Background The genetic aetiology of non-aneuploid fetal structural abnormalities is typically investigated by karyotyping and array-based detection of microscopically detectable rearrangements, and submicroscopic copy number variants (CNVs), which collectively yield a pathogenic finding in up to 10% of cases. We propose that exome sequencing may substantially increase the identification of underlying aetiologies.

Methods We performed exome sequencing on a cohort of 30 non-aneuploid fetuses and neonates (along with their parents) with diverse structural abnormalities first identified by prenatal ultrasound. We identified candidate pathogenic variants with a range of inheritance models, and evaluated these in the context of detailed phenotypic information.

Findings We identified 35 de novo single nucleotide variants (SNVs), small indels, deletions or duplications, of which three (accounting for 10% of the cohort) are highly likely to be causative. These are de novo missense variants in FGFR3 and COL2A1, and a de novo 16·8 kb deletion that includes most of OFD1. In five further cases (17%) we identified de novo or inherited recessive or X-linked variants in plausible candidate genes, which require additional validation to determine pathogenicity.

Interpretation Our diagnostic yield of 10% is comparable to, and supplementary to, the diagnostic yield of existing microarray testing for large chromosomal rearrangements and targeted CNV detection. The de novo nature of these events enables couples to be counselled as to their low recurrence risk. This study outlines the way for a substantial improvement in the diagnostic yield of prenatal genetic abnormalities through the application of next generation sequencing.