Related Experiment Videos
Sculpting proteins interactively: continual energy minimization embedded in a graphical modeling system
M C Surles1, J S Richardson, D C Richardson
1San Diego Supercomputer Center, California 92186-9784.
Protein Science : a Publication of the Protein Society
|February 1, 1994
Summary
Sculpt is a new protein modeling system that combines interactive graphics with fast energy minimization for physically realistic simulations. This approach enables large conformational changes and aids understanding of protein folding and design.
Area of Science:
- Computational Biology
- Structural Biology
- Biophysics
Background:
- Protein modeling traditionally involves either interactive manipulation or energy minimization.
- Existing methods often lack real-time physical realism or direct user control for complex modifications.
Purpose of the Study:
- Introduce a novel protein modeling paradigm combining interactive computer graphics with continual physical validation.
- Develop a system enabling direct user control and visualization of physically realistic protein structures.
Main Methods:
- Developed the Sculpt system featuring a fast energy minimization algorithm for real-time performance.
- Implemented rigid constraints for bond lengths, angles, and planar groups, with elastic restraints for energy terms.
- Utilized an augmented Lagrange-multiplier method for efficient local energy minimization, achieving linear time complexity with atom count.
Main Results:
- Sculpt achieves real-time updates (11/sec on 20 residues, 2/sec on 80 residues) on standard hardware.
- Demonstrated applications in protein design, including reversing helix bundle packing and folding beta-ribbons.
- Successfully designed a peptide sequence and conformation mimicking protein subunit interactions.
Conclusions:
- The Sculpt paradigm offers significant advantages for modeling large conformational changes and de novo protein design.
- Interactive, physically realistic models enhance user understanding of energy term contributions to protein stability.
- This approach represents a productive integration of interactive modeling, energy minimization, and physical principles for advancing computational protein science.