Why use it

Duplications are systematically harder to classify than deletions and are where unstructured judgement drifts most. Applying the published metric keeps the reasoning explicit and comparable between reviewers, and makes it obvious when a duplication is uncertain because the evidence is genuinely thin rather than because nobody wrote the reasoning down.

About

The ACMG/ClinGen technical standard scores a constitutional copy-number GAIN across the same five sections used for losses, but with a section 2 built around a different question. For a deletion the question is haploinsufficiency; for a duplication it is triplosensitivity — and established triplosensitive regions are far rarer, which is why duplications so often land in the uncertain band. Section 2 also handles a situation with no equivalent for losses: a duplication whose breakpoints fall inside a haploinsufficient gene, which can disrupt that gene even though copy number went up. As in the loss metric, point values are assigned rather than fixed — each criterion has a suggested value and an allowed range, and section 4 criteria are individually capped so no single well-published observation can dominate the total.

When to use

Classifying a duplication reported on chromosomal microarray, exome or genome sequencing
Deciding whether a duplication that interrupts a haploinsufficient gene should be treated as a loss-of-function event
Applying a laboratory judgement that deviates from a suggested point value, within the standard's allowed range
Re-evaluating a previously reported duplication after new case, segregation or population data
Teaching why duplications and deletions of the same interval do not get the same classification

Formula

Total = Section 1 + Section 2 + Section 3 + Section 4 + Section 5 Every scored criterion contributes an ASSIGNED value inside its allowed range; blank uses the published suggestion. Section 4 criteria are individually capped. Pathogenic ≥ 0.99 Likely pathogenic 0.90 to 0.98 Uncertain significance −0.89 to 0.89 Likely benign −0.90 to −0.98 Benign ≤ −0.99

Pearls & pitfalls

Points are assigned, not fixed. Each criterion has a suggested value and a range; leaving the assignment blank here uses the suggestion.
Triplosensitivity is established for far fewer regions than haploinsufficiency. A duplication of a well-known haploinsufficient gene is not evidence of pathogenicity on its own — 2A requires an established TS region, and the HI-gene-inside-a-gain case is 2H, which scores zero.
A duplication can still be a loss-of-function event. If both breakpoints fall inside one gene (2I), score it with the PVS1 decision tree, exactly as you would an intragenic deletion.
Criterion 2K (suggested 0.45) covers the other breakpoint scenario: one breakpoint inside an established HI gene, with a phenotype consistent with loss of function. If the phenotype is not consistent, 2J scores nothing.
Section 4 criteria are CAPPED. Five reported probands under 4A do not score 2.25 — the criterion tops out at 0.90 however many observations support it.
Section 4 covers cases reported in the literature and databases; section 5 covers the patient in front of you. A de novo observation in a published proband and a confirmed de novo in your own patient (5A) are different evidence, scored separately.
The gene-number tiers are more permissive than for losses: 0–34 genes suggests nothing, 35–49 suggests 0.45, and 50 or more suggests 0.90. Deletions cross the same suggestions at 25 and 35 genes.
A duplication identical in gene content to an established benign gain (2C) suggests −1.00 and is benign on its own. So does one smaller than a benign gain whose breakpoints do not interrupt genes (2D).
Inheritance from an apparently unaffected parent lowers the score but does not settle it — incomplete penetrance and variable expressivity look exactly like this.

Next steps

Pathogenic / Likely pathogenic: report, correlate with phenotype, offer genetic counselling and parental testing. VUS: do not use for clinical decision-making; parental studies and scheduled re-review are the routes out. Likely benign / Benign: generally not reported as causative. Record every criterion applied AND the value assigned to it — without both the total cannot be reproduced or challenged.

Evidence

Riggs ER, Andersen EF, Cherry AM, et al. Technical standards for the interpretation and reporting of constitutional copy-number variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics (ACMG) and the Clinical Genome Resource (ClinGen). Genet Med. 2020;22(2):245-257. Suggested values, allowed ranges and per-criterion caps follow the reference implementation at cnvcalc.clinicalgenome.org/cnvcalc/cnv-gain.

References

Riggs ER et al. Technical standards for the interpretation and reporting of constitutional copy-number variants: a joint consensus recommendation of ACMG and ClinGen. Genet Med. 2020;22(2):245-257. — link
Rehm HL et al. ClinGen — The Clinical Genome Resource. N Engl J Med. 2015;372(23):2235-2242. — link

ClinGen CNV Classification — Copy-Number GAIN

Calculator