Fig. 4: Clonal structure of SARS-CoV-2 epitope-specific response.

a Clusters of CDR3β sequences of epitope-specific CD8+ T cells. Each node represents a unique CDR3β amino-acid sequence or public sequences, the node size is proportional to the number of identical clonotypes. Lines connect similar CDR3β sequences, the thickness indicates Levenshtein distance = 1 or 2. Colors indicate donors. Only clusters with two or more members are shown. Clusters with 6 or more nodes highlighted with color; b proportion of clustered epitope-specific T-cell clonotypes; c fraction of public (dark red) and unique (dark green) epitope-specific clonotypes. Numbers of public clones are listed at top; d histograms of V (upper plots) and J-gene usage (lower plots) usage in LLY- (left) and YLQ-specific clones (right); e position-weight matrices for CDR3β sequences with the most common length observed in LLY and YLQ clusters. Cluster numbers corresponds to numbers shown in a; f probability of V(D)J-recombination with the observed V- and/or J-gene usage as calculated by the OLGA algorithm (57). Solid line represents the mean and dashed line represents the median Pgen. Mann-Whitney test, p-values are indicated, numbers of clones are listed at bottom; g Spearman correlation between various parameters of SARS-CoV-2 epitope-specific T-cell response. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001.