Interpreting Pathogenic and Benign Variants

Interpreting variants is the process of deciding whether a DNA change is likely to cause disease or is harmless. The standard approach places each sequence variant into one of five tiers, from benign to pathogenic, by weighing multiple independent lines of evidence rather than relying on any single feature. The same framework underlies modern clinical reporting of variant pathogenicity.

Definition

Variant interpretation is the structured assignment of a sequence variant to one of five pathogenicity categories by combining weighted lines of evidence according to a standardized classification framework.

Scope

The topic covers the five-tier classification scheme (benign, likely benign, uncertain significance, likely pathogenic, pathogenic), the categories of evidence used (population frequency, computational predictions, functional studies, segregation, de novo occurrence, and others), the rules for combining evidence strengths, and refinements such as the Bayesian reformulation of the criteria and gene-specific guidance. Copy-number variants are interpreted through a parallel standard. The treatment is methodological and does not provide clinical management direction.

Core questions

• What categories of evidence inform whether a variant is pathogenic or benign?
• How are evidence criteria weighted and combined into a final classification?
• How do the five tiers relate to one another and to certainty?
• How are special variant types, such as copy-number and loss-of-function variants, handled?

Key concepts

• Five-tier classification (benign to pathogenic)
• Evidence criteria and their strengths
• Population allele frequency
• Computational and functional evidence
• Segregation and de novo evidence
• Loss-of-function (PVS1) interpretation
• Copy-number variant interpretation

Key theories

Bayesian framework for variant classification

The ACMG/AMP combining rules can be modeled as a naturally Bayesian system in which each evidence criterion contributes a strength-weighted odds of pathogenicity, formalizing the qualitative rules into quantitative posterior probabilities.

Mechanisms

Each variant is evaluated against a defined set of evidence criteria, including its frequency in reference populations, computational predictions of effect, experimental functional data, co-segregation with disease in families, and de novo occurrence; each criterion carries a strength (supporting, moderate, strong, very strong) and a direction (toward pathogenic or benign), and the combination rules map the accumulated evidence onto one of five tiers (Richards et al., 2015). Subsequent work reformulated these qualitative rules as a Bayesian framework, showing that the criterion strengths correspond to multiplicative odds and yield an underlying continuous probability of pathogenicity (Tavtigian et al., 2018). Specialized guidance refines difficult criteria, such as how to apply the very strong loss-of-function rule (Abou Tayoun et al., 2018), while copy-number variants are scored against a parallel technical standard (Riggs et al., 2020).

Clinical relevance

Variant classification determines how a laboratory reports a finding and therefore shapes how genetic results are understood in counseling and evidence appraisal. The topic describes how pathogenicity is judged and reported; it is a reference account of the classification process and not guidance for acting on a specific result.

History

Before standardization, laboratories used heterogeneous and often incomparable schemes for calling variants pathogenic or benign. The 2015 ACMG/AMP consensus recommendation established a shared five-tier framework and evidence-combining rules (Richards et al., 2015). The field then refined it: the Bayesian reformulation gave the rules a quantitative footing (Tavtigian et al., 2018), criterion-specific and gene-specific recommendations addressed recurring difficulties (Abou Tayoun et al., 2018), and parallel standards were issued for copy-number variants (Riggs et al., 2020).

Debates

How should evidence criteria be weighted and quantified?

Whether the qualitative criterion strengths should be replaced or supplemented by explicit quantitative thresholds, and how to calibrate computational and functional evidence, remains an active area, with Bayesian modeling offering one route to consistency.

How conservatively should loss-of-function be called pathogenic?

Applying the very strong loss-of-function criterion requires confirming that loss of function is an established disease mechanism for the gene and that the predicted effect is real, prompting detailed rules to avoid over-classification.

Key figures

• Sue Richards
• Heidi Rehm
• Sean Tavtigian
• Leslie Biesecker

Seminal works

• richards-2015
• tavtigian-2018
• riggs-2020

Frequently asked questions

What are the five variant classification categories?

Benign, likely benign, uncertain significance, likely pathogenic, and pathogenic. A variant is placed in a tier by combining several independent lines of evidence rather than any single observation.

Why is no single piece of evidence usually enough?

Each evidence type has limitations, so the framework requires combining criteria of defined strength and direction; this guards against over-interpreting any one feature, such as a computational prediction or rarity alone.