Predicting the Immunogenicity of T cell epitopes: From HIV to SARS-CoV-2

Abstract

We describe a physics-based learning model for predicting the immunogenicity of Cytotoxic T Lymphocyte (CTL) epitopes derived from diverse pathogens, given a Human Leukocyte Antigen (HLA) genotype. The model was trained and tested on experimental data on the relative immunodominance of CTL epitopes in Human Immunodeficiency Virus infection. The method is more accurate than publicly available models. Our model predicts that only a fraction of SARS-CoV-2 epitopes that have been predicted to bind to HLA molecules is immunogenic. The immunogenic CTL epitopes across all SARS-CoV-2 proteins are predicted to provide broad population coverage, but the immunogenic epitopes in the SARS-CoV-2 spike protein alone are unlikely to do so. Our model predicts that several immunogenic SARS-CoV-2 CTL epitopes are identical to those contained in low-pathogenicity coronaviruses circulating in the population. Thus, we suggest that some level of CTL immunity against COVID-19 may be present in some individuals prior to SARS-CoV-2 infection.

Identification of immunodominant linear epitopes from SARS-CoV-2 patient plasma

“…Another epitope abundantly detected only in healing patients was the 787–822, a peptide segment extending at the periphery of the solvent exposed part of the protein ( Fig 3A and 3B ). It has also been experimentally observed in the SARS-CoV-1 [ 9 , 13 ], SARS-CoV-2 [ 38 , 39 ] and predicted bioinformatically [ 26 , 27 , 30 , 31 , 33 , 36 ]. Interestingly, this epitope includes the S2’ cleavage site of the spike protein ( Fig 4D–4F ), which has been reported to activate the protein for membrane fusion via extensive irreversible conformational changes [ 53 , 65 ].…”

“…Other antibodies with neutralizing activities have been discovered through different methodologies [20][21][22][23][24][25]. The rapid propagation of SARS-CoV-2 stimulated several studies predicting the antigenic parts of the viral proteins in silico [26][27][28][29][30][31][32], and analyzing SARS-CoV-1 epitopes that were conserved in this new coronavirus [33][34][35][36]. More recently, the first reports of experimental epitope mapping of the SARS-CoV-2 were deposited on repositories [37][38][39][40][41][42].…”

Evaluating performance of existing computational models in predicting CD8+ T cell pathogenic epitopes and cancer neoantigens

“…Accurate identification of vaccine targets from viral genomes would fine-tune this process to help mitigate and/or prevent future pandemics. Multiple studies have employed predictive algorithms to identify potential SARS-CoV-2 targets, ranging from publicly available binding predictors to novel systematic workflows [26][27][28][29][30][31][32] . For predictions of immunogenicity it is unclear which publiclyavailable models are the best-performing, and in particular for cancer peptides, evidence is emerging that available predictors of immunogenicity are suboptimal 3,19,24 .…”

In silico T cell epitope identification for SARS-CoV-2: Progress and perspectives

“…In addition to Vaxijen-2.0 and Calis et al, one SARS-CoV-2 study [ 21 ] used iPred [ 81 ] for immunogenicity prediction, which is also based on physicochemical properties of the amino acids in the peptide. A novel method to predict the immunogenicity of SARS-CoV-2 HLA-I restricted peptides was also proposed in Gao et al [ 82 ] which utilizes a physics-based model and takes into account factors such as peptide-HLA binding affinity and similarity of a peptide with pathogen-derived and human-derived peptides. For model training and testing, a well-characterized dataset of immunogenic HIV T cell epitopes was used.…”