Updated penetrance estimates for recurrent copy number variants - an improved definition and formula

Many copy number variants (CNVs) are implicated in neurodevelopmental disability, but exhibit incomplete penetrance. The definition of penetrance is often unclear. In published literature, penetrance typically includes the background risk of disease, while clinicians tend to exclude risks unrelated to the genetic variant. We propose a more clinically relevant definition of penetrance and develop a new formula for this. These changes are applied to existing data sources to produce updated penetrance estimates. Our findings indicate that most CNVs studied have significantly lower penetrance than previously published. Eleven CNVs, previously described as low-penetrant, are recalculated as having a penetrance close to 0% for intellectual disability. These include 1q21.1 proximal duplications [RBM8A], 2q11.2 deletions [TMEM127], 2q13 proximal deletions and duplications [NPHP1], 6q16 duplications [SIM1], 13q12 deletions [CRYL1], 15q11.2 duplications [NIPA1, NIPA2], 15q13.3 duplications [CHRNA7], 16p12.2 duplications [CDR2], 16p13.11 duplications [MYH11] and Xp22.3 duplications [SHOX]. Previous estimates of CNV penetrance, which ranged from 10-40% have been recalculated as 1-10%. In conclusion, many previously published estimates of CNV penetrance are inflated. Re-evaluation of existing data reveals lower and more accurate penetrance estimates for intellectual disability. This has important implications for diagnosis, genetic counselling, and prenatal reporting of recurrent CNVs.