p-MHC

Affinity-tag extraction [106]

Animal tumor tissues with specific MHC type tagged

To precisely extract known and neo-antigens in situ

RNA-seq and WGS [107, 108]

Patient tumor tissues

To obtain complete serial sequence information of one patient

peptide-MHC libraries [109]

Specific TCR or acquired T cells, and constructed vector libraries

To undifferentiately screen one TCR- recognizable known epitopes

MANAFEST, T-SCAN [110, 111]

Specific TCR

To high-throughput screen recognizable epitopes

SABR [112]

Specific TCR

To screen homologous epitopes

Trogocytosis [114]

T cells fluorescently labeled with membrane proteins

To trace target cells binding and then sequence involved TCRs

Hansolo system [115]

Patient T cells and immortalized B cell lines

To construct unbiased mutanome minigene recognizable library of the patient

Predictive modeling of neoantigens

Acquisition of patient's MHC molecular profile (individual-specific MHC typing)

In silico analysis and prediction of deliverable epitopes combined with simulation of realistic multi-step parameter optimization, with attention to distortion or overestimation of the predicted epitope library

Aim: capture of possible key antigens for usable TCR design

TCR and antigen prediction

1. Personalized information and MHC typing

2. Computerized prediction models:

i. HLA typing

ii. mutation typing and calling

iii. HLA binding prediction

iv. TCR prediction

v. TCR priority

vi. TCR-recognizing HLA screen [135]

3. Design of the corresponding TCR at the optimal epitope-MHC

Databases for in silico pre-analysis:

• whole-genome sequencing and WES

• RNA-seq

• proteomics

• MS

To predict epitopes and also exclude self-reactive antigens on a large-scale, use sequence information and select models