Effective Population Size
Current understanding
Effective population size (Ne) is a central parameter in population genetics, summarizing the rate at which genetic drift operates within a population. It is routinely estimated from genomic data via coalescent-based approaches, often using pairwise nucleotide divergence (π) between lineages under the assumption of panmixia. A critical caveat, however, is that Ne estimated this way reflects the entire ancestry of a sample — including shared lineage histories that pre-date the current population boundaries. When populations have been spatially structured in the past, this ancestral identity-by-descent (IBD) can cause Ne estimates to dramatically exceed the actual census size (Nc).
Simulations of stepping-stone models make this inflation concrete. At low migration rates (m = 0.001), end species in a linear array of demes accumulate high pairwise divergence from historical barriers to gene flow; when that divergence is fed into a standard panmictic Ne estimator, the resulting Ne (~7,370) is roughly 7-fold larger than the known census size of Nc = 1,000. Populations near the geographic center of the same array are much less affected, reaching only Ne ≈ 1,255 under the same conditions. This “ghost of structure past” means that range-edge or recently isolated lineages are particularly susceptible to Ne overestimation, with direct consequences for conservation assessments that rely on Ne as a proxy for population viability.
The magnitude of the bias depends on the ratio of divergence time to population size (TD/ND) and on migration rate: the 7-fold inflation specifically applies under TD/ND = 50 and m = 0.001. Other parameter regimes produce different degrees of inflation, but the directional effect — ancestral structure pushing estimated Ne above Nc — is robust across the scenarios examined.
Supporting evidence
- Hancock & Blackmon 2020, Finding 1: Quantifies Ne inflation due to ancestral IBD in a stepping-stone model. At m = 0.001 and TD/ND = 50, end-species pairwise divergence yields Ne ≈ 7,370 versus a true Nc of 1,000, while center-species Ne rises only modestly to ~1,255. The inflation arises because the panmictic estimator (Ne = (π₁₂ − 2TD·μ) / 4μ) cannot distinguish within-deme coalescence from among-deme coalescence accumulated over structured history.
Contradictions / open disagreements
The primary tension is methodological rather than empirical: the Ne inflation documented above is an artifact of applying a panmictic estimator to data generated under a structured model. Coalescent frameworks that explicitly incorporate population structure (e.g., the structured coalescent or isolation-with-migration models) would in principle recover unbiased local Ne values. Whether practitioners routinely apply such corrections — and whether the necessary demographic information is available to do so — remains an open question in applied conservation genetics. The magnitude of overestimation also varies with parameter choices (TD/ND ratio, m), so generalizing the exact 7-fold figure to real systems requires caution.
Tealc’s citation-neighborhood suggestions
- Papers developing or benchmarking structured-coalescent Ne estimators (e.g., MIGRATE-N, ∂a∂i, fastsimcoal2) would contextualize when panmictic approximations break down.
- Empirical conservation-genetics studies comparing Ne estimated from π vs. direct demographic censuses in fragmented populations would test whether the simulated bias manifests in real data.