Increasing reproducibility, robustness, and generalizability of biomarker selection from meta-analysis using Bayesian...

Area of Science:

• Biostatistics
• Genomics
• Biomarker Discovery

Background:

• Gene expression biomarker studies often lack reproducibility in diverse populations.
• Frequentist meta-analysis is a common solution but has limitations like data requirements and outlier sensitivity.

Purpose of the Study:

• To develop and evaluate a Bayesian meta-analysis framework for gene expression data.
• To address the limitations of frequentist meta-analysis in biomarker studies.

Main Methods:

• Developed a Bayesian meta-analysis framework for gene expression data analysis.
• Compared the Bayesian framework with a frequentist approach using real-world data from three diseases.
• Created a publicly available R package for the Bayesian framework.

Main Results:

• The Bayesian method demonstrated greater robustness to outliers.
• It provided more informative estimates of between-study heterogeneity.
• Reduced false positive and false negative biomarkers, leading to more generalizable results with less data.

Conclusions:

• The Bayesian meta-analysis framework offers a more robust and accurate approach for gene expression biomarker discovery.
• This method enhances the generalizability of biomarkers identified from heterogeneous populations.
• The developed R package facilitates the application of this advanced statistical technique.