Related Experiment Videos
Combining the GOR V algorithm with evolutionary information for protein secondary structure prediction from amino
A Kloczkowski1, K-L Ting, R L Jernigan
1Laboratory of Experimental and Computational Biology, NCI, NIH, Bethesda, Maryland, USA.
Proteins
|September 5, 2002
Summary
This study enhances protein secondary structure prediction using an improved GOR algorithm with evolutionary information and triplet statistics. The refined method achieves higher accuracy, paving the way for the GOR V release.
Area of Science:
- Computational Biology
- Bioinformatics
- Structural Biology
Background:
- Protein secondary structure prediction is crucial for understanding protein function.
- Previous GOR algorithms have limitations in accuracy and scope.
- Evolutionary information and advanced statistical methods can improve prediction.
Purpose of the Study:
- To significantly improve the GOR algorithm for protein secondary structure prediction.
- To incorporate evolutionary information from multiple sequence alignments and triplet statistics.
- To expand the prediction capabilities to shorter protein sequences.
Main Methods:
- Modified and optimized the GOR algorithm.
- Integrated PSI-BLAST multiple sequence alignments.
- Utilized an expanded database of 513 non-redundant protein domains.
- Introduced a variable-size window for sequence analysis.
Main Results:
- Achieved an average prediction accuracy of 73.5% using multiple sequence alignments.
- Increased accuracy to 74.2% for sequences with at least 50 PSI-BLAST alignments.
- The improved algorithm without multiple sequence alignments reached 67.5% accuracy, a 3% gain over GOR IV.
Conclusions:
- The enhanced GOR algorithm demonstrates superior performance in protein secondary structure prediction.
- The integration of evolutionary information is key to improved accuracy.
- This work lays the foundation for the GOR V online prediction server.