Fig. 2. Sequence sampling.

Local Distance Difference Test (lDDT) of AlphaFold predicted structures to the reference scaffold monomers for sequences generated using CARBonAra with, as objective, maximum sequence identity, minimum sequence identity, and low sequence similarity, against the (a) sequence identity and (b) sequence similarity. c lDDT of the AlphaFold predicted structures as a function of the highest expect value (E-value) of the generated sequences from a BLAST²⁶ search. d Scaffold, in white, from the birch pollen allergen Bet v 1 protein (PDB ID: 6R3C). AlphaFold predicted structure, shown in red, from the generated sequence has a lDDT of 70 from the reference scaffold. The generated sequence has a 7% identity and 13% similarity with the original scaffold protein.