FuNeL

FuNeL infers functional networks from rule-based evolutionary classifiers (BioHEL) by extracting co-predictive gene interactions from expression data to provide a perspective complementary to gene co-expression networks.

Key Features:

• Rule-Based Inference: FuNeL applies the rule-based evolutionary classifier BioHEL to describe expression samples through sets of rules that identify groups of co-predictive genes.
• Complementary Biological Knowledge: FuNeL constructs co-prediction networks that capture functional relationships complementary to traditional gene co-expression networks.
• Evaluation and Validation: The protocol was tested on synthetic datasets and evaluated across eight real-world human cancer datasets to assess network relevance.
• Biological Relevance and Disease Associations: FuNeL networks are compared to gene co-expression networks of equal size generated by three methods and analyzed for enriched biological terms and relationships among known disease-associated genes.
• Case Study Validation: A prostate cancer case study demonstrated that FuNeL-captured biological knowledge is consistent with specialized literature and with an independent dataset not used in inference.

Scientific Applications:

• Systems biology and omics analysis: Inferring functional gene networks from high-throughput expression data to explore complex gene interactions.
• Disease research and biomarker discovery: Identifying biologically relevant network elements, disease-gene associations, and potential biomarkers or therapeutic targets in cancer.

Methodology:

FuNeL applies BioHEL to derive rules describing expression samples and extracts gene interactions from those rules to infer co-prediction functional networks, which are then compared to co-expression networks and analyzed for enriched biological terms and disease-associated gene relationships; evaluations include synthetic datasets, eight human cancer datasets, and a prostate cancer validation with an independent dataset.