Related Experiment Videos
Secondary structure prediction for modelling by homology
P E Boscott1, G J Barton, W G Richards
1Physical Chemistry Laboratory, Oxford, UK.
Protein Engineering
|April 1, 1993
Summary
This study introduces a novel consensus algorithm for protein secondary structure prediction. The method enhances accuracy by customizing predictions based on homologous protein families, improving upon existing methods like GOR.
Area of Science:
- Computational Biology
- Structural Bioinformatics
- Protein Science
Background:
- Accurate protein secondary structure prediction is crucial for homology modeling.
- Existing methods like Garnier, Osguthorpe, and Robson (GOR) have limitations in predictive accuracy.
Purpose of the Study:
- To develop an improved method for protein secondary structure prediction using consensus algorithms.
- To enhance the accuracy of secondary structure prediction by tailoring algorithms to specific protein families.
Main Methods:
- Developed consensus algorithms by scaling and combining data from four published prediction methods.
- Used homologous proteins of known structure to tune and validate the customized consensus algorithms.
- Compared the new method against GOR and sequence alignment using 31 proteins from five families.
Main Results:
- The consensus algorithm demonstrated improved statistical secondary structure prediction accuracy.
- Accuracy improvements over GOR ranged from 3% to 7%, correlating with alignment significance scores.
- The method showed potential for integration into more comprehensive prediction systems.
Conclusions:
- The developed consensus-based approach offers a significant improvement in protein secondary structure prediction.
- Customizing prediction algorithms to homologous protein families enhances accuracy.
- This method provides a valuable tool for protein modeling and structural bioinformatics.