Variants with pathogenic/likely pathogenic disease associations¶
This criterion enables data from external sources to be used to prioritise variants with known pathogenic/likely pathogenic associations that would not otherwise be prioritised. This also enables variants that do not occur in transcripts with a considered biotype (see Criterion 3) to be prioritised.
The population frequency filter (Criterion 2) or predicted functional impact filter (Criterion 3) may result in potentially diagnostic variants with known pathogenic/likely pathogenic associations not being prioritised, e.g. if they are known to be pathogenic but above allele frequency threshold used during variant tiering.
ClinVar¶
ClinVar aggregates information about genomic variation and its relationship to human health. An individual submission to ClinVar asserting a relationship between a variant and condition are represented by a SCV record. All submissions about the same variant are aggregated into a VCV record.
A variant passes Criterion 6 if all of the criteria outlined below are fulfilled:
-
The variant has an exact match with the genomic position and alternate allele of variant with a VCV record in ClinVar OR the variant has a HGVSp match1 with a variant with a VCV record in ClinVar
-
The ClinVar VCV record has the following properties:
A clinical significance that is one of:
1. Pathogenic
2. Likely pathogenic
3. Conflicting interpretations of pathogenicity
A review status that is one of:
1. Practice guideline
2. Reviewed by expert panel
3. Criteria provided, multiple submitters, no conflicts
4. Criteria provided, conflicting interpretations2
5. Criteria provided, single submitter
6. No assertion criteria provided -
The variant has a population frequency in the custom Genomics England frequencies dataset less than 0.05 (i.e. 5%)
Note
The ClinVar version utilised in the pipeline is available in software and database versions.
The Rare Disease tiering pipeline does not consider ClinVar variants that are Established risk alleles
or drug response
.
Clinical Variant Ark (CVA)¶
CVA is a Genomics England database which stores information assigned to genomic variants for samples processed by Genomics England bioinformatics pipelines, including:
- variant classifications
- variant interpretation logs
- tiered variants
- clinical indications upon referral
- data reported in outcome questionnaires and summary of findings
Details of all the information available are further elaborated on the CVA online help page
A variant passes Criterion 6 if:
-
The variant has at least one pathogenic or likely pathogenic classification in CVA at the time of interpretation, or
-
A variant with the same protein change, based on HGVSp annotation for the same transcript, has at least one pathogenic or likely pathogenic classification in CVA at the time of interpretation
-
The variant must have a population frequency in the custom Genomics England frequencies dataset less than 0.05 (i.e. 5%)
Note
HGVSp annotations are only available for variants classified in cases run following the NGIS Izar release in March 2023
Note
Queries against the CVA database are done in real-time at the time of analysis; CVA is updated daily, therefore reanalysis at a later date can produce different results.
Note
Variant classifications in CVA are continually updated as GMS referrals are reviewed and findings returned. A variant with no pathogenic or likely pathogenic classifications in CVA at the point of interpretation will not pass criterion 6 for CVA even if the variant is subsequently classified as pathogenic or likely pathogenic before being reviewed. However, the variant will display the current CVA classification in the Interpretation Portal.
-
Variants matched through exact (#1) or protein-matched (#2) logic will be displayed in the same way in the interpretation portal. Users must manually review these events to determine if they were prioritised through the known pathogenic tiering criterion by exact match or through protein-matching. ↩
-
Variants with a review status of “criteria provided, conflicting interpretations” must have at least one associated SCV record with a pathogenic/likely pathogenic assertion. ↩