Penetrance and Expressivity

Penetrance and Expressivity

By Leos Kral

Penetrance and expresivity are two concepts that are different yet related. They are often confused for one another. One distinguishing characteristic is that "penetrance" is a qualitative concept and "expressivity" is a quantitative concept. Simply put, "penetrance" refers to whether a phenotype is expressed for a particular genotype, and "expressivity" refers to the degree to which a phenotype is expressed when it is expressed. Note that expressivity is dependent on penetrance. It is not possible to measure the degree of expression if a genotype is not expressed as a phenotype, that is, if the genotype does not penetrate to the phenotype.

Expressivity is not really a problem in genetic analysis as long as one is aware of it. For example, the merle trait in Australian Shepherds displays a wide expressivity range. Bart, my merle aussie, at first glance looks like a black tri. However, if one looks closely, there is a bit of merling on one side of his face, along some of his legs, and a few grey hairs can be found in small patches in the black coat. The merle gene in Bart would be said to display low expressivity. Bart's dad, on the other hand, is predominantly light gray with a random distribution of small black splotches. The merle gene in Bart's dad would be said to display very high expressivity.

In both Bart and in Bart's dad the merle gene pentrates to the phenotype. That is, we would say that both dogs are merle dogs, regardless of the degree of merling. If we were to draw a genetic pedigree of Bart and his parents and track the inheritance of the merle trait, we would fill in both Bart's square as well as Bart's dad's square to indicate that they are both merle. We would not care to what degree the merle trait is expressed.

Lack of penetrance in some individuals, on the other hand, can pose a problem to the analysis of genetic pedigrees. When pedigrees are analyzed to determine how a particular trait is inherited, the phenotypic pattern in the pedigree is matched with all possible variations of hypothetical genotypic patterns to determine which genotypic pattern could account for the observed phenotypic pattern. Obviously, if some individuals have genes for the expression of a particular trait but don't express that trait, then the genotypic vs. phenotypic pattern matching will be confounded.

As an example of how lack of penetrance in some individuals can confound pedigree analysis, let's examine the hypothetical pedigree of the inheritance of hairy eyeballs from the previous pedigree analysis article. The following pedigree is very suggestive of sex linked recessive inheritance of hairy eyeballs.

Suppose that owners of the dogs pointed to by arrows in the figure above decided to breed those dogs (certainly not recommended). The mating resulted in the production of 5 offspring as shown in the figure below.

Individuals labeled "1" and "2" do not have hairy eyeballs, but their daughter labeled as "3" does have hairy eyeballs. At first glance, the fact that these two individuals produced an affected daughter would seem to invalidate the previous conclusion that hairy eyeballs is inherited as a sex linked recessive trait. If the trait was sex linked recessive, then it would be necessary that the father (labeled "2") also have hairy eyeballs. (Sex linked recessive inheritance has the following characteristic pattern: An affected daughter has to have an affected dad.) At this point, we would have to revise our previous conclusion of sex linked recessive inheritance to the more likely mode of autosomal recessive inheritance. (Autosomal recessive inheritance has the following characteristic pattern: two unaffected individuals can produce affected daughters [and sons] in addition to unaffected daughters and sons. If the trait were sex linked recessive, then the characteristic pattern is that two unaffected individuals can produce affected sons but can not produce affected daughters.)

Now suppose that the following breeding was also performed:

Affected individuals labeled "3" and "4" were bred (also not recommended) and produced both affected and unaffected offspring. This is really a big problem for our analysis, because this pattern invalidates our previous conclusion of autosomal recessive inheritance of hairy eyeballs. Why? Because we are now faced with two characteristic patterns ("rules of pedigree analysis") which are contradictory:

If two unaffected parents ("1" and "2") produce affected offspring then the trait must be recessive.
If two affected parents ("3" and "4") produce unaffected offspring then the trait must be dominant.

So, what do we do now? A trait can not be both dominant and recessive. Well, first we must realize that the two "rules of pedigree analysis" assume that the trait in question is 100% penetrant. That is, that all individuals that have a genotype that should produce hairy eyeballs actually will have hairy eyeballs. If a trait is not 100% penetrant, then the two rules above are not absolute. For example, if the trait is 70% penetrant, that means that only 7 out of 10 individuals (on avearge) that have the hairy eyeball genotype will actually have hairy eyeballs. The other 3 individuals will not have hairy eyeballs even though they have the hairy eyeball genotype. Therefore, since the pedigree in question presents us with contradictory patterns of inheritance, it is likely that hairy eyeballs is not 100% penetrant.

So, to analyze pedigrees of traits that are not 100% penetrant, we would re-examine those pedigrees and note those individuals who would have to be genotipcally affected but are not phenotypically affected for each of the different modes of inheritance. In the next two samples of such re-examined pedigrees, those individuals that should have hairy eyeballs but do not are symbolized with light gray symbols.

If trait is sex-linked recessive:

If trait is sex-linked dominant:

To determine which is the actual pattern, we would have to perform statistical likelihood analyses taking into account such factors as different levels of penetrance, frequency of affected individuals in outcross matings, and sex of affected individuals in outcross matings (the last two would have to be derived from other pedigrees not shown). Since genetic analysis is primarily an application of probability and statistics, pedigrees must be large enough or there must be a large enough number of different pedigrees tracing the same trait such that the conclusions can be statistically significant. These sample pedigrees just do not contain enough information to allow us to make a determination of how the trait hairy eyeballs is inherited.

Note that even with a larger sample size the mode of inheritance may not be unambiguous. For example, we may find that the mode of inheritance is either autosomal dominant if penetrance is 30%, or autosomal recessive if penetrance is 70%. This sort of places us in a Catch-22 because we can't determine the degree of penetrance if dominance/recessiveness is not known and we can't determine dominance/recessiveness if the degree of penetrance is not known. While not having an absolute answer may not be totally satisfying, narrowing down the choices is really all that is required if the ultimate goal is to find the gene responsible for the trait such that a diagnostic test can be devised.

Participation of all aussies in the Australian Shepherd Health Registry will ensure that a large enough resource of genetic pedigree information will be available for future genetic analyses of those genetic diseases that effect the Australian Shepherd.