Population structure

Nonrandom distribution of genetic variation among biological populations, shaped by ancestry, movement, reproduction and drift.

For example, an allele can be common in one ancestry group and rare in another even without a causal effect on a measured trait. If trait prevalence also differs between groups, a pooled association can reflect ancestry. Modelling structure helps distinguish that source of association from a locus-specific effect. ^[ref]

Genetic differentiation is a difference in genetic composition between biological populations, often measured through allele-frequency differences. It can result from genetic drift, restricted gene flow or different selection pressures. Detecting differentiation alone does not identify which cause dominates. ^[ref]

One descriptive form of FST is $(H_T-H_S)/H_T$, where $H_T$ is expected heterozygosity in the pooled population and $H_S$ the average within-population value. For two equally weighted populations with allele frequencies 0.2 and 0.8, $H_T=0.5$ and $H_S=0.32$, giving 0.36. This constructed example shows how differing frequencies can reduce within-group diversity relative to pooled diversity. Sampling estimators, multiple alleles and genomic comparisons require additional choices; one locus is not a complete measure of biological separation.

Increasing genetic differentiation with geographic separation when gene flow tends to occur over shorter distances.

When most dispersal is local, nearby biological populations exchange more ancestry than distant ones, allowing genetic drift to generate a spatial pattern. The relationship is a population-level tendency rather than a rule that every pair must be ordered exactly by distance. Barriers and recent gene flow can modify it. ^[ref]