Gene expression rule discovery and multi-objective ROC analysis using a neural-genetic hybrid

Area of Science:

• Bioinformatics
• Computational Biology
• Systems Biology

Background:

• Microarray data offers deep insights into cellular biochemical mechanisms but extracting actionable information remains challenging.
• Current methods for analyzing gene expression data often require gene pre-filtering or pre-selection, potentially limiting discovery.

Purpose of the Study:

• To develop and evaluate a novel hybrid method for analyzing gene expression data.
• To optimize the trade-off between Artificial Neural Network (ANN) classifier accuracy (sensitivity, specificity) and model size (number of genes).
• To demonstrate the method's ability to discover biological rules and generate plausible results without prior gene selection.

Main Methods:

• A multi-objective genetic algorithm was employed to evolve a near-optimal solution.
• The algorithm optimized a hybrid model integrating Artificial Neural Networks (ANNs).
• The method was tested on four established gene expression datasets from existing literature.

Main Results:

• The hybrid method successfully balanced ANN classifier accuracy and the number of genes.
• The approach demonstrated rule discovery capabilities on the tested datasets.
• Results were biologically intelligible and plausible, validating the method's effectiveness.

Conclusions:

• The developed hybrid method offers an effective approach for extracting meaningful information from complex gene expression data.
• This technique eliminates the need for gene pre-filtering, simplifying the analysis pipeline.
• The method shows promise for advancing biological discovery through intelligent analysis of microarray data.