Prediction of SARS-CoV-2 epitopes across 9360 HLA class I alleles

Katie M. Campbell; Gabriela Steiner; Daniel K. Wells; Antoni Ribas; Anusha Kalbasi

In order to better understand the class I antigenic profile of the SARS-CoV-2 proteome, we focused on pMHCs with a binding affinity of less than 500nM (n=6,748). Of the 1,103 unique peptides comprising these pMHCs, there were between 1 and 55 (median 3) alleles that could bind each peptide (Figure 1B) , indicating that multiple HLA types could present the same antigens. Antigens were annotated by their corresponding protein (Methods) to identify which viral proteins may be responsible for class I recognition. There were between 12 and 684 (median 31) high-binding peptides observed in each of the 11 viral proteins (Orf1ab, Surface/Spike Glycoprotein, Orf3a, Envelope (Orf4; E), Membrane Glycoprotein, Orf6, Orf7a, Orf7b, Orf8, Nucleocapsid (Orf9), and Orf10; Figure 1C, Supplementary Figure 1) . These predicted peptides were distributed throughout each protein ( Figure 1D) . Orf1ab was the largest protein and had the highest number of peptides (n=684), followed by the Surface/Spike Glycoprotein (n=109) and the Nucleocapsid (Orf9; n=60). When normalized by amino acid length, the largest proteins had the lowest density of peptides.