Compact Representation of Continuous Energy Surfaces for More Efficient Protein Design

Area of Science:

• Computational biology
• Structural bioinformatics
• Macromolecular design

Background:

• Macromolecular design accuracy relies on modeling atomic motions around low-energy states.
• Computational cost of energy function calls limits the scope of these models.

Purpose of the Study:

• To develop a method to overcome the computational bottleneck in macromolecular design.
• To improve structural accuracy by efficiently modeling continuous atomic motions.

Main Methods:

• Consolidated conformational energy evaluations using a local polynomial expansion.
• Developed Energy as Polynomials in Internal Coordinates (EPIC) for continuous degrees of freedom.
• Applied EPIC to protein design, modeling side chain and backbone flexibility.

Main Results:

• EPIC efficiently represents energy surfaces for molecular-mechanics and quantum-mechanics energy functions.
• Demonstrated EPIC's effectiveness in accelerating macromolecular design calculations.
• Successfully applied EPIC to protein design, enhancing the modeling of conformational flexibility.

Conclusions:

• EPIC effectively removes computational bottlenecks in macromolecular design.
• The method significantly improves structural accuracy by efficiently modeling atomic motions.
• EPIC is a versatile approach applicable to various energy functions and macromolecular systems.