A New Genomic Wave Adjustment Method for DNA Copy Number Variation Detection

Area of Science:

• Genomics
• Bioinformatics
• Cancer Research

Background:

• DNA copy number variations (CNVs) are vital for understanding cancer genome alterations and DNA replication errors.
• Genomic waves in copy number data complicate accurate CNV detection across platforms.
• Existing algorithms struggle with the noise introduced by genomic wave patterns.

Purpose of the Study:

• To enhance the fused Lasso algorithm for precise DNA copy number variation detection.
• To develop a method that accounts for and corrects genomic wave patterns in copy number data.
• To improve the identification of change-points and genomic characteristics associated with wave patterns.

Main Methods:

• Proposed a partially linear model to differentiate piecewise-constant signals from genomic waves.
• Employed nonconvex penalized regression for multiple change-point identification.
• Validated the approach using simulated data and real-world bladder tumor aCGH, SNP genotyping, and breast tumor NGS data.

Main Results:

• The enhanced algorithm provides more precise estimators for the number and locations of change-points in the presence of genomic waves.
• Successfully applied the method to diverse cancer datasets (bladder, breast) for accurate CNV detection.
• Demonstrated the method's utility as a preprocessing step to correct biases in other CNV calling algorithms.

Conclusions:

• The proposed method effectively addresses genomic waves, significantly improving DNA copy number variation detection accuracy.
• This approach enhances the reliability of CNV analysis in cancer genomics.
• The algorithm offers a robust tool for cancer genome research and diagnostic applications.