Network-Based Copy Number Variation Analysis
Also known as: network CNV analysis, CNV network propagation, graph-based CNV analysis, network-integrated copy number analysis
Network-based copy number variation analysis integrates genome-wide CNV data with biological interaction networks — such as protein-protein interaction (PPI) or pathway networks — to identify functionally coherent regions, driver genes, and altered subnetworks that raw CNV calling alone would miss. By propagating CNV signals through the network graph, the method reveals coordinated genomic dosage imbalances that converge on common biological functions, making it especially powerful in cancer genomics and rare-disease studies.
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 network-based CNV analysis when you have genome-wide copy number profiles from a cohort of samples — particularly cancer specimens or patients with genomic diseases — and want to identify biologically meaningful driver alterations beyond the most recurrent focal events. It is well-suited to situations where individual CNV events are heterogeneous across patients (low recurrence) but converge on common pathways. A reliable biological network for the organism under study is required; results are only as informative as the network coverage. Do NOT use this method if you have only a handful of samples (fewer than ~20 for cancer studies), if you lack a suitable interaction network, if your primary goal is precise breakpoint mapping (use structural variant calling instead), or if the CNV data are low-resolution array data that cannot distinguish individual gene-level alterations reliably.
Strengths & limitations
- Identifies functionally coherent driver modules even when individual gene alterations are rare across the cohort.
- Integrates prior biological knowledge (PPI, pathway databases) to focus on interpretable, mechanistically relevant findings.
- Reduces the multiple-testing burden compared to gene-by-gene testing by collapsing many signals into subnetwork units.
- Compatible with multi-omics extension — CNV scores can be combined with mutation or expression data in the same propagation framework.
- Applicable to both somatic (cancer) and germline (rare disease) CNV datasets.
- Results are highly sensitive to the choice and completeness of the background network; poorly covered genes or tissues will be under-detected.
- Network propagation smooths over heterogeneous alteration patterns, potentially merging distinct biological lesions into a single module.
- Requires a moderately large cohort to generate stable recurrence statistics; very small cohorts yield unstable subnetwork scores.
- Computationally intensive permutation testing can be prohibitive for very large networks without high-performance computing resources.
Frequently asked
Which network should I use as the propagation substrate?
The choice matters considerably. Curated PPI databases such as STRING, BioGRID, or HINT are common defaults. For cancer studies, tissue-specific or cancer-specific networks derived from co-expression data can improve specificity. Pathway-based networks (KEGG, Reactome) give more interpretable but sparser results. Always report which network version was used and check sensitivity to network choice as part of your analysis.
How is this different from standard GISTIC2 analysis?
GISTIC2 identifies statistically recurrent focal and broad CNV events at the genomic coordinate level — it is excellent for finding the most frequently altered peaks. Network-based analysis adds a second layer: it asks whether genes that are individually altered at moderate frequency form functionally connected modules. The two methods are complementary; GISTIC2 findings can serve as input scores for network propagation.
What sample size do I need?
For stable subnetwork scores in cancer cohorts, at least 50–100 samples is recommended. Studies with 20–50 samples can yield exploratory findings but should interpret results cautiously. Below 20 samples the recurrence statistics are unreliable and the method is not appropriate; standard single-sample CNV characterisation is preferable.
Can I apply this to germline CNV data from case-control studies?
Yes, but with modifications. Instead of recurrence frequency across tumours, you score genes by case-vs-control enrichment of CNV burden. The propagation and subnetwork steps are the same, but the null permutation must preserve case-control labels to estimate false discovery correctly. Tools such as NETBAG and network-constrained burden tests have been developed for this setting.
Does the method account for the size of the CNV?
It depends on the scoring scheme. Gene-level scores from GISTIC2 already incorporate segment size and amplitude. If you use simple recurrence frequency as the input weight, large CNVs affecting many genes will inflate scores for all contained genes indiscriminately — applying a focal-event filter or using amplitude-weighted scores before propagation is recommended to mitigate this.
Sources
- Vandin, F., Upfal, E., & Raphael, B. J. (2012). De novo discovery of mutated driver pathways in cancer. Genome Research, 22(2), 375–385. DOI: 10.1101/gr.120477.111 ↗
- Leiserson, M. D. M., Vandin, F., Wu, H.-T., Dobson, J. R., Eldridge, J. V., Thomas, J. L., Papoutsaki, A., Kim, Y., Niu, B., McLellan, M., Lawrence, M. S., Gonzalez-Perez, A., Tamborero, D., Cheng, Y., Ryslik, G. A., Lopez-Bigas, N., Getz, G., Ding, L., & Raphael, B. J. (2015). Pan-cancer network analysis identifies combinations of rare somatic mutations across pathways and protein complexes. Nature Genetics, 47(2), 106–114. DOI: 10.1038/ng.3168 ↗
How to cite this page
ScholarGate. (2026, June 3). Network-Based Copy Number Variation Analysis. ScholarGate. https://scholargate.app/en/bioinformatics/network-based-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
- Gene Set Enrichment AnalysisBioinformatics↔ compare
- Network-based GWASBioinformatics↔ compare
- Network-based RNA-seq differential expressionBioinformatics↔ compare
- Pathway Enrichment AnalysisBioinformatics↔ compare
- Variant CallingBioinformatics↔ compare