Performance of preliminary SDHB specifications for ACMG/AMP sequence variant classification v3: A pilot assessment
ASHG 2026
Lauren Desrosiers-Battu, Roderick Clifton-Bligh, Jean-Pierre Bayley, Terri McVeigh, Paul Newey, Aparna Prasad, Susan Richter, Jamie Weyandt, Endocrine Tumor Predisposition Variant Curation Expert Panel, Tobias Else
Background
American College of Medical Genetics and Genomics (ACMG) and Association for Molecular Pathology provide general guidance for sequence variant classification (SVC v3); however, gene-disease expertise is required to optimize variant classification. Many labs offering genomic sequencing report on ACMG secondary findings genes, including SDHB, however may not have deep experience with endocrine tumor predisposition syndromes (Endo-TPS).
Methods
The Clinical Genome Resource (ClinGen) Endo-TPS Variant Curation Expert Panel (VCEP) is developing and evaluating SDHB specifications (specs) per the ClinGen VCEP protocol. Specific codes, e.g. PP3, were calibrated with odds of pathogenicity calculations using known pathogenic/likely pathogenic (P/LP) (n=14) and benign/likely benign (B/LB) (n=7) variants. SDHB specs are being piloted for P/LP, B/LB, and uncertain or conflicting variants (VUS).
Results
Eighteen variants were evaluated via preliminary VCEP specs, with 6 classified as P/LP, 7 as VUS and 5 as B/LB. All P/LP VCEP classifications (n=6) were consistent with the ClinVar aggregate classification summary. One variant with conflict in ClinVar was downgraded to LB. All 3 variants with conflicts introduced were VUS by VCEP specs and B/LB in ClinVar. Phenotype (PS4), and in silico (PP3/BP4) codes were most frequently applied. Strength of PS4 was variable, dependent on available data, with 1 very strong (VS), 3 strong (S) and 2 moderate (M). PP3_M was applied for 5 variants with REVEL scores >0.857 and BP4 was applied for 3 variants with scores <0.627. Benign phenotype code BS2 was applicable for 2 variants.
Discussion
Performance of preliminary VCEP specs for SDHB-related TPS in a pilot of 18 variants was comparable to aggregate classification in ClinVar. Of those that differed, none generated new clinically-significant conflicts in ClinVar and 1 provided more certainty (VUS to LB). VCEP application of PP3 and BP4 differed from ClinGen genome-wide calibration as SDHB variants, regardless of classification, tend to have high REVEL scores. None of the 3 variants that met the VCEP BP4 rule would have had BP4 applied when following ClinGen genome-wide calibration guidance. Additionally, use of BS2 was limited due to moderate penetrance of dominant phenotypes and hypomorphic SDHB variants associated only with recessive mitochondrial disease. Functional data, particularly for B/LB variants, will be critical for definitive classification of SDHB variants.
Related publications