Criteria for folding in structure-based models of proteins

TL;DR

This paper investigates how allowing a small percentage of unformed native contacts in structure-based protein models can significantly improve folding temperature predictions without major effects on folding times, highlighting the importance of backbone stiffness descriptions.

Contribution

It demonstrates that relaxing the all-contacts-formed assumption enhances model accuracy and compares effects of backbone stiffness across six different structure-based models.

Findings

Allowing partial native contact formation boosts folding temperature.

Backbone stiffness significantly influences characteristic temperatures.

Models without backbone stiffness perform similarly to those with stiffness.

Abstract

In structure-based models of proteins, one often assumes that folding is accomplished when all contacts are established. This assumption may frequently lead to a conceptual problem that folding takes place in a temperature region of very low thermodynamic stability, especially when the contact map used is too sparse. We consider six different structure-based models and show that allowing for a small, but model-dependent, percentage of the native contacts not being established boosts the folding temperature substantially while affecting the time scales of folding only in a minor way. We also compare other properties of the six models. We show that the choice of the description of the backbone stiffness has a substantial effect on the values of characteristic temperatures that relate both to equilibrium and kinetic properties. Models without any backbone stiffness (like the self-organized polymer) are found to perform similar to those with the stiffness, including in the studies of stretching.