Evaluating variants classified as pathogenic in ClinVar in the DDD Study

Area of Science:

• Genomics
• Medical Genetics
• Bioinformatics

Background:

• Automated variant filtering is crucial for diagnostic genome-wide sequencing.
• Standard filtering pipelines may produce false negatives, excluding true pathogenic variants.

Purpose of the Study:

• To investigate if standard variant filtering pipelines routinely exclude previously identified pathogenic variants.
• To assess the diagnostic yield of variants missed by automated filters.

Main Methods:

• Evaluated ClinVar pathogenic/likely pathogenic variants in developmental disorder genes.
• Used exome sequence data from the Deciphering Developmental Disorders (DDD) study.
• Assessed variants against DDD's filtering pipelines and clinical review.

Main Results:

• 3.6% of ClinVar pathogenic variants were found in DDD probands.
• 1134/1352 (83.9%) were previously identified by DDD filters.
• 112 additional variants (0.8% of probands) were identified as potential diagnoses after clinical review of filtered-out variants.

Conclusions:

• Lower minor allele frequency and higher ClinVar review status predict diagnostic variants.
• Approximately half of excluded pathogenic variants were not disease-causing, emphasizing the need for clinical evaluation.
• Human review remains essential for accurate variant interpretation in diagnosing developmental disorders.