Select all applicable evidence criteria. The framework combines evidence strength to generate a 5-tier variant classification (P → LP → VUS → LB → B). This is a decision-support tool — final classification requires expert molecular pathologist review.
Pathogenic Evidence Criteria
Benign Evidence Criteria
Guidelines & Evidence
Clinical Details
Section 1
When to Use
When to Use
Interpreting variants identified in clinical genetic testing panels, exomes, or whole-genome sequencing.
Classifying germline variants from hereditary cancer, cardiovascular, neurological, or metabolic gene panels.
Structuring evidence for VUS (Variant of Uncertain Significance) review and reclassification decisions.
Standard framework used by all ACMG-accredited clinical laboratories and required for CAP/CLIA certification.
Section 2
Formula & Logic
5-Tier Classification
Class
Designation
Clinical Action
5
Pathogenic (P)
Report; act on clinically
4
Likely Pathogenic (LP)
Report; act on clinically (≥90% probability pathogenic)
3
Variant of Uncertain Significance (VUS)
Report; do NOT act on alone; family studies
2
Likely Benign (LB)
Do not report as actionable; may include as note
1
Benign (B)
Do not report
Evidence Criteria (Select)
Code
Criterion
Strength
PVS1
Null variant (LOF) in gene where LOF is a mechanism
Very Strong Pathogenic
PS1
Same amino acid change as established pathogenic variant
Strong Pathogenic
PS2
De novo (confirmed) in patient with disease, no family history
Strong Pathogenic
PS3
Functional assay shows damaging effect
Strong Pathogenic
PM2
Absent or low freq in gnomAD (< 0.1% for recessive, < 0.01% for dominant)
Moderate Pathogenic
PM5
Novel missense at position previously established as pathogenic
Moderate Pathogenic
PP3
Multiple computational tools predict pathogenic
Supporting Pathogenic
BS1
Allele frequency too high for disorder
Strong Benign
BA1
MAF > 5% in gnomAD
Stand-Alone Benign
Classification Rules (Combining Evidence)
Pathogenic: (1 Very Strong + ≥1 Strong) OR (≥2 Strong) OR (1 Strong + ≥3 Moderate) OR (1 Strong + 2 Moderate + ≥2 Supporting)
Likely Pathogenic: (1 Very Strong + 1 Moderate) OR (1 Strong + 1–2 Moderate) OR (≥3 Supporting Pathogenic criteria)
VUS: Does not meet LP or LB criteria; conflicting or insufficient evidence
Benign/Likely Benign: ≥1 Stand-Alone Benign, or ≥2 Strong Benign, or (1 Strong + 1 Supporting) Benign
Section 3
Pearls/Pitfalls
Key Pearls
VUS ≠ negative — patients and clinicians must understand a VUS cannot be used to guide management; avoid acting on it as if pathogenic.
VUS reclassification rates: ~10–20% of VUS are reclassified annually; recontact protocols are essential.
Gene-specific rules matter — BRCA1/2, TP53, RB1 LOF interpretation differs from MYH7 or SCN5A.
ClinGen gene curation status determines if gene-disease association is gene-specific (e.g., BRCA1-HBOC vs. BRCA1-Fanconi Anemia).
Important Limitation
The ACMG/AMP framework applies to germline variant classification only. Somatic (tumour) variants use separate classification systems (e.g., AMP/ASCO/CAP 2017 Somatic Variant Classification). Do not conflate the two.
Section 4
Next Steps
Clinical Actions
01
P or LP variant: Disclose to patient with genetic counselling; initiate surveillance/intervention per gene-specific guidelines (e.g., NCCN, ACMG guidelines).
02
VUS: Counselling to avoid anchoring bias; family segregation studies where possible; schedule reclassification review in 12–24 months.
03
All positive/LP findings: Offer cascade testing to 1st degree relatives at 50% risk.
04
LB/Benign: Disclose as non-actionable; reassurance without surveillance modification.
Section 5
Evidence Appraisal
Primary Reference
Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology
Richards S et al. • Genetics in Medicine. 2015;17(5): 405–424
Section 6
Literature
Development
Published in 2015 by a joint ACMG/AMP working group to replace an earlier 3-tier classification (disease-causing, unclassified, polymorphism) that had become inadequate as clinical sequencing expanded. The 2015 framework introduced 28 coded evidence criteria, combining evidence systematically to generate reproducible variant classifications across laboratories.
Evolution
ClinGen has since developed gene-specific variant interpretation guidelines (e.g., BRCA1/2, TP53, RASopathies, Channelopathies) that modify the 2015 framework. Key updates include the Bayesian statistical framework (Tavtigian et al., Genetics in Medicine 2018) formalising evidence weighting as odds of pathogenicity, and the PVS1 null variant decision tree (Abou Tayoun 2018). Over 2 million variants are now classified in ClinVar.
