Related Experiment Videos
Classification of protein quaternary structure by functional domain composition.
Xiaojing Yu1, Chuan Wang, Yixue Li
1Bioinformatics Center, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, 320 Yueyang Road, Shanghai 200031, China. xjyu@sibs.ac.cn
BMC Bioinformatics
|April 6, 2006
Summary
Classifying protein quaternary structure computationally is crucial for understanding biological functions. A novel domain composition approach using the nearest neighbor algorithm achieved high accuracy, outperforming other methods.
Area of Science:
- Bioinformatics
- Computational Biology
- Structural Biology
Background:
- Protein quaternary structure, the arrangement of subunits, is vital for protein function and biological processes.
- Oligomeric proteins play key roles in metabolism, signal transduction, and chromosome replication.
- Developing computational methods for automatic quaternary structure classification from protein sequences is highly desirable.
Purpose of the Study:
- To develop and evaluate a computational method for classifying protein quaternary structure based on functional domain composition.
- To assess the effectiveness of the nearest neighbor algorithm (NNA) for this classification task.
Main Methods:
- Proteins were represented by vectors derived from their functional domains in the PFAM database.
- The nearest neighbor algorithm (NNA) was employed for quaternary structure classification.
- A jackknife cross-validation test was performed on a non-redundant dataset (sequence identity < 25%).
Main Results:
- The domain composition approach achieved an overall success rate of 75.17% in cross-validation.
- Testing on an independent dataset yielded a higher success rate of 84.11%.
- This method demonstrated superior performance compared to amino acid composition and BLAST methods.
Conclusions:
- The domain composition approach is an effective and promising high-throughput method for classifying protein quaternary structure.
- This bioinformatics strategy offers advantages over traditional methods for analyzing complex protein structures.
- The findings highlight the utility of domain information in predicting protein quaternary structure.