DNA Sequencing and Variant Annotation

Extracted genomic DNA was prepared for whole-exome sequencing, and sequencing was performed at the Sequencing and Genomic Technologies Shared Resource, part of the Duke University School of Medicine (Durham, NC). A custom pipeline on the basis of GATK best practices was used for variant calling.¹⁶ All pathogenic (P) and likely pathogenic (LP) variants were confirmed using Sanger sequencing. Filtered variants and multiple-nucleotide variants (Data Supplement) were individually assessed for classification into three categories: (1) benign or likely benign (B/LB), (2) variant of uncertain significance (VUS), or (3) pathogenic or likely pathogenic (P/LP). VarSome (releases 9-10)¹⁷ was used to establish base American College of Medical Genetics (ACMG) classification criteria,¹⁸ which were then adjusted to align with Sequence Variant Interpretation-Working Group (SVI-WG) recommendations.¹⁹ Three researchers (A.S., C.S., and D.M.) reviewed all variants independently and, if needed, formed a consensus for conflicting classifications. Further details on sample inclusion, sequencing technology, and data processing are provided in the Data Supplement.