Comparative genomic hybridization (CGH) is a method in cytogenetics for detecting gains and losses of DNA across a genome. Differently labeled DNA from a test sample and a reference sample is allowed to bind to the same set of genomic targets, and their relative signal intensities reveal differences in DNA abundance. Originally performed on metaphase chromosomes, CGH was subsequently adapted to DNA microarrays, greatly increasing its resolution. Its principal purpose is to detect genomic imbalance rather than determine DNA sequence or identify every type of chromosome abnormality.
Principle
CGH relies on nucleic acid hybridization: complementary DNA strands associate through sequence-specific base pairing. Test and reference DNA are labeled with distinguishable fluorescent dyes, mixed, and hybridized to a common target. At a given genomic location, the relative amount of bound test and reference DNA reflects the relative abundance of that sequence in the two samples.
A region present in more copies in the test sample generally produces an increased test-to-reference signal ratio; a region present in fewer copies produces a decreased ratio. Such differences are called copy-number variants or copy-number alterations, depending on the biological and analytical context. They can encompass entire chromosomes, chromosome arms, or smaller segments containing one or more genes.
Results are commonly represented as a log-ratio profile:
where and are the background-corrected and normalized test and reference signals at genomic position .
In an idealized comparison with a diploid reference:
- Two copies in both samples give a ratio of , or a log-ratio of .
- Three copies in the test sample versus two in the reference give a ratio of , or approximately .
- One copy in the test sample versus two in the reference gives a ratio of , or .
These values are theoretical expectations, not universal reporting thresholds. Actual measurements depend on sample composition, assay chemistry, background signal, and normalization. Most importantly, CGH measures relative genomic dosage, not an unconditional count of chromosomes or DNA molecules.
Main forms
Metaphase CGH
In conventional, or metaphase, CGH, labeled test and reference DNA are hybridized to normal metaphase chromosome spreads. These chromosomes provide an ordered physical representation of the genome. Fluorescence microscopy and image analysis measure the signal ratio along each chromosome.
The method does not require dividing cells from the test specimen: the metaphase chromosomes are supplied separately as the hybridization target. This was particularly useful for analyzing solid tumors, which can be difficult to investigate through conventional chromosome preparation.
Metaphase CGH typically detects imbalances spanning several megabases, with performance varying by region, alteration size, and assay quality. High-level amplifications can sometimes be recognized at smaller sizes than low-level gains or losses. Repetitive and technically difficult regions, including areas near centromeres, can be challenging to assess.
Array CGH
Array comparative genomic hybridization, usually written array CGH or aCGH, replaces metaphase chromosomes with a DNA microarray containing probes at known genomic locations. Early arrays commonly used large cloned DNA fragments; many later platforms use synthetic oligonucleotides.
Each probe reports relative dosage at a particular genomic interval. The distribution and spacing of probes determine which regions can be assessed and how precisely an alteration can be localized. Arrays may provide broadly distributed genome-wide coverage, dense coverage of selected clinically relevant regions, or both.
Array CGH generally offers much finer resolution than metaphase CGH. However, the smallest reliably detectable alteration is not simply the distance between adjacent probes: reliable calls usually require support from multiple probes and adequate signal quality. Breakpoints are therefore usually localized to intervals rather than determined at the level of individual nucleotides.
Laboratory workflow and data analysis
A typical array-CGH workflow includes:
- DNA extraction and quality assessment. DNA is isolated from the test material and reference material. Concentration, integrity, and sample suitability are assessed.
- Labeling. Test and reference DNA receive different fluorescent labels.
- Competitive hybridization. The labeled preparations are mixed and applied to the same array. Blocking DNA may be used to reduce binding by repetitive sequences.
- Washing and scanning. Unbound material is removed, and a scanner measures the two fluorescence channels.
- Normalization and quality control. Processing corrects for systematic effects such as dye-dependent differences and spatial variation.
- Segmentation and copy-number calling. A computational algorithm groups neighboring probes with similar ratios into segments and identifies candidate gains and losses.
- Interpretation. Findings are evaluated using their genomic location, size, gene content, reference data, and relevant sample information.
Segmentation is necessary because individual probes are noisy. However, its settings affect the balance between detecting small events and generating false-positive calls. Probe sequence specificity, repetitive DNA, and cross-hybridization can also influence results.
Reference selection matters. An apparent difference may arise from a variant in the reference rather than an alteration unique to the test sample. Sex-chromosome dosage must also be considered when test and reference samples differ in chromosomal sex.
Applications
Cancer research and tumor characterization
CGH was developed as a genome-wide approach to studying copy-number changes in cancer. Tumor genomes frequently contain deletions, chromosome gains, and amplifications. Comparing many tumors can identify recurrently altered regions and help locate genes involved in tumor development.
CGH measures DNA dosage, not gene expression. An amplified gene is not necessarily highly expressed, and a deletion does not by itself establish its functional consequences. Copy-number data are therefore often studied alongside sequence, RNA, protein, and pathological findings.
Tumor samples commonly contain mixtures of malignant and nonmalignant cells. Tumors may also contain genetically distinct subclones. Both features can reduce or complicate the measured signal.
Constitutional genomic analysis
Array CGH is used in constitutional cytogenetics to investigate genomic deletions and duplications associated with developmental disorders, congenital anomalies, and other genetic conditions. It can identify imbalances too small to be visible through conventional chromosome analysis. Its use also extends to prenatal and pregnancy-loss investigations in appropriate laboratory settings.
A detected copy-number change is not automatically disease-causing. Interpretation may consider dosage-sensitive genes, overlap with established genomic disorders, population frequency, inheritance, and the observed phenotype. Findings may be classified as pathogenic, likely pathogenic, of uncertain significance, likely benign, or benign.
Inheritance is informative but not conclusive: an inherited variant may still be clinically relevant, while a newly arisen variant is not necessarily pathogenic.
Other research uses
CGH can investigate genomic stability in cultured cells and compare DNA content among strains or related organisms. In comparisons between divergent genomes, however, reduced hybridization may reflect sequence divergence as well as lower copy number. Such experiments require a different interpretive framework from comparisons between closely matched human samples.
Limitations
Balanced rearrangements
CGH generally does not detect balanced translocations or inversions because these rearrangements can change chromosome structure without changing the amount of DNA. It may detect an accompanying deletion or duplication, but the dosage result alone does not reveal the complete structural arrangement.
Similarly, a duplication detected by CGH does not necessarily reveal where the extra material is located or its orientation. Conventional chromosome analysis, fluorescence in situ hybridization (FISH), or sequencing-based methods may provide complementary structural information.
Sequence-level changes
CGH is not designed to detect single-nucleotide substitutions or small insertions and deletions. Detecting such mutations generally requires DNA sequencing or another sequence-specific assay. Some densely targeted arrays can detect small exon-level dosage changes, but this does not make them general sequence-analysis methods.
Mosaicism and sample mixtures
Mosaicism occurs when a specimen contains cell populations with different genetic constitutions. Their DNA contributes to a combined signal, so an alteration present in only some cells produces a smaller shift than the same alteration present in every cell.
For example, if a fraction of otherwise diploid cells carries a single-copy deletion, the idealized average copy number is . Relative to a two-copy reference, the expected log-ratio is
As decreases, the signal approaches zero and becomes harder to distinguish from noise. Detection limits depend on the platform, alteration size, data quality, and analysis method; there is no single mosaicism threshold applicable to all CGH assays.
Ploidy and allele information
Because CGH compares relative dosage, a uniform increase in the number of complete chromosome sets can be obscured by normalization. Pure CGH may therefore fail to recognize some forms of polyploidy, even though it can detect relative chromosome imbalance such as many forms of aneuploidy.
Conventional CGH also provides no direct information about which alleles are present. It cannot, by itself, identify copy-neutral loss of heterozygosity or regions of homozygosity. These can be investigated using genotype-sensitive platforms.
Biological interpretation
Even a technically reliable copy-number call may have uncertain biological significance. Genomic imbalance is one class of genetic variation, and the absence of an abnormal CGH result does not establish that a genome is free of clinically or biologically important alterations.
Relationship to other genomic methods
Chromosomal microarray analysis is a broader category that includes array CGH, single-nucleotide polymorphism arrays, and combined platforms. These terms are therefore not exact synonyms. SNP arrays measure allele-specific signals as well as DNA abundance and can reveal allelic patterns that pure array CGH cannot.
FISH examines selected genomic targets and can show their location within chromosomes or cells, whereas CGH surveys dosage across many genomic regions. Polymerase chain reaction–based assays can assess selected copy-number changes but usually cover fewer loci.
Sequencing-based approaches infer copy number from read depth and other evidence. Depending on the assay, they can also identify sequence variants and structural rearrangements. Their capabilities, like those of CGH, depend on coverage, sample quality, and the analytical pipeline.
Historical development
Comparative genomic hybridization was introduced in 1992 by Anne Kallioniemi and colleagues as a method for mapping DNA sequence copy-number changes in solid tumors. Its central advance was to obtain a genome-wide view of relative gains and losses without requiring metaphase preparations from the tumor itself.
The development of array-based CGH in the late 1990s shifted the hybridization target from whole chromosomes to mapped DNA probes. Increasing probe density and improved analysis subsequently enabled much smaller alterations to be detected, expanding CGH from predominantly cancer research into constitutional genomic testing and other applications.