Related Experiment Videos
Application of SVM to predict membrane protein types.
Yu-Dong Cai1, Pong-Wong Ricardo, Chih-Hung Jen
1Shanghai Research Centre of Biotechnology, Chinese Academy of Sciences, Shanghai 200233, China. y.cai@umist.ac.uk
Journal of Theoretical Biology
|February 5, 2004
Summary
This study introduces the support vector machine (SVM) for predicting membrane protein types, building on prior automated methods. SVM shows promise as a powerful tool for protein attribute prediction when combined with existing algorithms.
Area of Science:
- Computational Biology
- Bioinformatics
- Protein Science
Background:
- Automated prediction of membrane protein types is an ongoing challenge in bioinformatics.
- Previous work by Chou and Elrod laid the groundwork for computational approaches to protein attribute prediction.
Purpose of the Study:
- To introduce and evaluate the support vector machine (SVM) as a novel method for predicting membrane protein types.
- To assess the efficacy of SVM in conjunction with the covariant discriminant algorithm for enhanced protein attribute prediction.
Main Methods:
- Application of the support vector machine (SVM) algorithm.
- Validation using re-substitution, jackknife, and independent data set tests.
- Integration of SVM with the covariant discriminant algorithm.
Main Results:
- The SVM approach demonstrated significant promise in predicting membrane protein types.
- Testing across multiple datasets confirmed the effectiveness of the SVM method.
- SVM shows potential for accurate prediction of various protein attributes.
Conclusions:
- The support vector machine (SVM) is a viable and promising tool for automated membrane protein type prediction.
- Combining SVM with the covariant discriminant algorithm offers a powerful strategy for predicting protein types and other attributes.
- This integrated approach advances computational methods in protein science.