CNVkit

CNVkit infers and visualizes copy number variants (CNVs) and somatic copy number alterations (SCNAs) from targeted DNA sequencing data to provide genome-wide and exon-level copy number profiles for genomic analysis.

Key Features:

• Data sources: Uses both targeted reads and nonspecifically captured off-target reads from targeted DNA sequencing and massively parallel sequencing.
• Resolution within targets: Provides exon-level copy number resolution inside targeted regions.
• Genome-wide resolution: Achieves approximately 100-kilobase resolution genome-wide from platforms targeting as few as 293 genes, with signal in intronic and intergenic areas via off-target reads.
• Normalization: Normalizes read counts against a pooled reference to reduce sample-to-sample variability.
• Bias correction: Corrects read-depth biases associated with GC content, target footprint size and spacing, and repetitive sequences.
• Bias sources addressed: Accounts for variability introduced by target capture efficiency and library preparation that affect read depth.
• Read-depth analysis: Operates on sequencing read counts/read depth as the primary signal for copy number inference.
• Visualization and reporting: Produces visualizations and reports of identified copy number changes and significant features.
• Benchmarking: Has been evaluated against array comparative genomic hybridization (aCGH) for performance assessment.

Scientific Applications:

• Germline CNV detection: Identification of germline copy number variants relevant to syndromic conditions and genetic studies.
• Somatic CNV/SCNA analysis in cancer: Detection and characterization of somatic copy number alterations in cancer genomes from targeted sequencing data.
• Targeted sequencing studies: Generation of high-resolution copy number data from targeted re-sequencing efforts to support genomic research into genetic contributions to disease.
• Cross-platform comparison: Use in studies comparing sequencing-based CNV calls to array-based methods such as aCGH.

Methodology:

Leverages targeted and off-target reads, normalizes read counts against a pooled reference, corrects for GC content, target footprint size and spacing, and repetitive sequences, and infers copy number at exon-level within targets and at ~100-kilobase resolution genome-wide.