Population Bottleneck

A population bottleneck (or genetic bottleneck) is a sharp, transient or sustained reduction in the census and effective population size (N_e) of a species or deme caused by environmental upheaval, epizootic disease, habitat destruction, overharvesting, or a colonizing founder effect. Within population genetics, a bottleneck amplifies the binomial sampling variance of gametes (genetic drift), stripping standing heterozygosity, eliminating low-frequency alleles, increasing autozygous identity by descent, and weakening the efficacy of purifying natural selection against mildly deleterious mutations.

Even if a surviving population subsequently rebounds to a large census size (N), the genetic footprint of a bottleneck persists for thousands of generations because lost allelic richness can only be regenerated through slow de novo genetic mutation or incoming gene flow from diverged populations. Consequently, demographic bottlenecks play a central role in conservation genetics, island biogeography, infectious disease transmission bottlenecks, and crop and livestock domestication.

Interactive Simulation: Simulate an acute demographic crash and observe the differential loss of rare alleles versus expected heterozygosity in the Allele Drift Simulator.

Harmonic Mean Effective Population Size Across Fluctuating Generations

In a population whose effective size fluctuates over T discrete generations (N₁, N₂, ..., N_T), the cumulative retention of heterozygosity is multiplicative:

H_T / H₀ = ∏_t=1^T [ 1 − 1 / (2N_t) ]

Equating this multi-generational retention product to an idealized population of constant effective size N_e obeying [1 − 1 / (2N_e)]^T and applying the first-order Taylor approximation ln(1 − x) ≈ −x for 1 / (2N_t) ≪ 1 demonstrates that long-term effective population size is governed by the harmonic mean rather than the arithmetic mean of single-generation sizes:

1 / N_e ≈ (1 / T) · ∑_t=1^T (1 / N_t) ⇒ N_e = T / [ ∑_t=1^T (1 / N_t) ]

Because the reciprocal 1 / N_t is large whenever N_t is small, the harmonic mean is heavily weighted toward the minimum bottleneck generations. For example, if a population maintains N_t = 10,000 for 99 generations but crashes to N₁₀₀ = 20 for a single generation (T = 100), its arithmetic mean size is 9,900.2, yet its harmonic mean effective population size is N_e = 100 / (99/10,000 + 1/20) = 100 / 0.0599 ≈ 1,669. A single generation at N = 20 reduces heterozygosity by 1 / (2 × 20) = 2.5%, which equals the cumulative heterozygosity lost across 500 generations at N = 10,000.

Differential Loss of Allelic Richness vs Expected Heterozygosity

In a classic 1975 theoretical study, Masatoshi Nei, Takeo Maruyama, and Ranajit Chakraborty proved that a population bottleneck impacts allelic richness (the total number of distinct alleles A per locus) far more rapidly than it reduces expected heterozygosity (H_e = 1 − ∑ p_i²). If a locus initially carries k alleles at frequencies p₁, p₂, ..., p_k before a single-generation bottleneck of effective size N_b (representing 2N_b sampled chromosomes), the expected number of alleles surviving the bottleneck (E[A]) is:

E[A] = k − ∑_i=1^k (1 − p_i)^2N_b

Rare alleles (p_i < 0.05) contribute very little to initial expected heterozygosity because their squared frequencies p_i² are negligible; therefore, their loss during the bottleneck barely alters H_e = 1 − ∑ p_i². However, the probability that a rare allele with p_i = 0.02 is lost in a single-generation bottleneck of N_b = 10 individuals (2N_b = 20 gametes) is (1 − 0.02)²⁰ ≈ 0.668 (a 66.8% chance of immediate extinction), whereas expected heterozygosity drops by only 1 / (2N_b) = 5.0%.

Cornuet-Luikart Heterozygosity Excess Test

Because rare alleles are purged much faster than expected heterozygosity (H_e) decays during a demographic contraction, a recently bottlenecked population exhibits a characteristic transient disequilibrium between its observed number of alleles (k_obs) and its gene diversity (H_e). Specifically, the surviving population has a higher Hardy-Weinberg expected heterozygosity (H_e) than would be predicted for an equilibrium population at mutation-drift balance (neutral theory of molecular evolution) carrying the same number of alleles k_obs (H_eq).

Developed by Jean-Marie Cornuet and Gordon Luikart (1996), the heterozygosity excess test evaluates whether H_e > H_eq across a panel of unlinked polymorphic loci (such as microsatellites or SNPs) using a Wilcoxon signed-rank test. Note that this signature refers to an excess of H_e relative to mutation-drift equilibrium H_eq given k_obs, not an excess of observed heterozygotes H_o over H_e. This transient H_e > H_eq window persists for roughly 0.2N_e to 4N_e generations after the bottleneck onset, providing a powerful forensic tool for conservation monitoring.

Tajima's D, Site Frequency Spectrum Shifts, and Linkage Disequilibrium

At the whole-genome sequence level, bottlenecks distort the site frequency spectrum (SFS)—the distribution of derived allele frequencies across polymorphic sites—and alter Tajima's D statistic (Fumio Tajima, 1989):

D = (π − θ_W) / √[ Var(π − θ_W) ]

Here, π is mean pairwise nucleotide diversity (weighted toward intermediate-frequency alleles) and Watterson's estimator θ_W = S / (∑_i=1ⁿ⁻¹ 1/i) estimates 4N_eμ from the total number of segregating sites S (where rare and common variants count equally).

1. Recent / Ongoing Bottleneck (Tajima's D > 0): As rare alleles are selectively eliminated by sampling drift while intermediate-frequency alleles survive, the number of segregating sites S (and thus θ_W) drops sharply while π declines slowly. Consequently, π > θ_W and Tajima's D becomes strongly positive across the entire genome (unlike balancing selection, which produces positive Tajima's D only at a specific selected locus).
2. Post-Bottleneck Population Expansion (Tajima's D < 0): Once the population rebounds to a large size, newly arising mutations accumulate as low-frequency singletons and doubletons, inflating S and θ_W much faster than π recovers and driving Tajima's D negative.
3. Extended Linkage Disequilibrium: Because few ancestral chromosomes pass through the demographic constriction, fewer historical genetic recombination events are captured in the genealogy, generating long blocks of elevated genetic linkage disequilibrium (LD) and extensive runs of homozygosity (ROH) across diploid chromosomes.

Inbreeding Depression, Drift Load, and the Extinction Vortex

During and immediately following a bottleneck, elevated biparental inbreeding (ΔF = 1 / [2N_e] per generation) increases genome-wide homozygosity, exposing recessive deleterious alleles and causing inbreeding depression—manifested as elevated juvenile mortality, congenital malformations, cryptorchidism, and reduced sperm viability. Concurrently, because the threshold for effective purifying selection is |s| > 1 / (2N_e), mildly deleterious mutations with selection coefficients 10⁻⁴ < |s| < 1 / (2N_b) behave as effectively neutral during the bottleneck and can drift to high frequency or fixation, converting segregating masking load into permanent drift load. In severe cases, the synergistic interaction of demographic stochasticity, inbreeding depression, and loss of adaptive diversity at immune loci (such as the Major Histocompatibility Complex, MHC) traps small populations in an extinction vortex.

Empirical Examples in Natural and Human Populations

• Northern Elephant Seal (Mirounga angustirostris): Commercial hunting for blubber oil in the 19th century reduced the northern elephant seal to a single surviving colony of ~20–30 individuals on Guadalupe Island by 1892. Protected in the 20th century, the species rebounded to >150,000 individuals. However, allozymic, mitochondrial, and whole-genome sequencing (Hoelzel et al., 1993; Abascal et al., 2016) revealed that only 2 mitochondrial control-region haplotypes survive (compared to >23 in the unbottlenecked southern elephant seal), coupled with long runs of homozygosity and elevated fixed deleterious missense mutations.
• African Cheetah (Acinonyx jubatus): Genomic analyses show that cheetahs experienced two Pleistocene bottlenecks (~100,000 years ago during intercontinental range expansion and ~10,000–12,000 years ago during the Late Pleistocene megafaunal extinction), depleting genome-wide nucleotide diversity by 90–99% relative to other felids and reducing MHC Class I and II diversity to the point where unrelated individuals accept reciprocal skin allografts without acute immune rejection (O'Brien et al., 1985).
• Pingelapese Achromatopsia (Typhoon Lengkieki Bottleneck): Around 1775, Typhoon Lengkieki devastated Pingelap Atoll in Micronesia, leaving approximately 20 survivors. One surviving chief (Nanmwarki Mwanenised) was a heterozygous carrier for a recessive mutation in the CNGB3 photoreceptor cyclic nucleotide-gated channel gene. Due to the severe bottleneck and subsequent endogamy, complete rod monochromacy (achromatopsia, total color blindness and photophobia) rose from a global prevalence of ~1 in 30,000 to affect ~5–10% of the contemporary Pingelapese population, with ~30% carrying the allele as heterozygotes.

Why does allelic richness recover much more slowly than heterozygosity after a population rebounds?

Actually, when measured toward their respective new post-expansion mutation-drift equilibria, expected heterozygosity (H_e) recovers more slowly in absolute time (~1/[2μ] or ~4N_e generations) than the count of rare alleles, because new mutations immediately create new singleton alleles (increasing allelic richness A and Watterson's θ_W), whereas new mutations must drift to intermediate frequencies before they contribute substantially to H_e = 1 − ∑ p_i². However, during the bottleneck itself, allelic richness drops far more drastically than H_e, so a rebounding population that has not yet accumulated new mutations remains severely depleted in total allele count (such as MHC alleles or self-incompatibility alleles) even if H_e was only modestly reduced.

How can researchers distinguish a genome-wide population bottleneck from a selective sweep?

A selective sweep occurs when a strongly beneficial mutation rises to fixation, dragging linked neutral variants along via genetic hitchhiking. Both a selective sweep and a bottleneck reduce local nucleotide diversity (π), elevate linkage disequilibrium, and skew the site frequency spectrum. The diagnostic distinction is genomic scale: a demographic bottleneck affects every chromosome and locus across the entire genome simultaneously through shared demographic history, whereas a selective sweep produces a localized valley of reduced diversity and altered Tajima's D strictly within the recombination neighborhood surrounding the selected gene.

Do severe bottlenecks ever purge deleterious genetic load?

A bottleneck can partially purge highly deleterious or lethal recessive alleles (where selection coefficient s > 0.5 and dominance h ≈ 0.02) if inbreeding exposes them in homozygotes slowly enough for purifying selection to eliminate carriers before the entire population crashes. However, empirical and theoretical studies (Kirkpatrick & Jarne, 2000) demonstrate that acute, rapid bottlenecks (e.g., N_b < 20) are ineffective at purging mildly deleterious mutations (|s| < 0.05) because genetic drift overpowers purifying selection (|2N_bs| < 1), allowing mildly deleterious variants to increase in frequency and elevate fixed drift load.

1. Nei, M., Maruyama, T., & Chakraborty, R. (1975). "The bottleneck effect and genetic variability in populations." Evolution, 29(1), 1–10. doi:10.1111/j.1558-5646.1975.tb00807.x
2. Cornuet, J. M., & Luikart, G. (1996). "Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data." Genetics, 144(4), 2001–2014. doi:10.1093/genetics/144.4.2001
3. O'Brien, S. J., Roelke, M. E., Marker, L., Newman, A., Winkler, C. A., Meltzer, D., Colly, L., Evermann, J. F., Bush, M., & Wildt, D. E. (1985). "Genetic basis for species vulnerability in the cheetah." Science, 227(4693), 1428–1434. doi:10.1126/science.2983425
4. Hoelzel, A. R., Halley, J., O'Brien, S. J., Campagna, C., Arnbom, T., Le Boeuf, B., Ralls, K., & Dover, G. A. (1993). "Elephant seal genetic variation and the use of simulation models to investigate historical population bottlenecks." Journal of Heredity, 84(6), 443–449. doi:10.1093/oxfordjournals.jhered.a111370