President’s Corner - Penetrance and Risk Variants
Published on
We recently updated our internal variant interpretation guidelines at PreventionGenetics. One change was to institute a new interpretation category: “risk variant”. This change is consistent with the 2015 ACMG interpretation guidelines (Richards et al. 2015. PMID 25741868), but we elaborated upon the limited recommendations in these guidelines.
The concept of penetrance is essential to defining a risk variant. Penetrance for a specific variant is defined as the fraction of individuals who carry the variant that are affected. Note that penetrance
values are typically different for individuals who are heterozygous versus hemizygous or homozygous. Penetrance values are also often ae dependent and may be sex dependent.
Penetrance values are known today for only a relatively few variants, but this will gradually change as large databases with both sequence and clinical information become available (see for example Bycroft et al. 2018. PMID 30305743; and commentary in Science Kaiser and Gibbons. 2019. PubMed ID: 30606827). Note, however, that there will usually be ascertainment biases in these databases. Consider, for example, a modestly rare pathogenic variant for dominant Intellectual Disability (ID) with a true penetrance of 0.50. In cohorts consisting of relatively healthy volunteers, there are unlikely to be any subjects with ID. Therefore, the observed penetrance in this dataset will be 0.00. On the other hand, if researchers are studying a group of affected individuals with ID, then the observed penetrance will be 1.00. Neither cohort will provide an accurate estimate of the true penetrance.
For purposes of variant interpretation, our PG penetrance cutoffs are 0.25 and 0.01. If a variant has a penetrance >0.25 for a rare disorder, and there is no contrary evidence, then this variant is Pathogenic. If the variant is clearly known to modify risk and if the penetrance is >0.01 but ≤0.25, then we label the variant as a risk variant. If the variant has a penetrance ≤0.01, then we label the variant Benign. If penetrance estimates are not available, then we will not use the risk variant label. However, when this label is used, penetrance values should be clearly stated within the test reports.
Other labs will undoubtedly choose different penetrance cutoff values for risk variants. These values merit considerable thought and discussion within the clinical genetics community. Hopefully the community will reach a reasonable consensus prior to the next update of the ACMG guidelines.
Note that the results of Genome Wide Association Studies (GWAS) (also known as case/control studies), have shown us that many common variants will slightly modify the course of disease. Such variants may even be useful in risk algorithms. However, because these common variants individually have a small impact on disease, and have very low penetrance values, we feel it is best at this time to label them as Benign.
Note also that vast numbers of currently unidentified risk variants also undoubtedly exist. Therefore many of the variants we currently interpret as Likely Benign and Benign will eventually turn out to slightly increase (or decrease) a person’s risk of disease.
Risk variant interpretation represents a nice step forward in categorizing variants that have too low penetrance to be labeled pathogenic, yet are clearly involved in disease risk. Widespread implementation of this interpretation category will make our test reports even more helpful to patients.