Biologically weighted LASSO: enhancing functional interpretability in gene expression data analysis

Area of Science:

• Bioinformatics
• Computational Biology
• Genomics

Background:

• Feature selection is crucial for gene expression data analysis.
• Current methods often lack biological interpretability.
• Integrating prior biological knowledge can enhance analysis.

Purpose of the Study:

• To develop an integrative feature selection approach.
• To combine weighted LASSO with biological prior knowledge.
• To improve both predictive performance and biological interpretability.

Main Methods:

• Developed an embedded integrative approach for feature selection.
• Created a novel score of biological relevance.
• Integrated weighted LASSO with biological knowledge in a single step.

Main Results:

• The proposed approach identifies the most predictive genes.
• It significantly enhances the biological interpretability of results.
• Outperformed standard LASSO in experiments.

Conclusions:

• The integrative approach effectively balances predictive power and biological insight.
• This method offers a more interpretable alternative for gene expression analysis.
• Code is publicly available for reproducibility.