Combined covalent-electrostatic model of hydrogen bonding improves structure prediction with Rosetta

Area of Science:

• Computational Biology
• Biophysics
• Molecular Modeling

Background:

• Modeling polar atom interactions is complex due to the dual nature of hydrogen bonds (H-bonds) and electrostatics.
• H-bonds exhibit orientation preferences at short ranges and electrostatic interactions at longer ranges.

Purpose of the Study:

• To develop a refined energy function for molecular modeling that accurately captures both H-bond and electrostatic interactions.
• To improve the performance of the Rosetta molecular modeling program in simulating biological systems.

Main Methods:

• Refined an orientation-dependent H-bond model previously used in Rosetta.
• Combined the refined H-bond model with a distance-dependent Coulomb model for electrostatics.
• Physically motivated the functional form of the H-bond potential and fitted parameters to crystal structure data.

Main Results:

• The combined potential accurately models H-bond geometries, resembling those in high-resolution crystal structures.
• Demonstrated improved performance in benchmarks including decoy discrimination, side chain prediction, and native sequence recovery.
• Established a new standard energy function for the Rosetta program.

Conclusions:

• The new energy function provides a more accurate representation of polar atom interactions in molecular simulations.
• This advancement enhances the capabilities of Rosetta for protein design and other molecular modeling applications.