PDHC De Novo Design Initiative Roundtable Discussion

The meeting focused on computational approaches for designing peptides, with panelists discussing the trade-offs between physics-based and machine learning methods. Hannes, Yu-Shan, Stephen, Vikram, and Parisa shared their research on using generative models, molecular dynamics, cyclic peptides, and non-natural amino acids to improve peptide design. The group explored how computational methods can interact with experiments to identify starting points and progress structural activity relationships. They also discussed the challenges of designing peptides that bind to targets while being permeable across cell membranes. The panelists agreed that further work is needed to develop methods that can handle non-canonical amino acids and to integrate binding and permeability considerations simultaneously.