DTFP-Growth: Dynamic Threshold-Based FP-Growth Rule Mining Algorithm Through Integrating Gene Expression,...

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Association rule mining is crucial for uncovering gene pair relationships in biological data.
• Existing methods often struggle with single data sets and global thresholds, limiting multi-view analysis.

Purpose of the Study:

• To propose a dynamic threshold-based FP-growth algorithm for multi-view association rule mining.
• To identify novel gene pair associations by integrating diverse biological data profiles.

Main Methods:

• Developed a dynamic threshold-based FP-growth algorithm integrating gene expression, methylation, and protein-protein interaction data.
• Introduced Distance-based Variable Supports (DVS), Confidences (DVC), and Lifts (DVL) for rule evaluation.
• Computed and applied individual thresholds for each rule based on weighted shortest distance.

Main Results:

• The proposed algorithm successfully integrates multiple omics data types.
• It considers both quantitative and interactive significance among gene pairs.
• The method generates fewer rules with greater biological significance and reduced running time compared to existing approaches.

Conclusions:

• The dynamic threshold approach offers a more effective method for association rule mining in multi-omics datasets.
• This algorithm enhances the discovery of biologically relevant gene pair interactions.