OBJECTIVE: Cystic fibrosis is one of the most common autosomal recessive conditions. The disease is caused by mutations to the CFTR gene that include single-nucleotide variants (SNVs), short insertions or deletions (indels), and large copy-number variants (CNVs). Guidelines recommend routine carrier screening via targeted genotyping of 23 frequent variants, including SNVs and indels, but not CNVs; CNV screening is recommended only when a reproductive partner is a known carrier. Here we assess the performance and clinical impact of routinely screening for SNVs, indels, and CNVs in a next-generation sequencing (NGS)-based expanded carrier screen (ECS).
DESIGN: Retrospective analysis of pathogenic variants observed in a large deidentified cohort of ECS patients.
MATERIALS AND METHODS: Pathogenic variants in CFTR from a cohort of 103,718 patients were discovered via a validated NGS-based ECS. A custom algorithm identified CNVs via statistically significant relative deviations in NGS read depth, with downward depth deflections signifying deletions and upward deflections indicating duplications. Approximate CNV breakpoints were inferred from the NGS-depth profile across the gene. Positive CNVs were orthogonally assessed via multiplex ligation-dependent probe amplification (MLPA). CFTR CNV sensitivity was explored across a range of length scales via in silico simulations.
RESULTS: The observed carrier rate for cystic fibrosis was 3.3% (N=3,394). The 23 commonly screened variants accounted for approximately 79% of observed carriers, and 98.7% of carriers had pathogenic SNVs or indels in the CFTR gene. Critically, the remaining 1.3% of carriers harbored a pathogenic CNV spanning at least one exon. We observed a diversity of breakpoints (25 unique CNVs in total), suggesting that a bioinformatics pipeline must have the flexibility to detect novel CNVs to maximize the detection rate of carriers. Simulations demonstrate robust detection of CNVs across a range of length scales, and further analysis reveals that NGS-based CNV detection has expected accuracy comparable to MLPA.
CONCLUSIONS: CNV detection is required to maximize identification of cystic fibrosis carriers and can be applied to all patients undergoing NGS-based ECS validated to detect these complicated variants. Clinical guidelines recommending screening of only the 23 most frequent variants miss critical identification of carriers and should be revisited.