Fig. 2: Peptide selection and antigen design.

a Output graph of predicted B-cell linear epitopes showing amino acid position (x-axis) and Bepipred score (Y-axis). Residues (yellow) with scores above the default threshold (red line) of 0.35 have a higher probability to be part of an epitope. b Five predicted epitopes (AG1–AG5) of lengths ≥12 amino acids (start and end positions indicated). c Swiss model of hANKRD1 from Pro₁₂₂ to Thr₃₁₃ using ANK-N5C-281 (PDB entry 4qfv.2.B) as template. Antigen sequences, AG4 and AG5, are shown in red. The sequence upstream of Pro₁₂₂ could not be modeled. A schematic representation of this region harboring AG1 is depicted by dotted lines. d Overlay of predicted antigenic sequences (yellow) against titin (green) and calsequestrin (purple) binding sites and the coiled coil region (orange) on hANKRD1. Putative nuclear localization (red dotted boxes), nuclear exporting signals (black dotted box), caspase 3 (red line) and calpain 3 (black line) proteolytic cleavage sites are shown. e Amino acid sequences of Trx-fused immunogens harboring three-copy inserts of antigen sequences (yellow) flanked by long GSGSG linkers (black underlined). The middle insert is separated from its flanking counterparts by shorter GSG linkers. The corresponding 3D structure predicted by homology modeling against thioredoxin 1 (1oaz.1.A.) as template are shown below each sequence. The coloured regions correspond to thioredoxin from its amino terminus (deep blue) to its carboxy terminus (red). The regions in white comprise the three-copy peptide inserts and flanking GSGSG linkers.