BFMDDT: A Decision-Tree-Based Gene Regulatory Network Inference From Multi-Type Datasets.

Area of Science:

• Computational Biology
• Bioinformatics
• Systems Biology

Background:

• Gene regulatory network (GRN) reconstruction is a complex challenge.
• Existing methods often rely on single data types, limiting comprehensive analysis.
• Integrating diverse datasets can enhance the accuracy of GRN inference.

Purpose of the Study:

• To propose a novel method, BFMDDT, for GRN reconstruction using multi-type datasets.
• To improve upon existing decision tree (DT) based GRN inference methods.
• To enhance the accuracy and efficiency of GRN reconstruction through data integration.

Main Methods:

• A novel data-integration strategy is employed to create distinct training sets from multi-type datasets.
• Ensemble decision tree (DT) models are utilized to determine gene regulatory relationships (both incoming and outgoing).
• Feature importance is extracted from DT models and combined to generate a global ranking for GRN construction.

Main Results:

• BFMDDT demonstrates effective information extraction from multi-type gene expression datasets.
• Experimental results validate the effectiveness of data integration for GRN reconstruction.
• The proposed method shows high accuracy, efficiency, and extensibility in GRN inference.

Conclusions:

• BFMDDT offers a robust approach for GRN reconstruction by leveraging multi-type data integration.
• The method's design allows for extensibility and combination with other computational techniques.
• BFMDDT significantly advances the field of GRN inference with improved performance.