Population Size Does Not Influence Mitochondrial Genetic Diversity in Animals

See allHide authors and affiliations

Science  28 Apr 2006:
Vol. 312, Issue 5773, pp. 570-572
DOI: 10.1126/science.1122033


Within-species genetic diversity is thought to reflect population size, history, ecology, and ability to adapt. Using a comprehensive collection of polymorphism data sets covering ∼3000 animal species, we show that the widely used mitochondrial DNA (mtDNA) marker does not reflect species abundance or ecology: mtDNA diversity is not higher in invertebrates than in vertebrates, in marine than in terrestrial species, or in small than in large organisms. Nuclear loci, in contrast, fit these intuitive expectations. The unexpected mitochondrial diversity distribution is explained by recurrent adaptive evolution, challenging the neutral theory of molecular evolution and questioning the relevance of mtDNA in biodiversity and conservation studies.

Genetic diversity is a central concept of evolutionary biology that has been linked to organismal complexity (1), ecosystem recovery (2), and species ability to respond to environmental changes (3). A lack of diversity is typically considered as evidence for a small or declining, potentially endangered population (4, 5). Population genetics theory tells us that, for a neutral locus, the expected polymorphism at mutation-drift equilibrium is proportional to the effective population size, the equivalent number of breeders in an ideal, panmictic population. Other factors can of course affect the genetic polymorphism, including population structure (6), population bottlenecks (3), and natural selection [either directly or through genetic linkage (7, 8)], life cycle (9), and mating systems (10). These multiple influences complicate any attempt to interpret the genetic diversity of one particular species in terms of population size (11). Population size, however, presumably varies by several orders of magnitude between species and taxa, so that one would typically predict that abundant species should be, on average, more polymorphic than scarce ones despite the noise introduced by other evolutionary forces.

Meta-analyses of allozyme polymorphism studies were mostly consistent with this theoretical prediction (12, 13). In particular, invertebrate animals were found to be more polymorphic, on average, than vertebrates (13). It was noted, however, that the expected proportional relationship between diversity and effective population size was rarely met (14). DNA-based markers have now replaced allozymes in population genetics studies. Among these, the supposedly nonrecombining and evolutionary nearly neutral mitochondrial DNA (mtDNA) has been the most widely used marker of population history and diversity (15, 16), the general belief being that mtDNA diversity should reflect effective population size more accurately than allozymes (17). In this study, we approach the taxonomic and ecological determinants of effective population size by analyzing the distribution of the genetic polymorphism across animal taxa, focusing on mtDNA and comparing it to allozymes and nuclear DNA data.

Three exhaustive within-species polymorphism data sets were used: an allozyme data set (912 species) taken from the compilation by Nevo et al. (12), a nuclear sequence data set (417 species), and a mitochondrial sequence data set (1683 species), the latter two both built from the Polymorphix database (18, 19). We first calculated the average genetic diversity in eight largely represented animal taxa (hereafter called “groups”). The allozyme and nuclear data sets yielded highly similar results (Fig. 1): The average within-species diversity in all four invertebrate groups was higher than that of vertebrates, mollusks being the most diverse and mammals the least diverse, on average. This is essentially in agreement with our intuition about species abundance in these taxa. The mtDNA data diversity, however, was highly variable between species within a group, but remarkably homogeneous between groups (Fig. 1). Insect or mollusk species did not appear more polymorphic, on average, than mammals or birds, contradicting our prior beliefs about relative population sizes in these taxa. The average invertebrate mtDNA diversity (7.67%) was not appreciably different from the vertebrate one (7.99%), whereas the nuclear invertebrate average (2.46%) was four times as high as the vertebrate one (0.60%).

Fig. 1.

Average allozymic, nuclear DNA, and mtDNA diversity in eight animal taxa. x axis: allozyme average heterozygosity. y axis: circles, nuclear DNA average synonymous diversity (kendall test: τ = 0.87, P < 0.05); squares, mtDNA average synonymous diversity (kendall test: τ = –0.14, not significant). Ma: Mammalia (allozymes: 184 species; nuclear: 30 species; mtDNA: 350 species); S: Sauropsida (reptiles and birds: 116, 20, 378); A: Amphibia (61, 4, 96); P: Pisces (bony fish and cartilaginous fish: 183, 22, 270); I: Insecta (156, 73, 511); C: Crustacea (122, 2, 78); E: Echinodermata (sea stars and urchins: 15, 14, 47); and Mo: Mollusca (46, 9, 125). The nuclear averages of the little-represented Amphibia (four species) and Crustacea (two species) are shown but were not used for the statistical test.

A series of within-group analyses were conducted to examine the influence of specific ecological variables (Table 1). Allozyme data again agreed with our intuition about population sizes: Among mollusks, the terrestrial pulmonates were substantially less polymorphic than marine bivalves or gastropods, consistent with the enormous dispersal potential of the latter; among crustaceans, the microscopic, planktonic branchiopods (e.g., Artemia and Daphnia) appeared much more diverse than the larger decapods (shrimps, lobsters, and crabs); among fish, marine species showed a significantly higher heterozygosity than the geographically restricted freshwater species. The mtDNA diversity, in contrast, failed to reflect these differences in average population size. Again, a homogeneous average nucleotide diversity was found, irrespective of body size and ecology (Table 1). Freshwater fish species were even significantly more polymorphic than marine ones.

Table 1.

Ecological determinism of allozyme and mtDNA genetic diversity. The numbers of species used are shown in parentheses.

TaxonAllozymes (H, %)mtDNA (πs, %)
Fish Freshwater 4.7 (71) 8.7View inline (123)
Marine 6.1View inline (65) 3.7 (51)
Crustaceans Large benthic 4.6 (81) 10.1 (26)
Small planktonic 21.0View inline (8) 5.8 (6)
Mollusks Terrestrial 7.4 (23) 7.8 (8)
Marine 30.0View inline (17) 5.6 (34)
  • View inline* P < 0.05 (Student's t test).

  • View inline** P < 0.01 (Student's t test).

  • Variations in mitochondrial mutation rate among phyla could be invoked to explain the discrepancy between animal mtDNA diversity and effective population size. The mutation rate, however, would have to be inversely related to population size throughout animal taxa to explain the data—a pattern very unlikely to appear by chance. Demographic stochasticity, e.g., recurrent population bottlenecks, could remove the effect of equilibrium population size on genetic diversity (20). Demographic effects, however, should affect the nuclear genome as well, which is not what we observe. Natural selection, either purifying or adaptive, must therefore be invoked to explain the locus-specific behavior of mtDNA.

    Purifying selection against deleterious mutations (so-called background selection) decreases the diversity at linked loci through hitch-hiking. The strength of this effect depends on the distribution of fitness effects among mutations, and one generally still expects an increase of diversity with population size under background selection (21), which is not consistent with the homogeneous mtDNA diversity distribution. Our analytical results confirmed this statement: The conditions under which background selection can lead to a more or less independent relationship between diversity and effective population size appear implausible (fig. S2).

    The mtDNA pattern, however, appears to be in good agreement with the hypothesis of recurrent fixation of advantageous mutations leading to frequent loss of variability at linked loci (7, 22), a process recently named “genetic draft” by Gillespie (23). The population number of advantageous mutations per generation obviously increases with population size and compensates the decrease of genetic drift in Gillespie's (24) simulations, which predict an essentially flat, even negative, relationship between genetic diversity and population size. The gene-dense, nonrecombining context of the animal mitochondrial genome maximizes the potential impact of the genetic draft, as compared with that of the nuclear genome (25).

    To firmly distinguish between the two selective models, we examined the pattern of nucleotide substitution between species. The neutrality index (NI) (26) was first calculated when outgroup sequences were available. This index aims at comparing the ratio of nonsynonymous (amino acid–changing) to synonymous (silent) changes within species (πNS) and between species (dN/dS): NI is 1 when evolution is neutral, greater than 1 under purifying selection, and less than 1 in the case of adaptation. A significant shift toward values less than 1 was detected in invertebrate mtDNA loci, consistent with the adaptive hypothesis (Fig. 2 and fig. S1).

    Fig. 2.

    Neutrality index (NI) distributions (logarithmic scale). Medians are indicated by thick horizontal bars. Boxes include 50% of the distributions. The invertebrate mtDNA median NI (0.42) is significantly lower than the vertebrate one (0.88; P < 10–3, Mann-Whitney test). NI values greater than 20 were forced to 20 for clarity. Low-frequency (<0.125) polymorphic sites were excluded from the analysis.

    This result, limited to the genes for which polymorphism data are available, was confirmed by a whole-genome mitochondrial analysis. The dN/dS ratio was calculated for the 13 mitochondrial protein-coding genes in various animal taxa (Table 2). The average genomic dN/dS was significantly higher in invertebrates than in vertebrates. This is not consistent with a model invoking solely purifying selection, because the rate of fixation of deleterious mutations is expected to decrease with population size. Observing a higher rate of nonsynonymous substitution, but not a higher level of diversity, in large populations strongly corroborates the hypothesis that positive selection drives mitochondrial evolution in animals: Neither negative selection (which should decrease dN/dS and increase NI) nor a relaxation of constraints (which should increase the diversity) can explain this pattern. The additional amino acid substitutions detected in invertebrates would correspond to adaptive changes, plus the deleterious ones hitch-hiking to fixation—the rate of deleterious substitution is expected to increase with population size in the genetic draft model (24).

    Table 2.

    Mitochondrial genomic dN/dS ratio in animals.

    TaxonData setsdN/dS
    Mammalia 21 0.080
    Sauropsida 9 0.121
    Amphibia 12 0.086
    Teleostei 44 0.065
    Chondrichthyes 2 0.077
        Average 0.086
    Insecta 4 0.198
    Crustacea 5 0.084
    Mollusca 2 0.122
    Echinodermata 1 0.106
    Nematoda 2 0.219
    Chelicerata 6 0.138
    Platyhelminthes 1 0.140
    Urochordata 1 0.188
    Cnidaria 2 0.167
        Average 0.151View inline
  • View inline** P < 0.01 (Student's t test).

  • This study reveals that the mitochondrial diversity of a given animal species does not reflect its population size: No correlation between mtDNA polymorphism and species abundance could be detected, despite the large body of data analyzed. Nuclear data, in contrast, are fairly consistent with intuitive expectations. We conclude that natural selection acting on mtDNA contributes to homogenization of the average diversity among groups, in agreement with the genetic draft theory. mtDNA appears to be anything but a neutral marker (16) and probably undergoes frequent adaptive evolution, e.g., direct selection on the respiratory machinery (27), nucleo-cytoplasmic coadaptation (28), two-level selection (29), or adaptive introgression, perhaps hitchhiking with a maternally transmitted parasite (30). mtDNA diversity is essentially unpredictable and will, in many instances, reflect the time since the last event of selective sweep, rather than population history and demography. Low-diversity mitochondrial lineages, typically disregarded as important from a conservation standpoint, might sometimes correspond to recently selected, well-adapted haplotypes to be preserved.

    Supporting Online Material

    Materials and Methods

    Figs. S1 to S5

    References and Notes

    References and Notes

    View Abstract

    Navigate This Article