All articles

Reduced penetrance has to be argued.

An unaffected parent who carries the candidate variant is evidence against it. Reduced penetrance can answer that, if argued from the gene's own data.

A boy of seven with a cardiomyopathy and a conduction defect has a trio exome. The best candidate is a rare missense variant in a gene with a well-established dominant association, inherited from his father. The father is forty-one, has never had a cardiac complaint, and has never had an echocardiogram. The draft report classifies the variant as likely pathogenic and adds a sentence: 'inherited from an apparently unaffected parent, consistent with reduced penetrance'.

The sentence may be right. Nothing in the case supports it. The father has not been examined, the gene's penetrance in carriers has not been looked up, and the variant's absence from population data has not been weighed against the number of unaffected people who would carry it if penetrance were as low as the sentence implies.

What the field knows

Penetrance is the share of people with a genotype who show the phenotype, and for most dominant conditions it is below complete and above the figure a report would need for a casual appeal to it. Cooper and colleagues (2013) reviewed the mechanisms behind incomplete penetrance and variable expressivity, from modifier variants and allelic differences to age, sex and environment, and argued that penetrance is a property of a specific variant in a specific context. Population sequencing has since shown how much of it there is. Studies that start from carriers in biobanks rather than from affected families, such as the analyses described by Wright and colleagues (2019) and the gnomAD constraint work (Karczewski et al., 2020), find that many variants classified as pathogenic from family studies appear in unaffected adults at frequencies that imply far lower penetrance than the clinical literature suggested.

The interpretation framework treats segregation as evidence in both directions. Under the ACMG and AMP guidelines (Richards et al., 2015), co-segregation with disease in a family supports pathogenicity, and non-segregation, a variant in an unaffected relative, is evidence against it, with a caution that reduced penetrance should be considered where it is known for the gene. The Sequence Variant Interpretation working group's subsequent guidance (Jarvik and Browning, 2016, on segregation; Biesecker and Harrison, 2018, on evidence weighting) ask that the strength of segregation evidence be calculated from the pedigree rather than asserted, and that reduced penetrance be invoked only where the gene's penetrance is documented.

For the cardiac genes in particular, penetrance is now measured. Registry and biobank studies of the common cardiomyopathy genes report age-dependent penetrance by gene and sometimes by variant class, so that for a given gene it is possible to say what fraction of carriers are expected to have a detectable phenotype by forty. That figure is the thing the sentence in the report is standing in for.

Where practice breaks

Reduced penetrance is used as a rescue. The candidate was selected because it fits; the pedigree then contradicts it; the contradiction is absorbed by a phrase and never weighed. The same phrase would be written for any gene and any parent, which is the sign that it carries no information.

The unaffected parent is unaffected by report, with no examination behind the word. The father in the opening case has not had an echocardiogram or an electrocardiogram. He may have a subclinical phenotype that would turn non-segregation into segregation and settle the case. The cheapest piece of evidence in the investigation is not collected because the pedigree form recorded him as well.

Penetrance is appealed to without a figure. For the gene concerned there may be published penetrance by age; the scientist writing the report rarely has it to hand, and the pipeline does not supply it. So the sentence stays qualitative, and a reader cannot tell whether an unaffected carrier at forty-one is expected in one carrier out of three or one in fifty.

The appeal, once written, propagates. The variant goes into the classification archive as likely pathogenic with a segregation note; the next laboratory sees the classification, counts it as evidence, and the penetrance question is never asked again. An unexamined father becomes, two reports later, a supporting citation.

The family bears the consequence. A likely pathogenic classification in a dominant cardiac gene leads to cascade testing and surveillance for the father's siblings and their children. If the variant is in fact benign and the boy's condition has another cause, a wide family has been placed under surveillance on the strength of a sentence, and the real cause has not been looked for.

How Tulparus handles it

The workbench records the phenotype of every family member with the same four states it uses for the proband: present, absent, unknown, not tested. A parent recorded as well on the form, with no examination recorded, carries the relevant features as not tested, and the segregation evidence for a variant they share is computed from that: an untested carrier is a gap, and the case says so instead of counting it as a counter-example.

Where the parent has been examined and the feature is absent, the segregation evidence is weighed against the gene's documented penetrance, drawn from the knowledge store with its source and release. The hypothesis page shows the calculation: for this gene, at this age, the expected fraction of carriers with a detectable phenotype, and therefore how much an unaffected carrier counts against the variant. If no penetrance data exist for the gene, the page says that too, and the phrase 'reduced penetrance' cannot be used as if they did.

The next step writes itself. For the boy above, the case would list the father's echocardiogram and electrocardiogram as the most informative missing evidence, with what each outcome would do to the hypothesis: a subclinical phenotype would convert the pedigree into supporting evidence; a clean study at forty-one would weigh against the variant by an amount the penetrance data set. The report, rendered from the case, would say that the variant was inherited from a parent whose cardiac status has not been assessed, and that assessment is recommended.

The classification stored on the case carries the segregation reasoning as structured evidence, with the relatives counted, so that a later reader, or a later release of the penetrance data, can revisit it.

What you can do tomorrow

Most of this is a habit about the unaffected relative.

  • Before writing 'reduced penetrance', write the penetrance. If you cannot find a figure for the gene, say that the gene's penetrance is not established rather than that it is reduced.
  • Treat an unexamined relative as untested, in the pedigree and in the report. Recommend the examination that would settle it, and name it.
  • When a variant is inherited from a reportedly well parent, make that parent's clinical assessment the first next step, before any further sequencing.
  • Record segregation evidence as a calculation from the pedigree, with the relatives counted, so that the next laboratory sees the reasoning as well as the conclusion.

Sources

  1. Cooper DN, Krawczak M, Polychronakos C, Tyler-Smith C, Kehrer-Sawatzki H. Where genotype is not predictive of phenotype: towards an understanding of the molecular basis of reduced penetrance in human inherited disease. Human Genetics, 2013.
  2. Wright CF, West B, Tuke M, et al. Assessing the pathogenicity, penetrance, and expressivity of putative disease-causing variants in a population setting. American Journal of Human Genetics, 2019.
  3. Karczewski KJ, Francioli LC, Tiao G, et al. The mutational constraint spectrum quantified from variation in 141,456 humans. Nature, 2020.
  4. Richards S, Aziz N, Bale S, et al. Standards and guidelines for the interpretation of sequence variants: a joint consensus recommendation of the American College of Medical Genetics and Genomics and the Association for Molecular Pathology. Genetics in Medicine, 2015.
  5. Jarvik GP, Browning BL. Consideration of cosegregation in the pathogenicity classification of genomic variants. American Journal of Human Genetics, 2016.
  6. Biesecker LG, Harrison SM. The ACMG/AMP reputable source criteria for the interpretation of sequence variants. Genetics in Medicine, 2018.

One well parent, one variant.See what the pedigree really says.

A demo takes 20 minutes, screen shared, on synthetic cases. Bring the question from one of your own.