FWER and FDR control when testing multiple mediators.

Area of Science:

• Biostatistics
• Genomics
• Epidemiology

Background:

• Understanding biological pathways between environmental exposures and disease risk is crucial but often challenging.
• Biomarker mediation analysis is a key approach to elucidate these complex relationships.
• Existing methods face limitations in simultaneously analyzing numerous biomarkers and controlling error rates.

Purpose of the Study:

• To develop a novel Multiple Comparison Procedure (MCP) for identifying mediating biomarkers.
• To control the Family-Wise Error Rate (FWER) or False Discovery Rate (FDR) in high-dimensional biomarker studies.
• To enhance statistical power for detecting true biological mediators.

Main Methods:

• Proposed a new Multiple Comparison Procedure (MCP) for mediation analysis.
• Simultaneously tested hypotheses for a set of 'm' biomarkers.
• Validated the method through simulations and applied it to a real-world breast cancer dataset.

Main Results:

• The proposed MCP demonstrated higher statistical power in detecting true mediators compared to existing methods.
• The method effectively controlled the Family-Wise Error Rate (FWER) or False Discovery Rate (FDR).
• Identified nine specific metabolites potentially mediating the association between BMI and breast cancer risk.

Conclusions:

• The developed MCP offers a powerful and reliable tool for biomarker mediation analysis in exposure-disease studies.
• The findings provide novel insights into the biological pathways of breast cancer.
• The R package 'MultiMed' is available for implementing this advanced statistical approach.