Quantitative prediction of amyloid fibril growth of short peptides from simulations: calculating association constants to dissect side chain importance.

Quantitative prediction of the fibril growth properties of different peptides is conducted with a molecular dynamics approach. Association constants of small peptides used as a model for amyloid formation are calculated, and the results show very good agreement with experiments. Also the free-energy differences associated with two important interactions that characterize fibril growth, namely cross-beta-sheet and lateral interactions, are obtained. These two interactions show different dependencies on the physicochemical properties of the side chains, explaining the variation in fibril morphologies between different peptides.