Related Experiment Videos
Prediction of protein structure by simulating coarse-grained folding pathways: a preliminary report
1Searle Chemistry Lab, University of Chicago, 5735 South Ellis Ave #126, Chicago, Illinois 60637, USA. acolubri@uchicago.edu
Journal of Biomolecular Structure & Dynamics
|February 11, 2004
Summary
A new software tool, Folding Machine (FM), models protein folding pathways efficiently. It uses a novel Monte Carlo approach to generate 3D protein structures and study folding kinetics.
Area of Science:
- Computational Biology
- Structural Bioinformatics
- Biophysics
Background:
- Protein structure and kinetics are crucial for biological function.
- Accurate prediction of protein folding pathways remains a significant challenge in computational biology.
Purpose of the Study:
- To develop a software tool for studying protein structure and kinetics.
- To generate low-resolution protein folding pathways using modest computational resources.
- To gain insights into protein folding kinetics and generate 3D structural models.
Main Methods:
- Developed the Folding Machine (FM) software utilizing a coarse-grained kinetic ab initio Monte Carlo sampler.
- Incorporated secondary structure predictions and fragment libraries for enhanced accuracy.
- Discretized conformational space using Ramachandran basins and implicitly treated solvent by rescaling energy terms.
Main Results:
- Generated models within 6 Å backbone RMSD for alpha-helical protein fragments (60-70 residues) in a CASP5 blind test.
- Observed unique failure to converge for a natively unfolded protein target, providing kinetic insights.
- Presented a new metric for evaluating structure prediction accuracy.
Conclusions:
- The Folding Machine (FM) offers a hybrid ab initio/knowledge-based approach for protein folding simulation.
- The tool demonstrates potential for modeling protein structures and understanding folding kinetics.
- Further improvements in predicting beta-sheet structures are ongoing.