Main Problem with Dawkins' Selfish Gene

Richard Dawkins defined a gene as (The Selfish Gene, 1976, p. 38):

"The gene is defined as a piece of chromosome which is sufficiently short for it to last, potentially, for long enough for it to function as a significant unit of natural selection."

Actually, it is not necessary for something to last intact for it to be a unit of selection. Imagine for example two (not linked) genes A,B that occur in a frequency of 50% each. If there is random mating, then 25% of people in the population will have the AB combination.

The AB combination potentially does not survive meiosis, but there will always be AB individiuals in the population, or at least for a long enough time for them to be selected for.

Thus, if AB confers an advantage or disadvantage, it will be positively or negatively selected. If it's advantageous, then this will lead to a rise of A and B genes in the population. But that is a consequence of selection. One cannot say that A and B were indeed the things that were selected.

Consider for example the co-evolution of small palates and small teeth. Since there are people with over-crowed teeth, and people with very widely spaced teeth (both of whom have problems), it's likely that these two features are controlled by at least two separate genetic factors. It was the combination of these factors that was selected for.

Also, suppose that AB is a selective +, BC is a -, CD is a + and DA is a - (think of A,B,C,D in a circle). Selection will tend to increase AB's. This will lead to the inccrease of B's. But BC's will tend to decrease B's. Each gene participates in a "good" and a "bad" combination. By observing the gene-level frequencies of A,B,C,D we would know nothing about the process taking place.

In conclusion, it is not just genes that are the targets of selection, but also genetic patterns consisting of more than one genes. In essence, everything that isn't transient (*) but is re-formed often enough in a population to be the target of selection, will be selected.

Gene-level shifts of frequency should not then be viewed in term's of a gene's selfish "fitness" or advantage but as the sum of the advantages of gene combinations in which the gene participates in a population.

(*) Like an individual man's whole genome, which won't re-appear on the planet, unless there is cloning of course.

Posted by Dienekes at May 17, 2003 12:54 PM | PermaLink