A Tool Preference Choice Method for RNA Secondary Structure Prediction by SVM with Statistical Tests

Area of Science:

• Computational Biology
• Bioinformatics
• Genomics

Background:

• RNA secondary structure prediction is crucial in molecular biology.
• Existing prediction tools have limitations and varying strengths.
• Integrating multiple tools can potentially enhance prediction accuracy.

Purpose of the Study:

• To develop an improved method for RNA secondary structure prediction.
• To integrate three existing tools: pknotsRG, RNAStructure, and NUPACK.
• To enhance prediction accuracy by combining feature selection and classifier fusion.

Main Methods:

• A tool choice method based on support vector machines (SVM) was developed.
• Feature extraction and information-theoretic feature selection were employed for ranking.
• An incremental approach combined feature selection and classifier fusion.
• A dataset of 720 RNA sequences (225 pseudoknotted, 495 nested) was used for testing.

Main Results:

• The proposed incremental method achieved a base-pair accuracy of 75.5%.
• This accuracy is significantly higher than the best individual predictor (pknotsRG) at 68.8%.
• Statistical tests confirmed the significance of the performance improvements.

Conclusions:

• The developed method offers a superior approach to RNA secondary structure prediction.
• This tool choice method acts as an effective preprocessing step for RNA sequence analysis.
• The findings highlight the benefits of integrating multiple prediction tools and advanced feature selection techniques.