Home Molecular Testing Methods Variant Classification in Genetic Testing: Pathogenic, Benign, VUS, and Results

Variant Classification in Genetic Testing: Pathogenic, Benign, VUS, and Results

By
Vita Library
-
31
Genetic variant classification labels findings as pathogenic, likely pathogenic, VUS, likely benign, or benign. Learn the evidence behind each category and reclassification.

Variant classification is the process laboratories use to decide whether a DNA change is likely to cause disease. Most clinical germline reports use five categories: pathogenic, likely pathogenic, variant of uncertain significance, likely benign, and benign. These labels summarize the strength of evidence about the variant—not the severity of a person’s symptoms and not a guarantee of what will happen in the future.

Classification requires more than a computer prediction. Specialists evaluate population frequency, the type and location of the change, laboratory studies, affected families, inheritance, and the established disease mechanism of the gene. Evidence can conflict or be incomplete, which is why a variant of uncertain significance, or VUS, is common in broad genetic testing. A VUS is not a positive diagnosis and should not usually direct irreversible medical decisions. Classifications can change as new data become available, making the report date and laboratory’s reinterpretation policy important.

  • The five categories describe evidence for or against disease causation.
  • Pathogenic and likely pathogenic findings can be clinically actionable when they fit the gene, condition, and inheritance pattern.
  • A VUS means uncertainty; it does not mean “probably harmful.”
  • Benign and likely benign variants are not considered causes of the tested disorder.
  • Reclassification may occur when population, family, functional, or gene-specific evidence improves.

Table of Contents

Classification Is an Evidence Judgment

Every person’s genome contains millions of differences from a reference sequence. Most are harmless. The challenge in clinical genetics is to identify the small subset that disrupts a gene in a way that causes or contributes to disease.

A variant classification answers a focused question: how strongly does the available evidence support that this DNA change is disease-causing for a specified gene-condition relationship? It does not directly answer whether the person has the condition. That second question requires clinical correlation, the correct inheritance pattern, and sometimes more than one variant.

For example, a pathogenic variant in a recessive disease gene may indicate carrier status when only one copy is affected. The same classification can confirm disease if a second pathogenic variant is present on the other chromosome and the phenotype fits. A pathogenic variant in a dominant gene may confer substantial risk, but penetrance can be incomplete. Thus, pathogenic does not mean inevitable or severe.

Classification is also distinct from variant consequence. Terms such as missense, nonsense, frameshift, splice-site, and deletion describe what changed. They do not by themselves establish clinical significance. A truncating variant is often harmful in a gene where loss of function causes disease, but it may be benign if it occurs in the final exon, in a nonessential transcript, or in a gene where truncation is not the disease mechanism.

Clinical laboratories commonly use the framework developed by the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. The framework assigns evidence codes supporting pathogenicity or benignity at different strengths. ClinGen expert groups have refined many codes and created gene- or disease-specific rules.

The final label is a structured professional judgment. It reflects the evidence available on the date of interpretation, the laboratory’s validated procedures, and the condition under consideration. The same variant may be classified differently for two diseases if the gene has distinct mechanisms. Somatic tumor variants also use separate classification systems focused on diagnostic, prognostic, or therapeutic significance.

A report should name the gene, transcript, variant nomenclature, zygosity, classification, and clinical interpretation. Reading only the label can be misleading. The supporting paragraph often explains whether the finding fits the patient, whether a second variant is needed, and which limitations remain.

The Five Classification Categories

The five-tier system separates variants with evidence for disease, variants with evidence against disease, and variants for which the evidence is insufficient or conflicting.

Pathogenic

A pathogenic classification means the evidence strongly supports that the variant causes disease in the stated gene-condition context. Examples may include a well-established loss-of-function variant in a gene where loss of function is known to cause disease, a recurrent variant seen in many affected people with convincing functional data, or a change with overwhelming segregation and case evidence.

Pathogenic is not synonymous with fully penetrant. A person may carry the variant and remain asymptomatic because of age, reduced penetrance, mosaicism, sex-specific effects, or modifying factors. The category also does not grade severity.

Likely pathogenic

Likely pathogenic means the evidence strongly favors disease causation but does not reach the laboratory’s threshold for pathogenic. Professional frameworks generally intend this category to represent a high probability of pathogenicity. In practice, pathogenic and likely pathogenic findings are often managed similarly when the result fits the phenotype and inheritance pattern.

The word “likely” refers to certainty about the variant, not the likelihood that the carrier will develop symptoms. A likely pathogenic variant in a low-penetrance gene can still confer modest risk, while a pathogenic variant in another gene can confer high risk.

Variant of uncertain significance

A VUS has insufficient or conflicting evidence. It may be rare and biologically plausible, but rarity and prediction are not enough. Some VUS will eventually be upgraded; many are downgraded to likely benign or benign as population data expand.

A VUS should not be described as a mutation, positive result, or diagnosis. Clinical management should generally be based on personal and family history until the classification changes or independent evidence confirms the condition.

Likely benign

A likely benign variant has evidence indicating that it does not cause the specified disorder, although the evidence may not be absolute. It is not used to explain the patient’s phenotype and is often omitted from the main report.

Benign

A benign variant is considered not disease-causing in the stated context. Common population variants are often benign when their frequency is incompatible with a rare, highly penetrant disorder. Well-established neutral functional data or observation in many healthy people can also support benignity.

Benign does not mean the DNA position has no biological effect of any kind. A variant can influence a trait or modify risk without causing the Mendelian disorder being evaluated. Classification is tied to a defined clinical question.

Risk alleles and low-penetrance variants

Some variants do not fit neatly into a high-penetrance Mendelian framework. Risk alleles modestly increase susceptibility, and low-penetrance variants cause disease in only a minority of carriers. ClinGen has developed approaches for curating these findings, but laboratories may report them using terms outside the standard five categories. The report should quantify risk and explain how management differs from that of a classic pathogenic variant.

Evidence Used to Classify a Variant

No single evidence type is universally decisive. Laboratories combine independent lines of evidence while avoiding double counting the same underlying observation.

Population frequency

A variant that is too common in the general population is unlikely to cause a rare, highly penetrant disease. Analysts compare allele frequency with disease prevalence, penetrance, inheritance, and genetic heterogeneity. The relevant threshold differs by gene and condition.

Absence from a population database provides only limited support for pathogenicity. Every person carries rare variants, and underrepresented ancestry groups may have fewer reference data. A variant can be absent simply because it is newly arisen or geographically restricted.

Population databases also require quality review. Apparent carriers may reflect sequencing artifacts, somatic changes in blood, or mislabeled regions. Analysts examine coverage, allele balance, and cohort composition rather than relying on one frequency number.

Predicted molecular consequence

A nonsense, frameshift, canonical splice-site, or whole-gene deletion may cause loss of function. This is strong evidence only when loss of function is an established disease mechanism for that gene and the variant is expected to eliminate meaningful protein activity.

Analysts consider transcript relevance, exon usage, nonsense-mediated RNA decay, alternative start sites, rescue by splicing, and whether the affected region is essential. A premature stop near the end of a gene may escape RNA decay and have a different effect from an early stop.

Missense variants replace one amino acid with another. Their effect depends on conservation, protein domain, biochemical properties, and known disease hotspots. Computational predictions can support classification when calibrated, but they are not equivalent to experimental proof.

Functional studies

Functional evidence may measure protein activity, RNA splicing, gene expression, cellular localization, or an organismal phenotype. A well-validated assay that reflects the disease mechanism can provide strong evidence. A research experiment with unclear controls or a model unrelated to human disease may provide little weight.

Laboratories ask whether the assay distinguishes known pathogenic and benign variants, whether results are reproducible, and whether the tested construct matches the patient’s variant. Overexpression in an artificial cell system can produce effects that do not occur in native tissue.

RNA analysis is particularly useful for suspected splice variants. It can show exon skipping, intron retention, or use of a cryptic splice site. Tissue selection matters because the gene may not be expressed in blood, and normal alternative transcripts can be mistaken for abnormal splicing.

Case observations and phenotype

Finding a rare variant in several unrelated people with a highly specific phenotype supports pathogenicity. The strength depends on how well cases are documented, whether alternative diagnoses were excluded, and whether affected people were selected because they already had the variant.

Phenotype specificity can be powerful. A person with a distinctive biochemical signature strongly linked to one gene provides more evidence than a person with a common, nonspecific symptom. However, phenotype alone cannot transform a weak candidate into a definitive variant.

Segregation in families

Segregation asks whether the variant tracks with disease among relatives. Multiple informative affected carriers and unaffected noncarriers can support pathogenicity. The analysis must account for penetrance, age, phenocopies, and family structure.

A variant found in a healthy older relative may provide benign evidence for a fully penetrant childhood disorder. It is less informative for a late-onset or incompletely penetrant condition. Related individuals are not independent observations, so statistical approaches may be used.

De novo occurrence

A confirmed de novo variant is present in the patient but absent from both biological parents. This can support pathogenicity, especially when the phenotype is specific and the gene is known for dominant de novo disease. Parentage and sample identity should be confirmed when strong evidence is assigned. Parental germline mosaicism remains possible even when blood tests are negative.

Allelic and phase data

For recessive disorders, finding a pathogenic variant in trans with another pathogenic variant supports the candidate change. If both are in cis on the same chromosome, they may not explain the disease. Testing parents can establish phase.

For dominant disorders, observing the candidate in cis with a known pathogenic variant may provide benign evidence in some contexts, but the logic depends on disease mechanism. Analysts apply such evidence cautiously.

Reputable databases and literature

ClinVar aggregates classifications from laboratories and expert panels, but submissions can conflict and vary in age and supporting detail. An expert-panel review generally carries more weight than an older single submitter. Database status is evidence to investigate, not a substitute for independent evaluation.

Published case reports can contain errors, duplicate patients, incomplete phenotypes, or outdated classifications. Laboratories trace evidence to primary data and avoid counting the same family multiple times across papers.

Why Gene and Disease Context Change the Answer

Generic rules are necessary, but gene-specific knowledge improves accuracy. The same evidence code may deserve different weight in different genes.

A population frequency that is impossible for one rare syndrome may be compatible with a common, low-penetrance condition. A truncating variant may be pathogenic in a haploinsufficient gene, benign in a gene where disease requires a specific gain-of-function change, or associated with a different phenotype depending on its location.

Gene-specific expert panels define:

  • which transcripts are clinically relevant;
  • which exons can be skipped without harm;
  • valid loss-of-function regions;
  • mutational hotspots and critical domains;
  • calibrated frequency thresholds;
  • approved functional assays;
  • phenotype specifications;
  • rules for combining evidence;
  • disease-specific penetrance and inheritance.

This work can reduce VUS rates and disagreement. BRCA1/BRCA2, APC, ATM, RYR1, LDLR, and other genes have specialized recommendations. A laboratory should use current expert guidance when available rather than apply every generic criterion identically.

Disease mechanism is central. Variants in one gene can cause different disorders through loss of function, gain of function, dominant-negative effects, altered splicing, or repeat expansion. A variant may be pathogenic for one phenotype and not established for another.

Transcript choice also changes nomenclature and consequence. A variant may fall in a coding exon of one transcript but an untranslated region of another. Laboratories select clinically relevant transcripts and should list them on the report. Comparing two reports without matching transcripts can create the appearance of disagreement when the variant is the same.

Somatic oncology classification uses a separate question: what is the variant’s significance in this cancer? A tumor variant can be categorized by therapeutic, diagnostic, or prognostic evidence even when its germline pathogenicity is unknown. Conversely, a germline pathogenic variant may not be an actionable tumor target.

Copy-number and structural variants use additional frameworks that consider gene dosage, region content, size, inheritance, and breakpoint effects. The familiar five labels may still appear, but the evidence calculation differs from that for a single-nucleotide variant.

How to Handle a Variant of Uncertain Significance

A VUS is a statement of insufficient knowledge. It should be held open rather than forced into a positive or negative category.

The safest clinical principle is that a VUS should not determine irreversible management. Risk-reducing surgery, organ removal, pregnancy termination, or predictive testing of healthy relatives should not be based solely on a VUS. Screening may still be appropriate because of the person’s symptoms or family history, but the reason should be documented as clinical risk rather than the uncertain variant.

Several questions help put a VUS in context:

  1. Does the gene clearly match the patient’s phenotype?
  2. Is the inheritance pattern compatible with the family?
  3. Was the VUS the only finding, or is a second variant required?
  4. Is the variant in a gene with a strong disease relationship?
  5. Could family studies or RNA testing provide meaningful evidence?
  6. Does the laboratory offer reinterpretation?

Family testing is not automatically useful. Testing healthy relatives simply to see whether they carry the VUS can generate ambiguous data. Segregation analysis should be planned with a genetics professional or laboratory that can state which relatives are informative and how each possible result would affect classification.

In a dominant condition, testing several clearly affected and unaffected relatives may help. In a recessive condition, parental testing may establish whether two variants are in trans. In a de novo disorder, testing parents can provide important evidence. Small families and incomplete penetrance often limit conclusions.

Functional testing should also be targeted. A commercial “functional” assay is not necessarily validated for classification. The laboratory should confirm that the test measures the relevant mechanism and that its result can be incorporated into a recognized framework.

Patients may search online databases and find another laboratory calling the same variant pathogenic. Differences can arise from newer evidence, access to internal case data, transcript choice, disease context, or inconsistent application of criteria. The ordering laboratory should review the conflict before care changes.

A VUS can be emotionally difficult because it appears specific but offers no clear answer. Genetic counseling can help distinguish “not known” from “high risk” and prevent family members from treating uncertainty as a diagnosis.

Reclassification and Laboratory Disagreement

Variant classifications are not permanent. Reclassification occurs when evidence changes or when improved rules are applied. Large population databases can show that a rare-appearing variant is common and benign. New functional studies or multiple well-documented cases can support an upgrade. Gene-specific criteria can resolve variants that remained uncertain under generic rules.

Most VUS reclassifications reported in large cohorts are downgrades toward likely benign or benign, although upgrades do occur. The probability varies by gene, ancestry, test type, and laboratory. Patients should not assume that a VUS will eventually become pathogenic.

Reclassification can move in several directions:

  • VUS to likely benign or benign;
  • VUS to likely pathogenic or pathogenic;
  • likely pathogenic to VUS when evidence weakens;
  • pathogenic to a lower category after error correction;
  • likely benign to VUS or higher in unusual cases.

A clinically important change should prompt an amended report or formal notification according to laboratory policy. Responsibility for recontact is shared and varies by health system. Laboratories may not have current contact details, clinicians may have moved, and patients may no longer receive care at the ordering center.

Practical steps include keeping the original report, registering with the laboratory portal if available, updating contact information, and asking the genetics clinic about review intervals. Reanalysis of an exome or genome is broader than variant reinterpretation: it can include newly discovered genes, improved pipelines, and additional variant types.

Two laboratories can disagree without either being careless. One may have unpublished internal cases or gene-specific expertise. The other may use a different evidence threshold or have reviewed the variant more recently. ClinVar displays conflicts, but not all submissions are equally current or authoritative.

When disagreement affects care, a qualified professional can request the evidence summaries, compare transcript and disease context, check expert-panel assertions, and ask laboratories to reconcile. Repeating the DNA test usually does not solve an interpretive conflict because both laboratories may agree on the base change and disagree only on meaning.

Classification date matters. A report from several years ago should not be assumed current. At the same time, automated websites that recalculate classifications are not substitutes for a clinical laboratory because they may misapply criteria, double count evidence, or lack phenotype and internal data.

Turning a Classification Into Clinical Care

Clinical action requires three matches: the variant must be sufficiently classified, the gene must be valid for the condition, and the patient’s genotype and phenotype must fit the inheritance pattern.

For a pathogenic or likely pathogenic finding, the care team asks:

  • Is one variant enough, or are two required?
  • Are two variants in trans?
  • Is the finding germline, somatic, or mosaic?
  • Does the phenotype match the gene’s known spectrum?
  • What is the penetrance and age-related risk?
  • Are management guidelines available?
  • Which relatives should be offered targeted family testing?

A positive classification can confirm a diagnosis, guide surveillance, support targeted treatment, alter medication choice, or inform reproductive planning. The intervention should follow evidence for that specific gene and condition. “Pathogenic” alone does not specify what care is appropriate.

For a VUS, manage the patient based on clinical findings and established family risk. Document the uncertainty, avoid predictive testing of healthy relatives unless part of a structured segregation study, and establish a route for reinterpretation.

For a likely benign or benign finding, do not attribute symptoms to the variant. The diagnostic search may need to continue through another gene, another variant type, a broader sequencing test, or a nongenetic evaluation.

A negative report does not mean no variants were found. Laboratories filter and classify many changes and may report only those relevant to the indication. The report’s methods and limitations explain what was analyzed. An unreported benign variant has no bearing on whether difficult regions or structural changes were adequately tested.

Patients should share the full report rather than a screenshot of the classification line. The gene, transcript, nomenclature, zygosity, specimen, interpretation, and limitations are all needed. When a classification changes, relatives who were tested based on the old label may need updated counseling.

Variant classification is a living evidence process. Its value comes from disciplined uncertainty: acting when evidence is strong, withholding conclusions when it is not, and revisiting the answer as science improves.

References

Disclaimer

This article is educational and does not replace interpretation by a clinical laboratory, genetics professional, or treating clinician. Variant classifications can differ by gene, disease, laboratory, and date. Medical decisions should use the complete report, current evidence, personal and family history, and appropriate professional guidance.

RELATED ARTICLESMORE FROM AUTHOR