A powerful procedure that controls the false discovery rate with directional information

Area of Science:

• Genetics
• Statistical genomics
• Bioinformatics

Background:

• Multiple testing in genetics often involves analyzing directional information (e.g., gene up/down-regulation).
• Existing False Discovery Rate (FDR) control methods are primarily P-value based and do not utilize this directional data.

Purpose of the Study:

• To introduce a novel procedure that leverages directional information in test statistics.
• To control the False Discovery Rate (FDR) in finite samples while incorporating sign information.
• To demonstrate the power advantage of the new procedure over existing methods.

Main Methods:

• Development of the signed-knockoff procedure.
• Application of the procedure to genetic data analysis.
• Comparison with existing FDR control methods via simulation studies.

Main Results:

• The signed-knockoff procedure effectively utilizes directional information.
• The proposed method demonstrates a power advantage in simulation studies.
• Successful application in two real genetic datasets, showing improved performance.

Conclusions:

• The signed-knockoff procedure offers a powerful new approach for FDR control in genetics.
• Incorporating directional information enhances the discovery power of multiple testing procedures.
• This method provides a valuable tool for analyzing gene expression and other directional biological data.