Re: did the spartan breeding project leave any measurable effect?

Dear Cris,

Eugenics resurfaces from time to time throughout human history. It is an appealing idea to many, probably because of the success that people have had with the selective breeding of plants and animals, starting before recorded history. I suppose it is natural to want to improve our own species somehow.

Let's turn to the Spartans for a moment. Suppose that there are a few hereditary disabling conditions  in this, like all other, human populations. Let's really simplify the problem and say that there is a recessive mutation "a" that makes infants die under Spartan selection. The normal allele "A" is more common. Population geneticists use a term that is useful to us here called the allele frequency. Everyone has two alleles of this gene (it's not sex linked). Let's count up all the alleles in the entire population and see how many A vs. a alleles there are. We will set this up as a fairly extreme example and set the allele frequency in the population at A = 0.9, a = 0.1 (10% of the alleles are a). Let's assume for the moment that the starting population is in "equilibrium," meaning that there is random mating of people regardless of their genotype. There are four genotypes, but two are equivalent (Aa and aA).

The frequency of the different genotypes in the starting population will therefore be:

AA (homozygous normal) = 0.9 x 0.9 = 81%

Aa (heterozygous) = 0.9 x 0.1 = 9% x 2= 18%

aa (homozygous for disease) = 0.1 x 0.1 = 1%

So one in every hundred infants in the starting population are left out to die by the Spartans. The parents of the aa babies are not killed; they are allowed to have more children in the hopes that they will have "good" ones. The next generation of breeders gives us the starting frequency of genotypes:

AA = 81.81% (81/99 x 100)

Aa = 18.18% (100% - frequency of AA)

aa = 0% (killed)

So the allele frequency is now A = 0.904534 (square root of 0.8181), a = 0.095465. You can see the effect of selection on the allele frequency. This generation produces aa babies at the frequency of 0.911%. Carry this out a few generations, and note that the rate of decrease in the allele frequency of a decreases. In other words, selection against a fairly common recessive allele has a strong effect at first, then it tapers off. Why?

Well, imagine that the frequencies of the alleles are A = 0.99, a = 0.01. As diseases go, this is still a pretty high allele frequency (1% of the alleles are a). At these allele frequencies, the frequency of the different genotypes in the starting population will be:

AA (homozygous normal) = 0.99 x 0.99 = 98.01%

Aa (heterozygous) = 0.99 x 0.01 = .099% x 2= 1.98%

aa (homozygous for disease) = 0.01 x 0.01 = 0.01% (1/10,000 births)

If the population of Sparta is around 10,000, they might not even see one of these affected kids in a single generation. Hence, their selection will have virtually no effect.

You can read up on population genetics at:
 http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=Books

Try "Modern Genetic Analysis" or "An Introduction to Genetic Analysis" at the link above.

That takes care of the mathematical argument. Unless the Spartans are trying to rid themselves of a very common recessive mutation, they will be unable to do so by selecting against homozygous recessives. At best, they will drive down the allele frequency to the point at which their selection has no effect (depends on the population size), and then it will stay there.

Now let's talk about biology. One of the most common genetic disorders of populations of European descent is cystic fibrosis:
 http://www.ncbi.nlm.nih.gov/htbin-post/Omim/dispmim?219700

The gene is CFTR, described at:
 http://www.ncbi.nlm.nih.gov/htbin-post/Omim/dispmim?602421

The incidence of the disease in Caucasians is 1/3,300:
 http://www.urmc.rochester.edu/genetics/CysticFibrosis.htm

This gives an allele frequency of about 1.7% for the mutant allele. These people typically do not reproduce, so from the standpoint of allele frequencies, it is exactly like the Spartan situation. The allele frequency has stabilized at over 1.7%. How did it get this high in the first place? There is some speculation that heterozygotes for cystic fibrosis are more resistant to cholera, so AA (actually CFTR<+>/CFTR<+> individuals were selected against during cholera epidemics.

This points out that a "bad" gene is not always a "bad" gene. In some circumstances, it can be advantageous.

The other (nonscientific) question, is to ask what the Spartans gained from this practice? They did not have to care for children who were sickly (but they did nurse their war wounded, eh?), so they had energy to devote to other things. Probably more important was the harsh message that it sent to all of the members of their society that they did not coddle weakness, something that might have helped them attain social cohesiveness.

Thanks for an interesting question.

Yours,

Paul Szauter
Mouse Genome Informatics