Recessive Lethals Are Mainly Eliminated by Selection against Heterozygotes

This principle was first established for recessive lethals in Drosophila pseudoobscura, in a classic study by Wright et al. (1942). About 14% of third chromosomes from this fly carry a recessive lethal (see Box 13.2). Suppose that there are n genes that can mutate to recessive lethals. The frequency of chromosomes that do not carry recessive lethals is (1 – p)ⁿ ~exp(–np) = 0.86, where p = µ/s; thus, np is estimated as –ln(0.86) = 0.15. Now, the total rate of spontaneous mutation to recessive lethals was measured in the laboratory as nµ = 0.003, so we can estimate the selection coefficient acting to eliminate recessive lethals in the homozygous state as 0.003/(–ln(0.86)) ~ 2%. In other words, recessive lethals segregate for about 1/0.02 = 50 generations before being eliminated. Assuming random mating, in each generation, a fraction 0.02 × 2 np = 0.012 of the population fails to reproduce because it is heterozygous for a lethal at one of the n loci on the third chromosome.

We can now check that this is much greater than the loss of fitness due to homozygosity, as assumed in the calculation. Given that there are about n = 900 genes on the third chromosome that can mutate to recessive lethals, the allele frequency at each is p = 0.15/900 = 1.7 × 10^–4. Therefore, np² = 2.6 × 10^–5 are lost when the lethal is expressed in homozygotes, which is indeed far lower than the loss of fitness through the much more common heterozygotes. Similar selection is likely to act against recessive alleles that cause human disease. In some populations, however, recessive mutations are mainly eliminated in homozygotes that arise from matings between relatives. We consider the effects of inbreeding in more detail on pp. 515–518.

The number of lethal mutable genes on the third chromosome of D. pseudoobscura is about 900, which is much higher than the 285 estimated by Wright et al. (1942). The higher figure comes from the total gene number of 13,600 genes (Adams et al. 2000), assuming that one-fifth of the genome is on chromosome 3 (it is equivalent to 2R in D. melanogaster, one of five equal size arms), and taking one-third of the genes to be lethal mutable. The remaining figures come from Wright et al. (1942). (The discrepancy between the estimate of gene number from Wright et al. and that used here arises because of variation in the rate of mutation to lethals between genes. The figure of 285 is an effective number that essentially counts the more highly mutable genes.)