Abstracts

Rationale: Historically, mosaicism was challenging to detect using conventional sequencing methods (i.e. Sanger sequencing) and has likely represented an under-reported cause of genetic disease. Capture-based NGS with high read-depth, mosaic-aware bioinformatics, and reduced random sequencing error provides a more sensitive and quantitative approach for the identification of mosaicism. Methods: We conducted a retrospective study to identify genes with a high rate of mosaicism tested by capture-based NGS at our diagnostic testing laboratory. The study included only likely pathogenic or pathogenic variants (L/PV) in 188 genes associated with autosomal dominant (AD) or X-linked disorders (XL).  We also limited analysis to genes with a high clinical sensitivity (>9 L/PV observed in the 2 year time period). These genes with high clinical sensitivity were assigned to phenotypic groups based on the major disease association and rate of mosaicism was calculated by group. Results: We observed that 1.4% (111/7859) of L/PVs were classified as mosaic. Mosaic variants were identified in 59/188 genes and accounted for 4.8% (111/2314) of all L/PVs in these genes. Mosaic L/PVs were observed in 9-39% of reads in AD genes, 10-25% of reads in XL genes in females, and 13-88% of reads in XL genes in males. The highest rate of mosaicism was observed in genes associated with neurodevelopmental disorders (NDD) where epilepsy is a common comorbidity or a major clinical indication for affected individuals. Specifically, 8.4% (17/202) of L/PVs were mosaic in brain malformation syndromes (BMF) and 3.2% (57/1778) of L/PVs were mosaic in genetic epilepsy and intellectual disability disorders (Epi/ID). Notably, some BMF genes with high rates of mosaicism were PAFAH1B1 (15.8%, 3/19), FLNA (13.6%, 3/22), and TUBA1A (12.0%, 3/25), and Epi/ID genes included SCN2A (7.8%, 6/77), PCDH19 (6.3%, 3/48), and MECP2 (4.3%, 4/93). In contrast to this high rate of mosaicism in genes associated with NDD, we observed that only 1/2378 and 1/658 L/PVs were mosaic in genes associated with cardiomyopathy/arrhythmia and neuromuscular disorders, respectively. Parental mosaicism was observed in 32 cases; mosaic L/PVs were observed in 2-35% of reads including 8 parents who had mosaic L/PVs in Conclusions: Collectively, these data emphasize that mosaic L/PVs are more common than previously reported in some genes associated with NDD and epilepsy. Diagnostic testing for patients with NDD should include capture-based NGS with high read-depth and mosaic-aware variant calling.  Additionally, parental testing by NGS or a comparable quantitative method, should be considered when the child has a L/PV in a gene with a high mosaic rate. Funding: None