Representation of clinical significance in ClinVar and other variation resources at NCBI
- Clinical significance on ClinVar submitted records (SCV)
- Clinical significance on ClinVar aggregate records (VCV and RCV)
- Representation in other variation resources
- Terminology
Clinical significance on ClinVar submitted records (SCV)
Standardizing representation
ClinVar uses standard terms for clinical significance recommended by an authoritative source when available. These standards include:
- the five terms for Mendelian diseases recommended by ACMG/AMP
- the terms for low penetrance variants and risk alleles recommended by ClinGen
Other non-standard terms are also used to meet submitter needs. ClinVar plans to add standard terms for other scenarios (e.g. somatic, pharmacogenomic, and mitochondrial variants) when they are available.
Options for clinical significance
The list of terms currently used by both ClinVar and GTR is reported on GTR's ftp site: https://ftp.ncbi.nlm.nih.gov/pub/GTR/standard_terms/Clinical_significance.txt. The following table lists these options for clinical significance for an individual submission (SCV) in ClinVar and guidance for when each term should be used.
Clinical significance value | Guidance for use in ClinVar SCV records |
---|---|
Benign | As recommended by ACMG/AMP for variants interpreted for Mendelian conditions. |
Likely benign | As recommended by ACMG/AMP for variants interpreted for Mendelian conditions. |
Uncertain significance | As recommended by ACMG/AMP for variants interpreted for Mendelian conditions. |
Likely pathogenic | As recommended by ACMG/AMP for variants interpreted for Mendelian conditions. |
Pathogenic | As recommended by ACMG/AMP for variants interpreted for Mendelian conditions. |
Likely pathogenic, low penetrance | As recommended by ClinGen for variants with decreased penetrance for Mendelian conditions. |
Pathogenic, low penetrance | As recommended by ClinGen for variants with decreased penetrance for Mendelian conditions. |
Uncertain risk allele | As recommended by ClinGen for variants with decreased penetrance for Mendelian conditions. |
Likely risk allele | As recommended by ClinGen for variants with decreased penetrance for Mendelian conditions. |
Established risk allele | As recommended by ClinGen for variants with decreased penetrance for Mendelian conditions. |
drug response | A general term for a variant that affects a drug response, not a disease. We anticipate adding more specific drug response terms based on a recommendation by CPIC. |
association | For variants identified in a GWAS study and further interpreted for their clinical significance. | protective | For variants that decrease the risk of a disorder, including infections. |
Affects | For variants that cause a non-disease phenotype, such as lactose intolerance. |
conflicting data from submitters | Only for submissions from a consortium, where groups within the consortium have conflicting intepretations of a variant but provide a single submission to ClinVar. |
other | If ClinVar does not have the appropriate term for your submission, we ask that you submit "other" as clinical significance and contact us to discuss if there are other terms we should add. |
not provided |
For submissions without an interpretation of clinical significance. The primary goal of ClinVar is to archive reports of clinical significance of variants. Therefore submissions with a clinical significance of "not provided" should be limited to:
|
'-' | This value may not be submitted. It is used in the file variant_summary.txt.gz in the path https://ftp.ncbi.nlm.nih.gov/pub/clinvar/tab_delimited/. This file reports '-' in the ClinicalSignificance column for an allele that was submitted to ClinVar only in combination with another allele (e.g.a submission with an interpretation of a haplotype or a compound heterozygote) and was not interpreted explictly. ClinVar thus has no interpretation specific to that allele. To find the the interpretation that includes that allele, you can query ClinVar by the AlleleID,e.g. https://www.ncbi.nlm.nih.gov/clinvar/?term=38420[alleleid]. |
Assertion score
Submitters are encouraged to provide documentation describing the criteria they use to classify variants, referred to as assertion criteria. If the assertion criteria include a point-based scoring system, the final score, or point value, can be submitted as the Assertion score. The ACMG/ClinGen CNV Guidelines, 2019 is an example of a point-based scoring system for variant classification.
Clinical significance and mode of inheritance
Submitters should provide a value for clinical significance in the context of a submitted mode of inheritance. Thus a disease-causing variant for a recessive disorder should be reported as Pathogenic, not Benign, even though the single copy will not cause disease.
When a submitter specifies a mode of inheritance, that value is displayed on the variant page on the "Clinical assertions" tab in parentheses in the Conditions column, as in this example .
Source of the interpretations of medical relevance
The value of clinical significance that ClinVar represents on a submitted record is not calculated by NCBI. The interpretation of variation is provided only by the submitter. The interpretation on the submitter's record, represented by an accession starting with SCV, is used to calculate conflicts among interpretations when
- all submissions about the same variation and condition are aggregated into a record with an accession starting with RCV
- all submissions about the same variation are aggregated into a record with an accession starting with VCV
In other words, ClinVar staff do not curate submitted interpretations. If you believe an interpretation is in error, you are invited to submit your data with your interpretation so the conflict will be noted and external groups such as ClinGen can focus on evaluation of all evidence.
There are two exceptions to the statement that ClinVar processes only what is explicitly submitted about clinical significance:
-- Data provided by OMIM do not include an explicit interpretation of clinical significance. When ClinVar processes data from OMIM, we map the disorders or phenotypes associated with each record into clinical significance values according to the following guidelines provided by OMIM:
Key Word(s) in Title of OMIM Entry or description in OMIM's genemap file | Clinical Significance |
---|---|
non-disease | Affects (implemented in the October, 2015 release) |
named protein variants | other, unless curated by NCBI staff to connect to a condition or phenotype |
"POLYMORPHISM" | Benign |
"PROTECTION AGAINST" | protective |
"QUANTITATIVE TRAIT" | association |
"RESPONSE" not "SUSCEPTIBILITY" | drug response |
"SOMATIC" not "RESPONSE" | pathogenic (implemented in the September, 2014 release) |
"SUSCEPTIBILITY TO" |
risk factor Note: For the case of variants in BRCA1 and BRCA2 related to breast cancer, OMIM asserted these should be treated as pathogenic. |
"VARIANT OF UNKNOWN SIGNIFICANCE" | Uncertain significance, also reviewed as time permits to determine if there is a disorder for which the variant has an uncertain significance. |
contains an identifier from dbSNP | Reviewed as time permits to determine if the accompanying text is reporting an association, a polymorphism, or a causative variant. |
key title word(s) that do not match any of the above | Pathogenic |
-- In some early submissions, some submitters provided clinical significance using their own values. In this case, the SCV record reflects the submitted clinical significance value, but ClinVar staff worked with the submitter to map the submitted value to one of NCBI's standard values. This standardization is necessary to support calculation of conflicts of interpretation. Current submissions to ClinVar must use the list of clinical significance terms that ClinVar processes.
Clinical significance on ClinVar aggregate records (VCV and RCV)
ClinVar aggregates the values of clinical significance provided in submitted records (SCV) by the variant (VCV records) or the variant/condition combination (RCV records). Some interpretations are given more weight in doing that aggregation, based on the review status of each submitted record represented in the VCV or RCV. Submitted records are used in this order of precedence:
-
practice guideline The clinical significance from the practice guideline record is used as the clinical significance on the VCV and RCV recordss, no matter what other submitters may have reported.
-
reviewed by expert panel The clinical significance from the expert panel record is used as the clinical significance on the VCV and RCV recordss, no matter what other submitters may have reported.
-
criteria provided, single submitter The clinical significance from all submitted records with this review status is used to calculate the aggregate clinical significance on the VCV and RCV records.
For example, the interpretation on a single SCV record with review status "criteria provided, single submitter" supercedes interpretations on multiple SCV records with lower review statuses.
-
no assertion criteria provided The clinical significance from all submitted records with this review status is used to calculate the aggregate clinical significance on the VCV and RCV records, if there is no record with higher precedence.
Representing conflicts in interpretation
The clinical significance terms recommended by ACMG/AMP use five levels to describe pathogenicity. Five additional terms to classify risk alleles and low-penetrance Mendelian variants were introduced by ClinGen. See Options for clinical significance. A difference in classification among submitters within these ten terms only are reported as a conflict using the phrase 'Conflicting interpretations of pathogenicity'. Other values of clinical significance described in Options for clinical significance are not considered in the determination of conflicting interpretations.
RCV records
- Starting in June 2022, conflicts on RCV records are reported between these levels of pathogenicity:
- Any pathogenic/risk term vs any uncertain term
- Any pathogenic/risk term vs any benign term
- Any uncertain term vs any benign term
- Starting in May 2018, conflicts on RCV records were reported between three levels of pathogenicity, i.e. Benign or Likely benign vs Uncertain significance vs Pathogenic or Likely pathogenic.
- Before May 2018, any conflict within the five terms of pathogenicity from ACMG/AMP was reported as a conflict. e.g. Pathogenic and Likely pathogenic were reported as a conflict.
VCV records
- Starting in June 2022, conflicts on VCV records are reported between these levels of pathogenicity:
- Any pathogenic/risk term vs any uncertain term
- Any pathogenic/risk term vs any benign term
- Any uncertain term vs any benign term
- Before June 2022, conflicts were reported between three levels of pathogenicity, e.g. Benign or Likely benign vs Uncertain significance vs Pathogenic or Likely pathogenic.
When ClinVar aggregates submissions from groups that provided a standard term not recommended by ACMG/AMP (e.g. drug response), those values are reported after the ACMG/AMP-based interpretation (see the table below).
Conflicts on aggregate records are reported in the XML extraction, the tab-delimited variant_summary.txt file, and on the web pages.
When there is a submission from an expert panel or a practice guideline, only the interpretation from that group is reported as the aggregate clinical significance even if other submissions provide different interpretations. The assertions on the SCV records from other submitters are not changed, but they do not contribute to the aggregate interpretation.
Combination of terms from different submitters | Reported as |
---|---|
|
The combination of terms, separated by ‘/’. e.g. Pathogenic/Established risk allele |
Uncertain significance AND Uncertain risk allele | The combination of terms, separated by ‘/’ |
Benign AND Likely benign | The combination of terms, separated by ‘/’ |
|
Conflicting interpretations of pathogenicity. The conflicting terms and the number of submissions are provided. e.g. Pathogenic(1); Uncertain significance(1) |
Any single ACMG or ClinGen term AND any other term, e.g. Pathogenic and drug response | The combination of terms, separated by ‘;’ |
Conflicting ACMG or ClinGen terms AND any other term, e.g. Pathogenic and Uncertain significance and drug response | Conflicting interpretations of pathogenicity; drug response The conflicting terms and the number of submissions are provided. |
No ACMG or ClinGen term AND multiple other terms, e.g. drug response and other | The combination of terms, separated by ‘;’ |
A consortium does not agree on which ACMG assessement is valid | Conflicting data from submitters |
Representation in other NCBI variation resources
ASN.1 files
ASN.1 files from NCBI with clinical variant annotation use a different set of clinical significance terms than those displayed in ClinVar and VCF, with the following mapping:
ASN.1 terms | ClinVar and VCF |
---|---|
0 – unknown | Uncertain significance |
1 – untested | not provided (includes the cases where data are not available or unknown) |
2 - non-pathogenic | Benign |
3 - probable-non-pathogenic | Likely benign |
4 - probable-pathogenic | Likely pathogenic |
5 – pathogenic | Pathogenic |
6 - drug-response | drug response |
7 – histocompatibility | histocompatibility |
255 - other | other |
confers sensitivity | |
risk factor | |
association | |
protective | |
Affects |
Terminology
Several different terms have been used in ClinVar for the field that represents the relationship between the variant and a condition. These terms include “clinical significance,” “assertion,” and “interpretation,” and were used based on feedback from ClinVar's community of users. There does not appear to be a standard that is used uniformly across all of clinical genetics.
Regardless of the term used, this relationship is considered a variant‐level interpretation, not a patient‐specific interpretation. It may represent the relationship between a variant and Mendelian disease, cancer, drug response, or other condition. As the relationship may be asserted through clinical testing, research, or curation, an “interpretation” is neither limited to a clinical context, nor to classifications of pathogenicity for disease. PMID 30311387