Virgin birth, genetic variation and inbreeding

Watts et al. (2006) and Chapman et al. (2007) recently reported facultative parthenogenesis in Komodo dragons (Varanus komodoensis) and hammerhead sharks (Sphyrna tiburo), respectively. In both cases, the parthenogenic events resulted in progeny homozygous for all loci, suggesting automictic parthenogenesis with terminal fusion and no recombination (Lenk et al. 2005). The genetic consequences of this type of parthenogenesis are analogous to intragametophytic selfing in plants (for the population genetics of this system, see Hedrick 1987a,b). For the hammerhead shark, Chapman et al. (2007) concluded that females are XX and males are XY and that both the mother and her parthenogenetic offspring were XX females. On the other hand, in Komodo dragons, females are ZW and males are ZZ (Watts et al. 2006) and the mothers were ZW and all their parthenogenetic progeny were ZZ males. Both reports suggest that facultative parthenogenesis could potentially be adaptive when no male mates are available but that it may also have negative consequences and lower genetic variation in populations of endangered species.

To examine this situation, assume that the frequency of allele A_i is p_i and that frequency of heterozygotes (A₁A₂) in the unmated female mother is 2p₁p₂. Parthenogenetically produced offspring from an A₁A₂ mother are expected to be 1/2 A₁A₁ and 1/2 A₂A₂ (when she is homozygous, her progeny are homozygous and identical to her). Let us first examine the situation in the hammerhead shark where the progeny are all female XX. When there are n independent parthenogenetically produced female offspring from a single female, then their expected heterozygosity (because there are no actual heterozygotes, this is called diversity) is 2p₁p₂(n−1)/n. For example, if there is only one offspring, then the diversity is 0, illustrating the combined effect of both automictic parthenogenesis and sampling. If there are two offspring, then the diversity is p₁p₂, still 1/2 that in the mother's population. Of course, this diversity is over multiply parthenogenetically reproduced homozygous females rather than in a population of breeding individuals. Finally, if there is one parthenogenetically produced female and she eventually mates with a random male, the expected heterozygosity in their offspring is p₁p₂, reflecting the homozygosity in the female offspring.

Now, let us examine the situation in Komodo dragons where the parthenogenetically produced progeny are all ZZ males. Of course, these sons cannot reproduce parthenogenically but they can potentially mate with their mother and subsequently establish a breeding population. If the female is heterozygous and a son is A₁A₁, then their progeny are 1/2 A₁A₁ and 1/2 A₁A₂, and p₁=0.75. As a result, the diversity is reduced from 2p₁p₂ in the mother to 3p₁p₂/4 in the progeny of the mating between the mother and her son, a 62.5% reduction. If she mates with multiple sons, then the diversity in their progeny is expected to approach p₁p₂, still a 50% loss.

This type of parthenogenesis can also result in purging of detrimental genetic variation, a potentially positive outcome (Hedrick 1994). For example, assume that genotype A₂A₂ is lethal. In this case, the only viable offspring from an A₁A₂ female are A₁A₁. Therefore, the frequency of the A₂ allele is reduced from p₂ in the population to 0 in her progeny, and the genetic load from lethals is eliminated. To examine purging in Komodo dragons, again assume A₂A₂ is lethal so an A₁A₂ female would only have A₁A₁ sons. Therefore, the frequency of the A₂ allele is reduced from 1/2 in the mother to 1/4 in the progeny of the mother and her sons, and the genetic load from lethals is halved.

However, when the female is heterozygous for multiple, detrimental variants, automictic parthenogenesis may result in a low probability of viable offspring and greatly reduce the likelihood of a descendant population. If there are m heterozygous lethals in the female, then the probability that each offspring is viable is only 1/2^m. For example, with m=3, only 12.5% of the progeny would be expected to be viable.

The recently documented parthenogenesis in hammerhead sharks and Komodo dragons may be expected to result in both a substantial loss of genetic variation and in purging of genetic load in descendant individuals. It is not clear how common this type of parthenogenesis is, because even other reptiles vary greatly in types of parthenogenesis (Schuett et al. 1997; Lenk et al. 2005). Further, in fishes, there are a number of different sex-determination mechanisms (Mank et al. 2006) and it is even possible that a ZW system could produce viable, parthenogenetic female WW offspring when there is not a fully developed sex chromosome system. For the Komodo dragon, it appears that a single female and her male progeny could start a population, which would have both lowered genetic variation and genetic load. Unknown parthenogenesis in captive populations could result in unexpectedly high inbreeding, loss of genetic variation and changes in founder contribution.