iDRBP-EL: Identifying DNA- and RNA- Binding Proteins Based on Hierarchical Ensemble Learning

Area of Science:

• Bioinformatics
• Computational Biology
• Molecular Biology

Background:

• Accurate identification of DNA-binding proteins (DBPs) and RNA-binding proteins (RBPs) is crucial for understanding protein-nucleic acid interactions.
• Existing machine-learning methods often use limited information and struggle to predict both DBPs and RBPs simultaneously.

Purpose of the Study:

• To develop a computational predictor, iDRBP-EL, capable of identifying both DBPs and RBPs from protein sequences.
• To enhance prediction performance by integrating multiple levels of information through hierarchical ensemble learning.

Main Methods:

• Proposed iDRBP-EL, a multi-label model integrating diverse features, machine learning algorithms, and data.
• Employed hierarchical ensemble learning to combine information from different sources.
• Conducted ablation experiments to validate the contribution of fused information.

Main Results:

• The fusion of different information sources significantly improved prediction performance.
• The method successfully overcame the cross-prediction problem (predicting DBPs and RBPs simultaneously).
• iDRBP-EL demonstrated superior performance compared to existing methods on independent datasets.

Conclusions:

• iDRBP-EL offers an effective and accurate approach for identifying both DBPs and RBPs using only protein sequences.
• The developed webserver provides a user-friendly platform for DBP and RBP prediction.
• Hierarchical ensemble learning is a powerful strategy for integrating multi-source information in bioinformatics prediction tasks.