Identification of differentially expressed genes using multi-resolution wavelet transformation analysis combined with...

Area of Science:

• Bioinformatics
• Genomics
• Statistical Genetics

Background:

• Identifying differentially expressed genes is crucial for understanding biological processes and disease mechanisms.
• Existing statistical methods often struggle to balance sensitivity with the control of false discovery rates (FDR).
• The need for more robust and efficient methods for gene expression analysis persists.

Purpose of the Study:

• To develop and evaluate an improved statistical method for identifying differentially expressed genes.
• To enhance the accuracy of gene expression analysis by effectively managing noise and controlling the FDR.
• To apply the proposed method to a real-world microarray dataset.

Main Methods:

• A novel approach combining multi-resolution wavelet transformation with Significance Analysis of Microarrays (SAM).
• Adjustment of the delta (Δ) parameter and computation of the false discovery rate (FDR).
• Application to a microarray dataset comparing adenoma patients and normal subjects.

Main Results:

• The wavelet transformation method, combined with SAM, identified a greater number and higher quality of differentially expressed genes.
• Increasing the delta (Δ) value led to a gradual reduction in the number of identified differentially expressed genes.
• Wavelet transformation significantly reduced the false discovery rate (FDR) at various delta (Δ) values.

Conclusions:

• The proposed wavelet transformation and SAM method offers superior performance in detecting differentially expressed genes compared to non-transformed data.
• This approach provides more controlled and reduced false discovery rates (FDR), increasing the reliability of findings.
• The method is effective for analyzing gene expression data, particularly in complex biological samples like those from adenoma patients.