Identification of differential aberrations in multiple-sample array CGH studies.

Area of Science:

• Genomics
• Bioinformatics
• Statistical Genetics

Background:

• Array comparative genomic hybridization (aCGH) is crucial for identifying genomic alterations.
• Existing aCGH analysis methods often focus on single samples, limiting comparative studies.
• Detecting common and distinct copy number variations across sample groups is essential for understanding disease heterogeneity.

Purpose of the Study:

• To develop a novel statistical method for analyzing multiple aCGH samples from distinct groups.
• To simultaneously identify shared and group-specific aberrant genomic regions.
• To improve the accuracy of aberration detection compared to single-sample methods.

Main Methods:

• A penalized regression approach utilizing a fused adaptive lasso penalty was developed.
• The method accounts for spatial dependence among genomic clones.
• Significance testing between neighboring clones and segments identifies aberrant regions.

Main Results:

• The proposed method accurately detects common aberrant regions within groups.
• It also identifies regions with differing copy number changes between groups.
• Simulation studies demonstrated superior performance over single-sample methods in reducing false positives and negatives.

Conclusions:

• The new penalized regression method offers a robust approach for comparative aCGH analysis.
• It effectively identifies both shared and unique genomic aberrations across sample groups.
• The method shows promise for applications in cancer subtyping, such as breast cancer grade subgroups.