CopyDetective: Detection threshold-aware copy number variant calling in whole-exome sequencing data

Area of Science:

• Genomics
• Bioinformatics
• Cancer Research

Background:

• Copy number variants (CNVs) are significant in disease development and progression.
• Detecting CNVs using whole-exome sequencing (WES) data presents considerable challenges.
• Existing methods often require supplementary experimental validation.

Purpose of the Study:

• To introduce CopyDetective, a novel algorithm for somatic copy number variant (CNV) calling.
• To enhance the accuracy and reliability of CNV detection in matched whole-exome sequencing (WES) data.
• To address the limitations of current CNV detection methods in WES.

Main Methods:

• CopyDetective employs a two-step procedure for CNV calling.
• Step 1: Quality analysis to establish individual detection thresholds for each sample.
• Step 2: CNV calling based on determined thresholds, evaluating variant allele frequency changes and reporting affected cell fractions.

Main Results:

• CopyDetective demonstrated superior performance in somatic CNV calling compared to established tools (ExomeCNV, VarScan2, ControlFREEC, ExomeDepth, CNV-seq).
• Analysis of four WES datasets (n=100) validated the algorithm's effectiveness.
• The algorithm accurately reports the fraction of cells affected by each CNV.

Conclusions:

• Not all WES datasets are equally suitable for CNV calling, highlighting the need for quality assessment.
• CopyDetective's initial quality analysis and individualized detection thresholds are crucial for accurate variant calling.
• Implementing quality analysis prior to CNV calling is recommended for robust results.