Related Experiment Videos
Using multiple structure alignments, fast model building, and energetic analysis in fold recognition and homology
Donald Petrey1, Zhexin Xiang, Christopher L Tang
1Howard Hughes Medical Institute, Department of Biochemistry and Molecular Biophysics, Center for Computational Biology and Bioinformatics, Columbia University New York, New York 10032, USA.
Proteins
|October 28, 2003
Summary
This study presents in-house software for protein structure prediction, focusing on generating accurate sequence-to-structure alignments and rapid model building. The methods, including HMAP and NEST, were refined during CASP5, highlighting interactive model building for improved accuracy.
Area of Science:
- Computational Biology
- Structural Bioinformatics
- Protein Structure Prediction
Background:
- Protein structure prediction is crucial for understanding biological function.
- CASP (Critical Assessment of protein Structure Prediction) is a community-wide experiment for evaluating methods.
- Accurate sequence-to-structure alignments are fundamental for reliable protein modeling.
Purpose of the Study:
- To evaluate in-house developed software for protein fold recognition and homology modeling in the CASP5 competition.
- To demonstrate a strategy combining sequence-to-structure alignment generation with rapid model building and evaluation.
- To identify areas for improvement in protein structure prediction methodologies.
Main Methods:
- Utilized HMAP (Hybrid Multidimensional Alignment Profile) for profile-to-profile sequence alignment.
- Employed NEST, a fast model building program using an artificial evolution algorithm.
- Integrated GRASP2 for visualization, multiple structure superposition, and domain database scanning.
- Applied energy-based functions (all-atom and simplified) for model evaluation.
Main Results:
- Successfully participated in the fold recognition and homology modeling sections of CASP5.
- Demonstrated the effectiveness of combining alignment generation with fast model building and energy-based evaluation.
- Identified advantages of interactive model building procedures through manual analysis.
Conclusions:
- The developed in-house software provides a robust framework for protein structure prediction.
- Interactive model building and refinement are essential for enhancing prediction accuracy.
- The study offers insights for future improvements in protein structure prediction algorithms and strategies.