Fig. 2: High DARPin mutation frequency in DANA is correlated to antigen complexity.

a Comparison of mean VH germline identity (GLI) of trimer- (n = 444) and V3 peptide-reactive (n = 431) clonotypes among 80,963 B cell receptor (BCR) sequences from 19 HIV-1 bnAb inducers (total 21 PBMC samples) analyzed by Libra-seq. Trimer- and V3-binding double-positive BCRs were excluded. Boxplots indicate median (middle line), upper and lower quartiles (box limits) and 1.5× interquartile ranges (whiskers). Mean VH GLIs were compared by two-sided Wilcoxon rank-sum test. b Generic DARPin design comprising amino (N-cap) - and carboxyterminal (C-cap) ankyrin repeats, flanking up to three internal ankyrin repeats. Conserved framework in cyan, positions randomized in the library in red. c Sequence analysis workflow. DARPins with valid ORF were categorized into N1C, N2C or N3C (117, 150, and 183 amino acids respectively)-type DARPins. DARPins with insertions/deletions ( > 1 aa) were considered mutated. DARPins were pairwise aligned to respective consensus sequences of the same DARPin type (N1C, N2C, N3C) and alignment scores calculated. DARPins with >5% framework substitutions were considered mutated. d DARPin length distribution for DANA 1–5. Red dotted lines indicate length for typical N1C, N2C and N3C DARPins. e DARPin framework mutation analysis by alignment scores and DARPin type distribution for DANA 1–5. Higher alignment scores indicate higher similarity to the DARPin-type consensus sequence. DARPins with insertions/deletions ( > 1aa) are not considered in this analysis. f DARPin frequency without mutations (i.e., typical DARPins), with insertions/deletions and with >5% framework mutations. All DARPins with valid ORF (n = 126–184) were analyzed.