Single-cell Copy Number Variation Analysis
Also known as: scCNV analysis, single-cell CNV, scCNA analysis, single-cell copy number aberration analysis
Single-cell copy number variation (scCNV) analysis detects gains and losses of genomic segments within individual cells, enabling researchers to resolve intratumor heterogeneity, reconstruct clonal evolution, and distinguish malignant from normal cells at single-cell resolution. It can be applied to single-cell whole-genome sequencing data directly or inferred from read-depth signals in scRNA-seq or scATAC-seq experiments.
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 scCNV analysis when the research question concerns intratumor heterogeneity, clonal evolution, or the identification of malignant subpopulations in cancer samples. It is the method of choice when bulk CNV profiling has already identified broad alterations and the goal is to dissect which cells carry which events, or when distinguishing cancer cells from normal stromal cells in a mixed-cell scRNA-seq experiment. Do not apply this method when cells originate from diploid normal tissue with no expected structural variation — the signal-to-noise ratio will be too low for meaningful calls. Also avoid scCNV inference from scRNA-seq when sequencing depth is very shallow (fewer than ~2,000 genes detected per cell), as read-depth signals become too sparse for reliable segmentation.
Strengths & limitations
- Resolves intratumor heterogeneity at single-cell resolution, revealing subclonal architecture invisible to bulk methods.
- Enables reconstruction of clonal evolution and identification of founding versus late copy-number events.
- Can infer CNVs from scRNA-seq data without additional sequencing assays, using tools such as inferCNV or CopyKAT.
- Distinguishes malignant from normal cells in mixed-sample scRNA-seq experiments based on copy-number divergence from the diploid baseline.
- Compatible with fixed-cell and FFPE material when using dedicated low-input scWGS protocols.
- Shallow per-cell sequencing depth (typical for scWGS droplet protocols) limits resolution; small focal amplifications or deletions may be missed.
- CNV inference from scRNA-seq is indirect and dependent on gene density across the genome; gene-poor regions yield unreliable estimates.
- Defining a clean diploid reference cell population for normalization can be difficult in highly aneuploid or contaminated samples.
- Doublets (two cells captured together) can produce artifactual CNV profiles and must be removed carefully.
- Computational pipelines differ substantially in assumptions and outputs; results can be tool-dependent.
Frequently asked
Can I perform scCNV analysis without whole-genome sequencing — for example, from a standard 10x Genomics scRNA-seq experiment?
Yes. Tools such as inferCNV, CopyKAT, and Numbat infer copy-number states from gene expression read-depth patterns in scRNA-seq data. The resolution and sensitivity are lower than direct scWGS because inference depends on gene density and expression variability, but these methods can reliably detect large chromosomal gains and losses (arm-level and above) and are widely used to separate tumor from normal cells in mixed samples.
How do I choose a reference (normal) cell population for normalization?
Ideally, normal cells come from the same individual — for example, immune or stromal cells within the same tumor sample that are expected to be diploid. If no internal normal cells are available, a matched normal tissue sample or a panel of normal cells from a public dataset can be used. The key requirement is that the reference cells are truly diploid and not admixed with aneuploid cells.
What bin size should I use for scWGS data?
Common choices range from 200 kb to 1 Mb depending on coverage depth. A practical rule is to choose a bin size such that the genome-wide average read count per bin per cell is at least 10–20 reads. With very shallow coverage (less than 0.05x per cell, typical for high-throughput droplet protocols), bins of 500 kb to 1 Mb are more appropriate. Ginkgo and similar tools can automatically suggest a bin size given the observed coverage.
How many cells do I need for reliable clonal inference?
There is no universal minimum, but reliable clonal clustering typically requires at least a few hundred cells per sample. With fewer than 100 cells the statistical power to distinguish true subclones from noise is low. For very rare subclones (less than 5% frequency), several thousand cells may be needed to detect them with confidence.
Is scCNV analysis appropriate for non-cancer samples?
It can be applied in non-cancer contexts where somatic mosaicism is expected — for example, in studies of neuronal genome diversity, aging-related somatic mutation accumulation, or early embryonic development. In diploid somatic tissue without expected structural variation, the signal-to-noise ratio is typically too low for meaningful calls and the method is not recommended.
Sources
- Garvin, T., Aboukhalil, R., Kendall, J., Baslan, T., Atwal, G. S., Hicks, J., Wigler, M., & Schatz, M. C. (2015). Interactive analysis and assessment of single-cell copy-number variations. Nature Methods, 12(11), 1058–1060. link ↗
- Gao, R., Bai, S., Henderson, Y. C., Lin, Y., Schalck, A., Yan, Y., Kumar, T., Hu, M., Sei, E., Davis, A., Wang, F., Shaitelman, S. F., Wang, J. R., Chen, K., Moulder, S., Lai, S. Y., & Navin, N. E. (2021). Delineating copy number and clonal substructure in human tumors from single-cell transcriptomes. Nature Biotechnology, 39(5), 599–608. link ↗
How to cite this page
ScholarGate. (2026, June 3). Single-cell Copy Number Variation Analysis. ScholarGate. https://scholargate.app/en/bioinformatics/single-cell-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.
- Copy Number Variation AnalysisBioinformatics↔ compare
- Single-cell epigenome-wide association studyBioinformatics↔ compare
- Single-cell GWASBioinformatics↔ compare
- Single-cell RNA-seq analysisBioinformatics↔ compare
- Single-cell RNA-seq differential expressionBioinformatics↔ compare
- Variant CallingBioinformatics↔ compare