Limited diagnostic impact of duplications <1 Mb of uncertain clinical significance: a 10-year retrospective analysis of reporting practices at the Mayo Clinic.

Abstract

Copy-number variants (CNVs) of uncertain clinical significance are routinely reported in a clinical setting only when exceeding predetermined reporting thresholds, typically based on CNV size. Given that very few genes are associated with triplosensitive phenotypes, it is not surprising that many interstitial duplications <1 Mb are found to be inherited and anticipated to be of limited or no clinical significance. In an effort to further refine our reporting criteria to maximize diagnostic yield while minimizing the return of uncertain variants, we performed a retrospective analysis of all clinical microarray cases reported in a 10-year window. A total of 1112 reported duplications had parental follow-up, and these were compared by size, RefSeq gene content, and inheritance pattern. De novo origin was used as a rough proxy for pathogenicity. Approximately 6% of duplications 500 kb-1 Mb were de novo observations, compared with approximately 14% for 1-2 Mb duplications (p = 0.0005). On average, de novo duplications had higher gene counts than inherited duplications. Our data reveal limited diagnostic utility for duplications of uncertain significance <1 Mb. Considerations for revised reporting criteria are discussed and are applicable to CNVs detected by any genome-wide exploratory methodology, including exome/genome sequencing.