Invitae’s method of variant classification
Accurate and reproducible
Invitae’s method of variant classification is a systematic process for assessing the evidence gathered during variant review and applying a formal variant classification based on this evidence.
Invitae’s team of scientists, physicians, and genetic counselors works together to provide high-quality and rigorous variant classifications. Our method (published in Genetics in Medicine) adheres closely to the recommendations from the American College of Medical Genetics (ACMG).1,2 Invitae communicates the evidence and logic behind the classification to the clinician through an interpretive report.
Invitae's interpretation and reporting process: Step by step
1
Variant research
The first step of variant interpretation is to collect and research evidence related to each variant.
The following evidence types are included in this review:
- Previous observations of the variant in individuals
- The type of variant and the mechanism of disease
- Experimental evidence
- Indirect and predictive evidence
- Clinical interpretations from other laboratories (variant classifications from other laboratories are noted but are not considered to be independent pieces of evidence)
2
Variant classification
After the relevant information has been gathered and the variant has been thoroughly researched, a formal variant classification is assigned.
A formal classification is meant to help answer two questions about the clinical significance of the variant.
- From a diagnostic perspective: Does this variant, in the correct genetic background, provide an explanation for disease in an affected individual?
- From a predictive perspective, narrowly applied to highly penetrant conditions: Is an individual who inherits this variant likely to develop disease?
The ACMG has recommended a five-tier classification system. According to this system, a variant can be classified as:
- Pathogenic: This variant directly contributes to the development of disease. Some pathogenic variants may not be fully penetrant. In the case of recessive or X-linked conditions, a single pathogenic variant may not be sufficient to cause disease on its own. Additional evidence is not expected to alter the classification of this variant.
- Likely pathogenic: There is a high likelihood (greater than 90% certainty) that this variant is disease-causing. Additional evidence is expected to confirm this assertion of pathogenicity, but there is a small chance that new evidence may demonstrate that this variant does not have clinical significance.
- Uncertain significance: There is not enough information at this time to support a more definitive classification of this variant.
- Likely benign: This variant is not expected to have a major effect on disease; however, the scientific evidence is currently insufficient to prove this conclusively. Additional evidence is expected to confirm this assertion, but we cannot fully rule out the possibility that new evidence may demonstrate that this variant can contribute to disease.
- Benign: This variant does not cause disease.
Four additional classifications are also periodically invoked to capture variable expressivity:
- Pathogenic (low penetrance): This variant is commonly accepted as a contributing factor of disease, but the penetrance of this particular change is sufficiently low to be often seen in individuals without disease. As a result, the predictive value of this information is considered to be low.
-
Increased risk allele: This variant is not expected to cause disease. However, substantial association data exists suggesting that the presence of this variant may be indicative of increased risk for either a related disease or a completely separate condition.
-
Pseudodeficiency allele: This variant can lead to false positive results on biochemical enzyme studies, but is not known to cause clinical symptoms or lead to disease. Enzyme studies cannot differentiate between true pathogenic variants and pseudodeficiency alleles, so these must be distinguished by molecular studies.
-
Benign reportable variant: This category of variant classification includes variants that can lead to abnormal biochemical studies or biomarker studies, but are not known to cause clinical symptoms or lead to disease. This category also includes variants that do not directly cause disease but may affect disease severity.
Variants are assigned a formal classification based on:
- Weighted evidence: The evidence gathered during variant review is scored based on the relative importance of the evidence type. Certain types of evidence may support a pathogenic classification while others may support a benign classification.
- Grouping by basic argument: If multiple pieces of evidence point to the same basic argument, only the strongest piece of evidence is considered.
- Example: A functional study that demonstrates reduced protein levels and an in silico algorithm that predicts that the variant is deleterious, both contribute to the general argument that “protein function may be disrupted.”
- Additive vs. non-additive evidence: Certain types of evidence can be considered more than once, and certain types cannot.
- Example: Independent observations of the variant in multiple affected individuals are additive.
- Example: Multiple experimental studies demonstrating that different aspects of protein function are disrupted are non-additive.
- Extensive review to ensure accurate and consistent classification: Each variant classification undergoes extensive review by our clinical reporting team. Invitae has developed proprietary software tools to efficiently record and share information, an important step to ensure reproducibility.
3
Reporting
Our clinical report:
- Highlights the most important findings for your patient
- Provides relevant clinical information about the diseases associated with the genes in which variants were identified
- Describes the relevance of the findings for your patient and their family
- Suggests possibilities for follow-up testing
- Describes in detail the evidence and logic supporting each variant interpretation
- Provides recurrence risk information when possible
The report includes:
- Summary: A clear statement summarizing the high-level result of the genetic test — positive, negative, or clinically inconclusive
- Clinical summary: A description of the relevance of the genetic results to the patient based on their clinical and family history
- Complete results table: A table of the genes in which genetic variants were identified, and a complete list of all genes analyzed in the test*
- Variant details: A summary of the evidence and logic used to justify the interpretation of each variant
- Methods and limitations: A description of Invitae's next-generation sequencing assay along with assay limitations
The field of genetics is constantly evolving, often revealing new evidence relevant to variant interpretation. When new evidence on a variant becomes available, we review our variant interpretation and, if indicated, we will reclassify the variant and issue an amended report to the ordering clinician. If a report is amended, the ordering clinician will receive a notification via phone or email.
Genetic testing can have health implications not only for an individual, but for an entire family. Invitae offers family variant testing at no additional charge for all first-degree relatives of patients who undergo gene or panel testing for a hereditary condition at Invitae and are found to have a pathogenic or likely pathogenic variant. To help resolve variants of uncertain significance (VUS) in our test results, Invitae also offers complimentary follow-up testing to select family members of patients tested at Invitae when informative data can be obtained.
To learn more about these programs, please visit our VUS resolution and family variant testing webpages.
*Benign variants and likely benign variants are not included in the report but are available upon request.
What's different?
A formal classification framework
While the ACMG-AMP guidelines were a major step toward establishing a common framework for variant classification, several aspects of the guidelines lack specificity or are subject to varied interpretations. Our clinical genomics group developed Sherloc: a set of semiquantitative, hierarchical evidence-based rules for locus interpretation.
Building on the ACMG-AMP guidelines
Using the ACMG-AMP guidelines to classify more than 40,000 clinically observed variants, we refined the classification criteria to add greater clarity to subjective rules, and capture edge cases and exceptions.
| Limitation | Sherloc solution |
|---|---|
| Certain ACMG-AMP rules conflate concepts that should be considered separately. | Expands overly encumbered criteria into a set of discrete but related rules that can be weighted separately. This reduces subjectivity and increases consistency in the variant classification. |
| Certain pairings of ACMG-AMP rules capture types of evidence that contribute to the same basic argument, which creates a “double counting” effect. | Groups evidence criteria into broader categories so that only the strongest piece of evidence from the category is counted. This prevents inflated classifications due to “double counting” weak pieces of evidence. |
| The “clinical criteria” style of the ACMG-AMP guidelines introduce difficulty in intuitively understanding the cumulative strength of the evidence. As a result, it is difficult to appreciate how much additional evidence is needed to move to a more definitive classification. | Substitutes the “clinical criteria” style with a categorical framework and numerically weighted evidence criteria. The sum of the weighted evidence represents the strength of the argument for pathogenicity, supporting a more intuitive understanding of the evidence required for a definitive variant classification. |
Example: Evidence for pathogenicity based on family segregation
In the case of family segregation, Sherloc takes a vague notion of supporting evidence that "may be used as stronger evidence” in certain situations, and distills it down into quantifiable scores that can be combined with other pieces of evidence to arrive at a final variant classification.
Functional modeling platform: Reducing uncertainty in genetic testing
Invitae’s state-of-the-art Functional Modeling Platform (FMP) provides clarity for patients with variants of uncertain significance (VUS). By pulling in many lines of evidence from both lab experiments and computational analyses, FMP can accurately predict how some VUS will affect gene function. For 1 in 40 (or 2.5%) of Invitae patients, that means we can provide a more definitive variant classification (benign, likely benign, likely pathogenic, or pathogenic), rather than a VUS.
To learn more about how Invitae delivers deeper, more informative results with FMP, read our blog post, flyer, or recent white paper.
