« The human brain evolution revolution | Gene Expression Front Page | What people should know about evolution »
April 04, 2005

Dissolving the dominance of dominance

Over the past few years one problem I have had when discussing issues of genetics with people are confusions issuing out of high school recollections of the dominant-recessive concept. Because this concept is introduced rather early on and emphasized through the ubiquitous Punnet Square there is often confusion about how polygenic additive traits can exist in light of the perceived centrality of dominance and recessive. Nearly a century ago R.A. Fisher theoretically reconciled (PDF) polygenic continuous variation (quantitative genetics) with a Mendelian paradigm, but obviously most people aren't going to be reading decades old population genetics papers, and I am not sure that such a mathematical exposition is easily internalized by most.

Recently I stumbled on to this philosophical diatribe against the concept of dominance in biology. The author basically seems to suggest that we should ditch the emphasis that we place on dominance and recessive and start from first principles and focus on the idea that in diploid organisms two copies of a given gene are expressed on any locus. To me this argument is made particularly powerful by the molecular revolution, for now we can illustrate how dominance might actually work in the context of the Central Dogma and a few basic concepts that are already partially digested from the cultural zeitgeist.

Consider that you have an organism which processes nutrient X via enzyme z. Enzyme z is synthesized like so:

Gene z - > transcription of mRNA -> translation of mRNA -> Enzyme z

Now, imagine that there is a "null" copy of Gene z, a mutation on the promoter prevents the initiation of transcription. In this case null-Gene-z results in lack of production of enzyme z, ergo, the inability to process nutrient X.

But in diploid organisms you have two copies of a gene. If you have a normal copy and a null copy there will be production of the enzyme, but only half the normal amount of enzyme will be generated because only half the normal amount of DNA is transcribed.1 If 1/2 of the normal concentration of enzyme is sufficient to break down enough of nutrient X so that the organism seems healthy one might say that the trait is "dominant," but at the molecular level the heterozygote is expressing only "half" the phenotype as a pure "wild type."

It might be that in a normal context there is almost never enough of nutrient X in the environment so that more than 50% of the enzyme z is necessary for optimal functioning (the abundance of nutrient X is the limiting factor). But there could be rare circumstances where nutrient X is very abundant and those organisms who have both copies of the functioning gene can take advantage of their enzymatic advantage and store the end product of nutrient X for future use when there is a deficit. In this case there is a clear fitness advantage that wild type homozygotes have over heterozygotes, but it is not normally observed over short time scales. Additionally, there is the possibility that gene z is prone to mutation in the somatic (body) cells that process nutrient X, so those individuals with some level of redunancy might have a long term life history advantage over heterozygotes who might harbor a multitude of sub-optimal cells which fail to process nutrient X by the end of their life because they have no "back up" plan.

In any case, I think that this narrative does have conceptual advantages over the dominance-recessive idea, and is simple enough that it can be introduced at the secondary educational level. As the author of the paper above notes it would also render terms like "codominance," "incomplete dominance," "penetrance," etc. redundant. And the idea of starting at the molecular level might also make an understanding of polygenic traits a bit more comprehensible, and in the future polygenic traits are going to loom large in the public imagination as the low hanging single locus fruit are picked clean and scientific press releases begin to skew toward more complex phenotypes.

1 - This is simplistic, there are a lot of issues relating to how/why/to what extent genes get expressed. One could perhaps concoct a model where the organism up regulates gene expression by producing a factor which induces more frequent binding of the polymerase when it "detects" that nutrient X is not being metabolized efficiently. We'll neglect such counter scenarios.

Posted by razib at 08:57 PM