Differential Copy Number Variation Analysis — Comparative CNV / SCNA Analysis
Differential Copy Number Variation Analysis · Also known as: dCNV analysis, comparative CNV analysis, somatic copy number alteration analysis, SCNA analysis
Differential copy number variation (dCNV) analysis identifies genomic regions where DNA copy numbers differ systematically between two conditions — such as tumor versus normal tissue, case versus control cohorts, or treated versus untreated cells. By combining probe-level read-depth or array-intensity data with statistical segmentation and group-level testing, it pinpoints somatic amplifications and deletions that may drive disease, and distinguishes recurrent driver events from passenger noise across a cohort.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Use differential CNV analysis when you have genomic copy-number data from at least two groups — most commonly tumor versus matched normal, or a disease cohort versus healthy controls — and the goal is to identify recurrently amplified or deleted loci that differ between groups. It is the standard approach for discovering somatic copy-number alterations (SCNAs) in cancer genomics. Minimum viable cohort size for cohort-level methods like GISTIC is roughly 30 samples per group; smaller paired designs can use CBS plus per-sample comparison. Do not use this method when your data consist of only qualitative presence/absence genotyping calls without probe-level signal, when samples lack clear group assignments, or when the primary question is phasing or inheritance rather than somatic alteration — structural variant calling or linkage analysis would be more appropriate there.
Strengths & limitations
- Detects both focal (gene-level) and broad (arm/chromosome-level) copy-number differences, giving a comprehensive view of genomic instability.
- GISTIC2 and related tools provide statistically rigorous FDR-controlled identification of recurrent driver regions across cohorts.
- Applicable to multiple data modalities — array CGH, SNP arrays, whole-genome sequencing, and whole-exome sequencing — using consistent analytic logic.
- Segmentation reduces dimensionality from millions of probes to thousands of interpretable segments, enabling efficient downstream analysis.
- Paired tumor-normal designs eliminate most germline CNV noise, dramatically improving specificity for somatic events.
- Tumor purity and ploidy confound absolute copy-number estimation; a highly aneuploid or low-purity sample can mask or mimic true alterations.
- Spatial resolution is limited by probe density (arrays) or sequencing depth; small focal events below ~50 kb may be missed at standard depths.
- Cohort-based methods require reasonably large, well-matched groups — small or heterogeneous cohorts yield inflated false discovery rates.
- Germline CNVs must be carefully filtered; without matched normals, distinguishing somatic from inherited copy-number changes is unreliable.
Frequently asked
What is the difference between CNV detection and differential CNV analysis?
CNV detection identifies copy-number variants in one sample or population relative to a reference genome, and may include germline variants. Differential CNV analysis explicitly compares two conditions or groups — typically tumor versus normal or disease versus control — to find loci where copy number systematically differs between groups. The differential step adds group-level statistical testing on top of per-sample copy-number calls.
Do I need matched normal samples?
Matched normals (e.g., blood DNA from the same patient as the tumor) are strongly preferred because they allow direct subtraction of germline CNVs, leaving only somatic alterations. Without matched normals you can use a panel of normals or population CNV databases to filter common germline variants, but sensitivity and specificity for somatic calls will be lower, especially in regions that are polymorphic in the population.
How does GISTIC differ from simply counting altered samples at each locus?
GISTIC weights alterations by their amplitude as well as frequency, so a small number of samples with very high amplification at a locus can still score highly. It also models the expected background rate of random alterations to distinguish focal, likely driver events from the broad wave of arm-level aneuploidy that is common in cancer. The result is a q-value that reflects statistical significance of recurrence rather than a raw count.
Can I apply this pipeline to whole-exome sequencing data?
Yes, but with caveats. Exome capture is not uniform across the genome, so read-depth normalization must account for capture efficiency per target region. Coverage in intronic and intergenic regions is absent, limiting detection to exon-proximal regions. Focal events smaller than a single exon or spanning only intergenic sequence may be missed. Using a matched panel of normals captured on the same kit reduces these artifacts substantially.
What sample size is needed for a GISTIC analysis?
GISTIC2 was designed and validated on cancer cohorts typically numbering 50–500 samples. In practice, 30 samples is often cited as a minimum for meaningful recurrence analysis, but statistical power to detect alterations present in fewer than 10% of the cohort requires much larger numbers. For smaller studies, per-sample CBS segmentation with group-level Wilcoxon or Fisher tests applied segment-by-segment is a more reliable alternative.
Sources
- Olshen, A. B., Venkatraman, E. S., Lucito, R., & Wigler, M. (2004). Circular binary segmentation for the analysis of array-based DNA copy number data. Biostatistics, 5(4), 557–572. DOI: 10.1093/biostatistics/kxh008 ↗
- Mermel, C. H., Schumacher, S. E., Hill, B., Meyerson, M. L., Beroukhim, R., & Getz, G. (2011). GISTIC2.0 facilitates sensitive and confident localization of the targets of focal somatic copy-number alteration in human cancers. Genome Biology, 12(4), R41. DOI: 10.1186/gb-2011-12-4-r41 ↗
How to cite this page
ScholarGate. (2026, June 3). Differential Copy Number Variation Analysis. ScholarGate. https://scholargate.app/en/bioinformatics/differential-copy-number-variation-analysis
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Genome-wide association studyBioinformatics↔ compare