President's Corner - Genetic Penetrance
Published on
Penetrance may be defined as the fraction of patients with a specific genotype who are affected. Penetrance is critically important in variant interpretation and in reaching medical decisions. As just one example, prospective parents who receive a lab report indicating that their fetus carries an incompletely penetrant sequence variant will want to know the probability that their baby will be sick.
Unfortunately, we have today even rough estimates of penetrance for only a handful of sequence variants (see for example Cooper et al. 2013 Hum Genet 132:1077-1130; Eloy et al. 2016, PLoS Genet 12(2):e1005888; Minikel et al. 2016 Science Translational Medicine 8(322):322ra9). For the vast majority of variants, we do not have even rough estimates. At best we may know that some relatives of probands carry the variant, but are apparently unaffected indicating that penetrance is “incomplete”.
Measurement of penetrance is complicated by variable definitions. For example, if a disorder has three primary and several secondary clinical features, how to we define affected? Does affected mean just one of the primary features or is it all three primary plus at least two secondary? Penetrance may also be age and gender dependent, as is the case for the HFE p.Cys282Tyr variant in hemochromatosis (Seckington and Powell 2015 Gene Reviews, https://www.ncbi.nlm.nih.gov/books/NBK1440/).
For dominant and X-linked recessive disorders, estimates of penetrance can often be achieved with reasonable effort. Many family members will carry the variant in question. It then is just a matter of determining if the carriers have the relevant clinical features. For recessive disease, measurement of penetrance is usually much tougher. Unless the homozygous genotype is relatively common, we may need to resort to large research population databases such as gnomAD. The main problem with such databases of course is that we’re not sure if the homozygous individuals are truly unaffected.
The most recent ACMG Guidelines for variant interpretation state “Until better guidance is developed, an interim solution is to report these variants (those with relatively low penetrance) as “risk alleles”….” (Richards et al 2015). Labs can follow this recommendation if we have decent numeric estimates of penetrance and if “low penetrance” is quantitatively defined. Perhaps in future we will use the standard five interpretation categories for sequence variants with penetrance values say ≥ 10%, and use the designation “risk allele” for variants with penetrance values say < 10% but ≥ 1%. For variants with penetrance < 1%, benign may be the best classification even if for example a GWAS study demonstrates they have some minor effect on disease risk.
We clearly need to devote many more resources toward measurement of penetrance. Perhaps the NIH’s Precision Medicine initiative will be a considerable help. Until then, labs and providers will continue to be plagued by uncertainties in penetrance.