CNV-FB: A Feature bagging strategy-based approach to detect copy number variants from NGS data

Area of Science:

• Genomics
• Bioinformatics

Background:

• Copy number variation (CNV) is a significant genomic structural variation linked to human disease pathogenesis.
• Next-generation sequencing (NGS) technology facilitates CNV detection, yet challenges remain for low-purity and low-coverage data.

Purpose of the Study:

• To introduce CNV-FB, a novel computational method for detecting CNVs from NGS data.
• To address the limitations of existing CNV detection algorithms, particularly in challenging data scenarios.

Main Methods:

• CNV-FB employs random sampling of read depth values from genome fragments.
• Outlier detection is performed on individual samples, with results aggregated into a final outlier score.

Main Results:

• CNV-FB demonstrated superior performance compared to five other CNV detection methods.
• The method was validated using both simulated and real-world NGS datasets.

Conclusions:

• CNV-FB shows promise as an effective algorithm for identifying genomic mutations.
• The proposed method offers an improved approach for CNV detection in bioinformatics.