Database-derived potentials dependent on protein size for in silico folding and design

Area of Science:

• Computational Biology
• Biophysics
• Structural Bioinformatics

Background:

• Knowledge-based potentials are crucial for protein simulations, offering computational efficiency for large-scale modeling.
• However, their accuracy is limited by dataset specificity and unclear physical interpretations.

Purpose of the Study:

• To investigate the theoretical underpinnings of knowledge-based potentials as mean-force potentials.
• To develop a protein size-dependent potential to enhance accuracy in structure prediction.

Main Methods:

• Analysis of distance-dependent amino acid pair potentials derived from protein structure datasets of varying lengths (N).
• Theoretical probing of potentials considering implicit solvent effects and entropic contributions.
• Development of a novel size-dependent potential based on observed trends.

Main Results:

• Distance-dependent potentials exhibit a 1/N behavior at large inter-residue distances, influenced by system boundaries and compressibility.
• Short-distance interactions show distinct patterns based on residue pair properties (electrostatic, cation-pi, pi-pi, hydrophobic).
• The novel size-dependent potential improved discrimination between native and decoy protein structures.

Conclusions:

• Knowledge-based potentials have a theoretical basis as mean-force potentials, incorporating entropic and solvation effects.
• Protein size significantly impacts potential behavior, necessitating size-dependent formulations.
• The developed potential enhances the reliability of protein structure prediction and design.