Using ERDS to infer copy-number variants in high-coverage genomes

Area of Science:

• Genomics
• Bioinformatics

Background:

• Accurate inference of copy-number variants (CNVs) from next-generation sequencing (NGS) data is crucial for understanding genetic variation and disease.
• Existing CNV detection methods vary in computational efficiency, sensitivity, and specificity across different CNV types and coverage levels.

Purpose of the Study:

• To introduce and evaluate a novel computational method, estimation by read depth with single-nucleotide variants (ERDS), for CNV detection.
• To compare the performance of ERDS against established CNV calling algorithms using various metrics.

Main Methods:

• ERDS utilizes read depth information combined with single-nucleotide variant data for CNV inference.
• Performance comparisons involved assessing sensitivity and specificity across common and rare CNVs in high-coverage genomes.
• The method was evaluated for its ability to handle unique and amplified genomic regions without requiring separate alignments.

Main Results:

• ERDS demonstrates performance comparable to the leading method (Genome STRiP) for common CNVs in high-coverage genomes.
• ERDS outperforms all tested methods for detecting rare CNVs in high-coverage genomes.
• The method efficiently processes unique and amplified genomic regions, integrating CNV calling with other variant analyses.

Conclusions:

• ERDS offers a computationally efficient and accurate approach for CNV detection in high-coverage genomes.
• This method provides a valuable tool for genetic research, particularly for identifying rare CNVs.
• ERDS simplifies the analysis pipeline by not requiring separate alignments for different variant types.