Assessing pathogenicity according to international standards
DECIPHER supports the annotation and sharing of variant pathogenicity using ACMG guidelines for sequence variants and ACMG/ClinGen technical standards for CNVs, which helps to standardise the classification of variants across centres. When interpreting a CNV it is possible for users to choose to assess the variant using sequence variant guidelines, which may be more applicable for small CNVs since the distinction between a sequence variant and a CNV is blurred (Brandt et al ., 2019).
Criteria selection: In both pathogenicity interfaces (Fig. 5A, 5B), types of evidence (such as population data and functional data) are displayed, along with the relevant evidence criteria used to determine if data supports the variant being pathogenic or benign. Relevant criteria can be selected with a single click. Some of the criteria have additional information links. These either provide information about how the criteria can be used according to the original paper (e.g. de novo CNV evidence), or in the case of sequence variants they provide information about ClinGen Sequence Variant Interpretation (SVI) Working Group guidelines (e.g. recommendation for functional assays (PS3/BS3) Brnich et al ., 2019). As new guidelines become available these pathogenicity interfaces are updated to provide the latest relevant recommendations. Criteria strengths can be modified as required in the interface.
Relevant evidence: Within the interfaces there is a customised section displaying ‘evidence to consider’ which provides information relating to the specific evidence type being assessed. For example, for computational and predictive data evidence, predictive pathogenicity scores (SIFT, PolyPhen-2, CADD, REVEL, SpliceAI) are displayed. Links are also provided to relevant DECIPHER interpretation interfaces, for example to the in-built tolerated population variation calculator, which can be used to determine if a variant observed in the reference sample is too common to cause a given Mendelian disorder of interest (Whiffinet al ., 2017). External links (e.g. PubMed literature search) are also provided.
Calculation of variant pathogenicity: As criteria are added, DECIPHER uses these to calculate the variant pathogenicity. For sequence variants, this is calculated according to the combining rules detailed in the original 2015 ACMG guidelines. In addition, DECIPHER calculates the posterior probability of pathogenicity and classification according to the ClinGen SVI Working Group’s Bayesian classification framework, which provides a mathematical foundation for the combining rules (Tavtigian et al ., 2018). DECIPHER highlights cases where these classifications disagree, and ultimately all pathogenicity assessments are made by depositors using their professional discretion. For CNVs, the evidence can be scored according to ACMG/ClinGen technical standards instead.
ClinGen Expert Panel specifications: For some genes there are ClinGen Variant Curation Expert Panel Specifications, which recommend adaptations of the sequence variant ACMG guidelines (e.g. Rett and Angelman-like Disorders Variant Curation Expert Panel for MECP2 ,CDKL5 , FOXG1 , UBE3A , SLC9A6, andTCF4 ). When interpreting variants in genes for which these recommendations exist, detailed information about how to apply the criteria is provided along with a link to the relevant Clinical Domain Working Group, so that patients with variants in these genes benefit from interpretation in accordance with these recommended standards.