View Categories

ACMG/AMP Framework in SeqSMART

5 min read

Introduction #

The ACMG/AMP framework defines the international standard for interpreting the clinical significance of genomic variants.
SeqSMART implements this framework as the foundation of its variant interpretation engine — combining automation, transparency, and expert oversight.

While SeqSMART adheres strictly to the official ACMG/AMP guidelines, it introduces an innovative simplified structure and interactive user interface to make variant classification more efficient and intuitive for professional users.

1. The ACMG/AMP Classification Concept #

The ACMG/AMP guidelines define five major categories of clinical significance:

ClassificationDescription
Pathogenic (P)Variant is clearly disease-causing.
Likely Pathogenic (LP)Variant has strong evidence suggesting pathogenicity.
Variant of Uncertain Significance (VUS)Evidence is currently insufficient or conflicting.
Likely Benign (LB)Variant has strong evidence against pathogenicity.
Benign (B)Variant is conclusively non-disease-causing.

Each classification is derived from evaluating 28 specific criteria categorized by their evidence strength and direction (pathogenic or benign).
SeqSMART operationalizes these criteria through data-driven automation and guided expert review.

2. SeqSMART’s Simplified Evidence-Level Structure #

To streamline variant interpretation, SeqSMART organizes the 28 ACMG criteria into eight evidence-level groups.
This approach helps users quickly locate relevant information and understand the context of each criterion within the classification process.

Evidence GroupCriteriaDescription
1. Population DataBA1, BS1, BS2, PM2, PS4Evaluates allele frequency in population databases and case enrichment.
2. Computational Predictive DataPP3, BP4, BP7Assesses computational predictions for variant impact and conservation.
3. Functional StudiesBS3, PS3Considers experimental data from functional assays supporting or refuting pathogenicity.
4. Segregation StudiesPP1, PM3, BS4, BP2Evaluates inheritance consistency across family members and co-occurrence data.
5. De Novo and Inheritance PatternPS2, PM6Reviews confirmed or assumed de novo variants and inheritance compatibility.
6. Variant Type and LocationPM1, PM5, PS1, PVS1, PP2, BP1, BP3, PM4Analyzes variant consequence, position within critical domains, and loss-of-function relevance.
7. Previous EvidencesPP5, BP5References previously reported classifications and curated interpretations from trusted sources.
8. Phenotypic EvidencesPP4Integrates phenotype–genotype correlations based on HPO and OMIM data.

This organization allows experts to navigate evidence categories more logically and efficiently — mirroring real-world interpretation workflows used in clinical genomics.

3. Visual Overview of Criteria #

At the top of each variant detail page, SeqSMART displays an interactive ACMG summary bar showing all 28 criteria grouped by their evidence type.

  • Each criterion is marked as met, unmet, or pending.
  • Users can hover or click on a criterion to explore detailed evidence, data sources, and decision logic.
  • Color-coded indicators visually distinguish pathogenic-supporting vs. benign-supporting evidence.

This design gives users an immediate understanding of the classification status while browsing variant-level details, reducing the need for manual tracking or external reference tables.

4. Evidence Strength and Integration #

Each ACMG criterion carries a defined strength level — Very Strong, Strong, Moderate, or Supporting — contributing quantitatively to the final classification.

Evidence TypeStrengthPrefixExample
PathogenicVery StrongPVSPVS1 – Null variant in LOF gene
PathogenicStrongPSPS1 – Same amino acid change as known pathogenic variant
PathogenicModeratePMPM2 – Absent from control populations
PathogenicSupportingPPPP3 – Computational evidence supports deleterious effect
BenignStrongBSBS1 – Allele frequency too high for disorder
BenignSupportingBPBP4 – Computational prediction indicates benign

SeqSMART applies a weighted logic model consistent with ACMG/AMP combination rules, calculating the cumulative evidence dynamically and transparently.

5. Automated vs. Expert Layers #

SeqSMART divides ACMG evaluation into two complementary layers:

Automated Layer #

Automatically determines criteria that can be supported through data-driven logic, such as:

  • Population frequency thresholds (BA1, BS1, PM2)
  • Computational predictions (PP3, BP4)
  • Known variant recurrence (PS1, PM5)
  • Functional or domain-based rules (PVS1, PM4)

Expert Layer #

Empowers analysts to assess criteria that require interpretation or contextual evidence:

  • Family segregation or de novo confirmation (PP1, PS2, PM6)
  • Phenotype matching and literature review (PP4, PP5)
  • Experimental data evaluation (PS3, BS3)

All analyst decisions, modifications, and justifications are logged automatically for traceability and quality control.

6. Phenotype Integration #

SeqSMART uniquely integrates phenotypic evidence into ACMG evaluation using curated HPO and OMIM data.
For each case, the platform:

  • Matches patient phenotypes to associated genes and disorders.
  • Assesses consistency between clinical presentation and variant function.
  • Applies PP4 automatically when phenotype–genotype alignment is statistically or clinically significant.

This tight integration ensures that variant classification always reflects the clinical context.

7. Transparency and Reproducibility #

Every ACMG decision in SeqSMART is fully transparent. Users can view:

  • The data source, logic, and confidence score for each rule.
  • Whether a rule was automatically applied or manually adjusted.
  • The version of the ACMG logic used for the classification.

All of this information is stored in the case audit trail, ensuring complete reproducibility for institutional audits or future reanalysis.

8. Customization and Institutional Alignment #

SeqSMART’s ACMG framework is fully customizable at the institutional level.
Administrators can:

  • Modify thresholds (e.g., population frequency cutoffs for BS1/PM2).
  • Enable or disable selected criteria based on laboratory policy.
  • Adjust rule weighting or combine rules into institution-specific templates.
  • Save custom ACMG schemes for consistent team-wide interpretation.

This flexibility allows SeqSMART to serve both clinical diagnostic and research-oriented applications without compromising compliance or reproducibility.

9. Continuous Updates #

SeqSMART’s ACMG engine evolves continuously to incorporate the latest scientific consensus and database updates.
Each platform version clearly documents:

  • Adjustments in rule logic or thresholds.
  • Integration of new reference datasets or tools.
  • Compatibility with the official ACMG/AMP recommendations.

All updates are version-controlled and listed in the Release Notes, ensuring backward traceability of every classification.

The ACMG/AMP framework in SeqSMART combines the full rigor of official guidelines with an intuitive, structured, and transparent implementation.
By dividing the 28 ACMG criteria into eight evidence-level groups and providing clear visual feedback for each variant, SeqSMART transforms complex genomic interpretation into a manageable and verifiable process.

SeqSMART ACMG Framework Principle:
 Scientifically standardized — practically simplified.