Robust meta-analysis of gene expression using the elastic net.

Area of Science:

• Bioinformatics
• Genomics
• Computational Biology

Background:

• Gene expression meta-analysis offers biological insights but faces methodological limitations.
• Existing methods struggle to systematically integrate diverse gene expression datasets.

Purpose of the Study:

• To develop and validate a novel meta-analysis method for gene expression data using the elastic net.
• To apply this method to lung cancer gene expression data for subtype classification.

Main Methods:

• Developed a meta-analysis framework employing the elastic net for classification and regression.
• Utilized publicly available lung cancer gene expression data from five datasets (629 samples).
• Trained a multinomial classifier to distinguish between four lung cancer subtypes.

Main Results:

• The meta-analysis-derived classifier identified 58 key genes.
• Achieved 91% accuracy in leave-one-study-out cross-validation.
• Validated the classifier's performance on three independent datasets with 91% accuracy.

Conclusions:

• The elastic net-based meta-analysis provides a systematic approach to gene expression data integration.
• This method enhances the scope of questions addressable by meta-analysis.
• The approach is valuable for distilling large public gene expression datasets into actionable knowledge.