CNVkit

CNVkit is a Python-based open-source toolkit for copy number variation analysis from targeted DNA sequencing data. It is commonly used in cancer genomics to identify amplifications and deletions from exome, targeted panel, or amplicon sequencing experiments. By leveraging depth-of-coverage information, CNVkit aims to provide robust copy number estimates even when the captured regions are sparse or unevenly covered.

The core approach involves computing read depth over defined target intervals and, when available, over off-target regions to enable broader genome-wide inference. The sample coverage is compared to a reference built from normal samples or a synthetic reference, producing log2 copy-number ratios. The workflow includes steps to correct for systematic biases such as GC content and target density, followed by segmentation to delineate contiguous regions with similar copy number. CNVkit supports multiple segmentation strategies and can yield allele-specific copy number estimates if SNP data are provided.

Output from CNVkit includes per-bin copy-number estimates and segmented copy-number profiles, which can be exported to standard formats for downstream analysis. Built-in visualization tools generate plots such as chromosome-wide heatmaps, genome-wide copy-number plots, and region-focused charts to aid interpretation. The software is designed to work with typical BAM inputs and BED-style target interval definitions, enabling straightforward integration into existing sequencing analysis pipelines.

CNVkit is released as an open-source project with a command-line interface and is suitable for researchers who need a practical, scalable solution for detecting copy number variation from targeted sequencing data.