Computational cyclic peptide design machine learning & Rosetta based methods

Macrocyclic peptides offer a promising alternative to antibodies and small molecules for targeting flat, intracellular surfaces that are often considered "undruggable." Compared to their linear counterparts, macrocycles also exhibit enhanced stability and binding specificity. Head-to-tail cyclic peptides have increased proteolytic protection due to the non-exposed charged termini, and the incorporation of non-canonical amino acids is straight forward given that they are chemically synthesized. However, the rational design of cyclic peptides remains a significant challenge due to conformational constraints and the complexity of cyclic backbone sampling. This chapter reviews current computational and experimental algorithms for cyclic peptide design.