Extended Data Fig. 5: SARS-CoV-2 epitope recognition and HLA binding position prediction.
From: Self-iterative multiple-instance learning enables the prediction of CD4+ T cell immunogenic epitopes
a, Summary of maximal responses to all peptides (n = 29) across tested donors (n = 8), grouped by DR status. Immune responses were quantified as spot-forming cells (SFCs) per 10,000 PBMCs and plotted on a \({\log }_{10}(\frac{{sfcs}}{\mathrm{10,000}}+1)\) scale. Each point represents the maximal response of a donor to a peptide, based on duplicate ELISpot wells. The dashed line at 1.415 notes the cutoff for donor-peptide responses, representing 25 sfcs/10,000 cells. The response rate is expressed as a percentage. An inset shows P value was calculated using a two-sided Fisher’s exact test, to compare the DR1+/DR4+ statuses (groups) with the positive peptide responses. b, Summary of DR1+/DR4+ status predictions by ImmuScope-IM for all peptides (n = 29). The P value was calculated using two-sided Wilcoxon signed rank test to compare the peptide responses associated with DR1+/DR4+ status. c, d, e, f, Binding positions and alignment scores on HLA-DR1-S511-530 (PDB: 8CMC, c), HLA-DR1-M176-190 (PDB: 8CME, d), HLA-DR1-nsp31350-1364 (PDB: 8CMF, e) and HLA-DR1-nsp146420-6434 (PDB: 8CMG, f), respectively.
