Guidelines for variant classification and interpretation

ACMG/AMP guidelines

The current guidelines for variant classification and interpretation stem from the expert opinion of the workgroup convened by the American College of Medical Genetics and Genomics (ACMG) in 2013 that comprised representatives from the ACMG, the Association for Molecular Pathology (AMP) and the College of American Pathologists (CAP). The latest guidelines were published in 2015. You will learn more about how to apply these guidelines throughout the course.

These guidelines are based on expert opinion and empirical data, though some criteria use published evidence.

When do you apply these criteria?

These guidelines are to be used for the interpretation of variants observed in patients with suspected inherited (primarily Mendelian) disorders in a clinical diagnostic laboratory setting. Please refrain from using them to interpret somatic variations, pharmacogenomic (PGx) variants, or variants in genes associated with multigenic non-Mendelian complex disorders. Ensure you apply these criteria for variants with an established gene-disease relationship. These rules should be applied to the gene and the specific disorder pair (and the inheritance pattern), and not to all the Mendelian disorders in which the gene might have a role.

Terminology and criteria

Although the basic premise is that the clinical significance of any given sequence falls along a gradient (continuous variable), the terminology uses a categorical approach: pathogenic, likely pathogenic, variant of uncertain significance, likely benign and benign.

There are two sets of criteria: one for the classification of pathogenic and likely pathogenic variants and the other for the classification of benign and likely benign variants. If a variant does not fit the criteria of pathogenic or benign, or the evidence for benign and pathogenic is conflicting, the variant defaults to being of uncertain significance. ‘Likely’ refers to a greater than 90% certainty that a variant is either disease-causing or benign.

What evidence is used for variant classification?

The variant classification process involves collecting and evaluating the data on several qualities of a variant. Most importantly, we look at the following:

• Allele frequency of the variant in the healthy population. A disease-causing variant is very rare, and a benign variant is frequent. Therefore, prioritise local databases to assess population data.
• A variety of software programmes, both publicly and commercially available, can help in the interpretation of sequence variants by predicting their consequences on a gene or protein function.
• Published functional data can provide evidence of the functional consequences of the variant in human cells and animal models.
• Segregation data and de novo data provide evidence of how the variant segregates in the family and other affected and unaffected individuals.
• Always search for information in mutation databases from reputable sources.

You will find several sections in the course addressing these issues in more detail.

Each criterion is weighted, but is flexible

The criteria for determining the pathogenicity of genetic variants are weighted as very strong, strong, moderate, or supporting for pathogenic criteria, and stand-alone, strong, or supporting for benign criteria. To enable some flexibility, these criteria can be adjusted between weights based on professional judgement and the evidence available. However, expert judgement needs to be applied when evaluating the full body of evidence to account for differences in the strength of variant evidence.

Click to expand

Figure 1. Evidence Framework. This chart organises each of the criteria by the type of evidence, as well as the strength of the criteria for a benign (left side) or pathogenic (right side) assertion. BS, benign strong; BP, benign supporting; FH, family history; LOF, loss of function; MAF, minor allele frequency; path., pathogenic; PM, pathogenic moderate; PP, pathogenic supporting; PS, pathogenic strong; PVS, pathogenic very strong. Source: Genetics in Medicine.

We recommend you download the original table from the manuscript, as this might be useful in other steps throughout the course.

How are these guidelines used in clinical management?

These guidelines assist in the clinical management of patients by informing decisions. They provide guidance on whether Pathogenic and Likely Pathogenic variants are likely to cause disease. It’s important that variants of uncertain significance (VUS) should not be used for clinical decision-making. A VUS represents a 10-90% chance of being pathogenic, and not all VUSs are the same. However, variant classification is just one of several pieces of evidence required for clinical decision-making. We will discuss VUS in more detail in Week 3.

Other critical factors include clinical information (external phenotype, results from routine and special clinical and laboratory investigations, and imaging findings); and family history (information about the patient’s family health history).

Although the ACMG/AMP framework has been central to the harmonisation of variant pathogenicity classification, there is still discordance in the use of criteria. Recommendations and guidance will continue to change due to the refinement of the framework and new/updated resources and datasets. Since the publication of these guidelines, several additional recommendations have been made to improve them. Interested readers can view these in the links below. There is a new version of the framework under development, which will mainly use the same evidence types, but will include changes in how evidence is considered and updated criteria names.

Share your examples: do you have any experience in this topic you would like to share? Add a link to resources you think can help deepen the understanding of the topic.