Genetics
Genetic Bottleneck
When a population crash leaves the survivors holding a tiny sample of the original gene pool
A genetic bottleneck is a sharp population crash that strips a species of much of its allelic variation, leaving the survivors and their descendants with reduced heterozygosity, elevated inbreeding, and a genome that looks more uniform than the original gene pool. Cheetahs carry roughly 10x less microsatellite diversity than other large felids, with inbreeding coefficients at many loci approaching ~99% identity-by-descent; northern elephant seals were reduced to about 20 individuals in the 1890s and now number ~150,000 from that handful of founders. Even after demographic recovery, the genetic scar persists for thousands of generations because new mutations accumulate at only ~10^-8 per site per generation.
- Heterozygosity loss/gen1/(2Ne)
- Cheetah diversity~10x lower than other felids
- N. elephant seal low~20 in 1890s → ~150,000
- Pingelap achromatopsia5–10% vs 1 in 30,000 world
- Recovery via mutationtens of thousands of generations
- DetectionPSMC, ROH, SFS skew
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
Why genetic bottlenecks matter
- Diversity vanishes faster than headcount. Heterozygosity declines by 1/(2Ne) every generation a population stays small. Twenty surviving northern elephant seals lose about 2.5% of their heterozygosity per generation; ten generations of bottleneck equals ~22% loss before the population can even start to rebound.
- Inbreeding depression hits fitness. Bottlenecked populations show reduced fertility, higher juvenile mortality, and elevated incidence of recessive disease. Florida panthers before the 1995 rescue had ~90% sperm abnormalities, kinked tails in ~80% of males, and atrial septal defects in roughly half of necropsied animals — all symptoms of homozygosity at deleterious loci.
- Conservation policy depends on it. The IUCN's "minimum viable population" guidance — typically Ne ≥ 500 to maintain evolutionary potential and Ne ≥ 50 to avoid imminent inbreeding depression (the "50/500 rule" of Franklin and Soulé, 1980) — is grounded in bottleneck mathematics. Most endangered-species recovery plans target the 50/500 thresholds explicitly.
- Disease vulnerability scales with MHC homozygosity. Cheetahs accept allogeneic skin grafts because MHC variation is so collapsed; the entire species could in principle be wiped out by a single novel pathogen. Tasmanian devils, bottlenecked twice in their recent history, are now being decimated by devil facial tumor disease — a transmissible cancer that the immune system treats as self.
- Adaptive potential is the silent cost. Even if a bottlenecked population recovers numerically, it has lost the rare alleles that natural selection would need to draw on for future environmental change. A species with low standing variation must wait for new mutations to fuel adaptation, which on human timescales is effectively never.
- Genetic load is unmasked. Recessive deleterious alleles that were hidden in heterozygotes get exposed in the homozygotes that bottlenecks produce, sometimes purging the load (as in some highly inbred island populations) but more often causing extinction debt — a fitness deficit that condemns a small population even after it numerically recovers.
- Bottlenecks can drive speciation. Mayr's 1942 founder-effect speciation hypothesis argues that a few founders carried to a new island can rapidly drift to a novel adaptive optimum, reproductively isolating from the source. Hawaiian Drosophila and Galápagos finches show signatures consistent with founder bottlenecks at the start of each radiation.
Common misconceptions
- Recovery means the diversity comes back. Demographic recovery is fast; genetic recovery is glacial. Northern elephant seals are at 150,000 individuals but still have essentially zero mtDNA variation. Mutation rate sets the ceiling: ~1 new variant per gamete per Mb per generation.
- Bottleneck is the same as founder effect. Both reduce Ne to a small number, but a bottleneck is a crash within one continuous population while a founder effect is a colonization event. The source population persists in the founder case, opening the door to gene flow rescue that the bottleneck case lacks.
- Census size tells you whether a species is "safe." Ne, not census N, drives diversity loss. Cheetahs have N ~7,000 but Ne closer to a few thousand because of skewed reproductive success and historical bottlenecks. The IUCN Red List uses both metrics; conservation geneticists weight Ne heavily.
- Inbreeding always reduces fitness. Sometimes it purges deleterious recessives — small island populations of Channel Island foxes appear to have shed much of their genetic load through repeated bottleneck-and-purge cycles. Purging is real but unreliable; you can't bank on it as a conservation strategy.
- Higher mutation rate would fix the problem. Mutation is largely deleterious. A higher rate replenishes diversity but also raises genetic load. Evolution has selected mammalian mutation rates at roughly 10^-8/site/gen as a balance; you cannot crank it up without paying a fitness tax.
- Zoo breeding rebuilds genomes. Captive breeding programs slow further loss but cannot recreate alleles that vanished pre-capture. The California condor program preserves all 14 founder lineages and their ~1.5 GB of distinct genome content; everything outside those 14 is gone.
How a bottleneck reshapes a gene pool
Take a population in Hardy-Weinberg equilibrium with allele frequency p at some locus. When the population crashes to Nb diploid survivors, those survivors carry a sample of 2Nb alleles drawn from the parental pool. Sampling variance is binomial: the frequency in the survivor cohort is approximately Normal with mean p and variance p(1−p)/(2Nb). Rare alleles (small p) are most likely to be missed entirely — the probability of total loss in a sample of 2Nb is (1−p)2Nb, which is roughly 13% for p = 0.05 and 2Nb = 40, dropping to 2% for p = 0.10. After the bottleneck, many low-frequency alleles are simply gone, while the alleles that survived are at frequencies different from where they started — the genetic signature of drift.
If the bottleneck lasts G generations at constant Nb before recovery, expected heterozygosity decays as HG = H0(1 − 1/(2Nb))G. For Nb = 20 and G = 10, that is H10 ≈ 0.776 H0 — a 22% loss before recovery even begins. After the population rebounds, mutation is the only source of new variation; at a per-site rate μ ≈ 10^-8 in mammals, restoring the lost variance takes order 1/μ generations, which for a 1-year generation time is ~100 million years. This is why bottleneck signatures persist as long as the species does.
The detected signature has three classic features. The site frequency spectrum is depleted of rare variants because they were lost; runs of homozygosity (ROH) are longer because all chromosomes trace to few common ancestors; and pairwise sequentially Markovian coalescent (PSMC) curves show a sharp drop in inferred Ne at the bottleneck date and only slow recovery afterward. PSMC reads heterozygous-site density along a single diploid genome and dates each crash by the depth of the trough — this is the method that confirmed the cheetah's Pleistocene bottleneck and pinned the northern elephant seal crash to the 1890s.
Founder effect vs genetic bottleneck
| Property | Genetic bottleneck | Founder effect |
|---|---|---|
| Geometry | Crash within one population | Few migrants found a new daughter population |
| Source population fate | Reduced or extinct | Persists, often unchanged |
| Gene flow rescue possible? | No (no source) | Yes (source still there) |
| Named by | Carlson 1942 / Wright drift theory | Mayr 1942 (Systematics and the Origin of Species) |
| Typical example | Northern elephant seal hunting crash | Hawaiian Drosophila colonization |
| Diversity loss formula | 1 − (1 − 1/(2Nb))G | Identical — same drift math |
| Speciation signature | Within-species genetic scar | Often a step toward new species |
| Conservation framing | Genetic rescue from related populations | Manage daughter as distinct unit |
Famous case studies
- Cheetah (Acinonyx jubatus). Stephen O'Brien's lab discovered ~10x reduced microsatellite heterozygosity and successful reciprocal allogeneic skin grafts in the 1980s, indicating MHC homozygosity. Whole-genome work in the 2010s dated a major bottleneck to ~10,000–12,000 ya, possibly with a more recent secondary crash. Sperm abnormalities run ~70% and juvenile mortality is high. Census ~7,000 wild; Ne in the low thousands.
- Northern elephant seal (Mirounga angustirostris). Hunted to ~20 individuals on Guadalupe Island by the 1890s, recovered to ~150,000 today after Mexican (1922) and US (1972) protections. Microsatellite heterozygosity is roughly 0.40 versus ~0.65 in the never-bottlenecked southern species; mtDNA variation is essentially zero. Classic textbook case for "demographic recovery ≠ genetic recovery."
- Pingelap atoll achromatopsia. Typhoon Lengkieki ~1775 left ~20 survivors on the Pohnpei atoll. One carried a recessive CNGB3 mutation. Today ~5–10% of the ~700 Pingelapese have complete achromatopsia versus ~1 in 30,000 worldwide; ~30% are carriers. Oliver Sacks documented the community in The Island of the Colorblind (1997). A bottleneck that elevated one rare allele to high frequency in the descendant pool.
- Florida panther (Puma concolor coryi). Reduced to ~20–30 adults by the 1990s with severe inbreeding markers — kinked tails, cryptorchidism, atrial septal defects, ~90% sperm abnormalities. Eight female Texas pumas were translocated in 1995; within two generations sperm quality, survival, and tail morphology all improved. Genetic rescue is now textbook conservation practice, with ~230 panthers as of recent counts.
- Tasmanian devil (Sarcophilus harrisii). Two bottlenecks in the species' recent history left it with very low MHC variation. Devil facial tumor disease (DFTD), a transmissible cancer first detected in 1996, spreads because the immune system fails to distinguish foreign tumor cells from self. Population declined ~80% from peak; insurance metapopulations on Maria Island and in captivity preserve diversity for future re-release.
Frequently asked questions
How is a genetic bottleneck different from a founder effect?
Both shrink the gene pool sampled by future generations, but the geometry differs. A bottleneck is a temporary crash within a single continuous population — a hurricane, an epidemic, hunting to near-extinction — followed by recovery from the survivors. The founder effect, named by Ernst Mayr in 1942, occurs when a small group migrates to a new area and starts a daughter population, with the source still intact. Mathematically the two converge: both reduce effective population size Ne to the survivor or founder count, and the expected loss of heterozygosity per generation is 1/(2Ne). The biological difference is that a bottlenecked population has no untouched parental gene pool to draw migrants from, while a founder population usually does, so subsequent gene flow can rescue founders but cannot rescue bottleneck victims.
Why do cheetahs have such low genetic diversity?
Cheetahs (Acinonyx jubatus) carry roughly 10x less microsatellite heterozygosity than other large felids and accept reciprocal skin grafts from unrelated individuals — a sign that MHC variation is so collapsed that immune systems treat strangers as self. The leading hypothesis, from work by Stephen O'Brien starting in the 1980s, is one or more bottlenecks: a Pleistocene crash around the end of the last ice age (~10,000–12,000 years ago) that eliminated most of the species, possibly compounded by a more recent crash in the Holocene. Whole-genome sequencing in the 2010s confirmed an extreme reduction in genome-wide diversity, with current Ne estimated in the low thousands despite a census of ~7,000 wild cheetahs. The result is reduced sperm quality (~70% abnormal), high juvenile mortality, and unusual susceptibility to feline infectious peritonitis.
What happened with northern elephant seals?
Hunted for blubber oil through the 19th century, northern elephant seals (Mirounga angustirostris) were reduced to roughly 20 individuals on Guadalupe Island, Mexico by the 1890s. Mexican and US protections starting in 1922 allowed the population to rebound to about 150,000 today — but every modern individual descends from those ~20 survivors. Microsatellite and mtDNA studies show essentially zero diversity at many loci compared to southern elephant seals, which were never bottlenecked to the same degree. The species is the textbook case that demographic recovery does not equal genetic recovery: the population is large but functionally homogeneous, and a single novel pathogen could in principle sweep through it.
What is the Pingelap achromatopsia case?
Pingelap is a Pacific atoll struck by Typhoon Lengkieki around 1775, which killed most of the population and left roughly 20 survivors. One of them carried a recessive mutation in CNGB3, a cone-photoreceptor gene whose loss causes complete achromatopsia — total color blindness with severe photophobia and reduced visual acuity. Today about 5–10% of the ~700 Pingelapese on the atoll are affected (versus a worldwide rate of about 1 in 30,000), with another ~30% carriers. Oliver Sacks profiled the community in The Island of the Colorblind (1997). The case is technically a founder effect followed by drift, but it illustrates the bottleneck mechanism: a single rare allele rode a small survivor sample to high frequency in the descendant population.
Can a bottlenecked species recover its diversity?
Slowly and only partially. New variation can only come from mutation, which adds about 1.0 × 10^-8 per nucleotide per generation in mammals — meaning a bottlenecked population needs tens of thousands of generations to regenerate the heterozygosity it lost in a single crash. Gene flow from unbottlenecked populations is much faster, but only works if such populations still exist. The Florida panther (Puma concolor coryi) was rescued in 1995 by introducing eight Texas pumas; sperm quality, kink-tail incidence, and survival improved within two generations. The northern elephant seal has no analogous rescue source, since the southern species is reproductively isolated. Conservation genetics now treats genetic rescue — moving a few individuals between populations — as one of the few interventions that work on management timescales.
How do scientists detect a past bottleneck from DNA?
Several signatures stack up. First, overall heterozygosity is lower than expected for the species' census size, since alleles were lost during the crash. Second, the site frequency spectrum shows a deficit of rare alleles — rare variants are the ones most likely to be lost, so a recovered population is enriched for common alleles. Third, runs of homozygosity (ROH) are longer and more frequent, because shared ancestry from few survivors leaves long identical chromosomal stretches. Methods like PSMC (pairwise sequentially Markovian coalescent) and SMC++ infer Ne through deep time from a single genome by analyzing heterozygous-site density along chromosomes; a sharp dip and slow recovery in the inferred Ne curve is the bottleneck fingerprint, and this is how the cheetah, northern elephant seal, and Tasmanian devil crashes were dated.