The Cost of Natural Selection

In summary, the Cost of Natural Selection is the number of selective deaths required to fix an allele in a population. In 1957 the well known scientist J. B. S. Haldane presented a doubt about Darwin's concept of natural selection, called "the cost of natural selection." Scientists, advancing Haldane’s idea of "cost of selection," nowadays name different varieties as "cost of mutation," "cost of segregation," and "cost of substitution." In this essay, we will briefly explain this concept of "cost of selection," which is one of the difficult topics for population geneticists and therefore is not a very much discussed topic. The unavoidable, selective death of a gene to be substituted was first understood by the biologist J.B.S. Haldane. According to his calculation, the higher the intensity of natural selection, the higher the rate of selective death (or infertility) must be. A population cannot tolerate an indefinitely high rate of selective death, namely, if selection is too strong it will drive the population to extinction.

John Sanford, a horticultural geneticist and professor at Cornell University, writes:

Haldane was the first to recognize that there was a cost to selection which limited what it could realistically be expected to do...He calculated that in man, it would take 6 million years to fix just 1,000 mutations (assuming 20 years per generation). He could not then know that the number of actual genetic units is 3 billion, and that at least 1 million new mutations would be entering any hypothetical pre-human population each generation - most of which would require selection. Man and chimp differ by at least 150 million nucleotides representing at least 40 million hypothetical mutations (Britten, 2002). So if man evolved from a chimp-like creature, then during that process there were at least 20 million mutations fixed within the human lineage (40 million divided by 2), yet natural selection could only have selected for 1,000 of those. All the rest would have had to have been fixed by random drift - creating millions of nearly-neutral deleterious mutations. This would not just have made us inferior to our chimp-like ancestors; it would surely have killed us.

Let us give an illustrative description about this process of natural selection and its cost. When you go to a shop of 1000 coats to choose the best one, you don't have to try all of them because you know which size you need and which color and material you like. You can simply try on a few coats and select the warmest. Because of its good quality, you can hand it down to your children later. This is all possible because this selection involves human intelligence.

Now let us see how this selection would work in the process of natural selection, which is not intelligent but based on chance. If the coats would be not arranged by size nor the size would be indicated, you would have to go through the trouble of trying all of them and eliminating them one-by-one except the one you like. However, even this trouble is not even closely the same as the complex selection in nature.

Expanding our example, suppose that two coats became warmer due to genetic mutation. Because you don't desire to try all the coats in the shop, you bring 999 of your friends and relatives to randomly accept one coat each. Consequently, 998 coats will be eliminated as bad when your relatives and friends will freeze to death in them. This scenario of elimination illustrates an oversimplified episode of natural selection, where one allele2 substitution occurs at the cost of the loss of almost a whole generation of population.

In actual cases, the cost of one allele substitution is usually paid in increments spread over many generations. The total cost may then be equated to the loss of many generations multiplied by the number of individuals in one generation of population. This fact, which is not intuitively obvious, was pointed out by Haldane 25 years ago.

Obviously, the above described minimum-cost selection-by-elimination process is quite complex. Firstly, you might most probably not survive it, because your chance of getting one of the two warmer coats is only 1 in 500. Moreover, getting one of the two coats would only slightly increase your chance of survival, because wearing the better coat cannot protect you from death if you happen to be in a wrong place at the wrong time. This implies that the alleles of the warmer, coated individuals would be lost and the mutation would have to reappear many times before becoming fixed in the population.

So, although you would get the warmer coat and luckily survive the severe cold, you might still not be able to hand over the coat to your children because the gene that determined coat warmth might be heterozygous4 in yourself. Choosing a heterozygous marriage-partner with a warm coat would be favorable for your future children. (To choose a marriage-partner with a superior coat, rather than one with similar genes like yours, will assure that your children will get a good coat. This will increase the chances of survival of your lineage. This is natural selection as it is. It hardly makes these theories acceptable: the concepts of fitness and selection, based on the community of genes rather than the superiority of individuals.)

It is important to note here that the actual evolution of an individual's coat in a complex environment involves many genes, not just one as given in the above example. Thus, it is very difficult to accept that the population can evolve a better coat without some intelligent guidance but through a rather random, complex, and costly allele substitution. Moreover, all this becomes even more difficult to accept when we take into account the endless number of other changes in physiology and behavior of individuals by selective elimination.

NOTE: Interestingly, Haldane’s dilemma and the problem of protein polymorphisms forced Kimura and friends to argue that the majority (say 90+%) of molecular evolution could not be Darwinian. Thus 90% of the genome must be subjected to neutrality, not selection.

The Description of the 'Cost of Selection' by Professor John Sanford

The fact that all people are mutant makes selection much more difficult. If we were to simply select against all "mutations," it would mean no one could reproduce - resulting in instant extinction. Obviously, this selection strategy creates a "reproductive cost" that is too high. It is widely acknowledged that we each inherit thousands of deleterious mutations, so collectively, as a population, we carry many trillions of deleterious mutations. However, to make the problem easier, let us limit our attention just to the 600 billion new mutations that entered the human gene pool within our own generation. Since we cannot simply select against "mutants," we will have to select between individuals who are "more mutant" versus those who are "less mutant" (as we will see, recognizing “more mutant” versus “less mutant” is a huge problem in itself). All this selection must cost us considerably less than 33% of the population per generation.

Let me try to illustrate the extent of the cost problem which is associated with selecting against 600 billion mutations. If we have a population of 6 billion people, then maximally one third of them could be "eliminated" (i.e. from having no children for some reason). This is 2 billion people (try to imagine that - this thought should be enough to make even the most cold-blooded eugenicist shudder). Eliminating 2 billion people from mating would only eliminate 100 x 2 billion = 200 billion new mutations. This would still leave 400 billion new mutations for the newly added genetic burden for the next generation. Even if we assume that two-thirds of the remaining mutations are perfectly neutral, we still have 133 billion deleterious mutations added to the population. We just cannot get rid of enough mutations and still maintain population size.

Even if two-thirds of the mutations are neutral, and in addition we doubled selection intensity (although we certainly cannot afford to spend two-thirds of our population), it would still leave 67 billion new deleterious mutations for the next generation. The cost of selection clearly limits the number of mutations we can eliminate per generation, and the known mutation rate for humans is way too high to be countered by any level of selection. Therefore, mutations will continue to accumulate, and the species must degenerate. Can you see that the cost of selection is rather a mind-boggling problem when viewed on the genomic level?

…we can only afford to "fund" progressive selection for beneficial mutations after we have paid for all other reproductive costs, including all costs associated with eliminating bad mutations. As we have already seen, there are so many bad mutations that we cannot even afford to pay just the reproductive cost of eliminating them. Since we cannot afford to stop degeneration, we obviously have nothing left to fund progressive selection. There is just one way around this. In the short run, we can fund progressive selection for a very limited number of traits - if we borrow "selection dollars" from our long-term struggle against bad mutations. However, this means that any short-term progress in terms of specific beneficial mutations is paid for by faster genomic degeneration in the long run.