Comparison of α-conotoxin free energy landscapes via simulation and dimensionality reduction*

Conotoxins are short, disulfide-rich peptides of interest as therapeutic candidates. Because of the combinatorial nature of the way their cysteines can covalently connect, they can exist in metastable ensembles of structurally distinct "disulfide isomers," which makes them particularly difficult to explore in silico. Traditional molecular dynamics cannot capture the breaking and reforging of disulfide bonds. Here, we simulate four α-conotoxins, containing four cysteines, with their disulfide bonds unconnected. We construct different low-dimensional models to compare their free energies and explore how we can transform between them with minimal additional simulation time. This work represents a first step in the creation of a unified search space for disulfide-rich peptides as therapeutic candidates.