Royal Society Publishing

Genealogies and Geography

N. H. Barton , I. Wilson


Any sample of genes traces back to a single common ancestor. Each gene also has other properties: its sequence, its geographic location and the phenotype and fitness of the organism that carries it. With sexual reproduction, different genes have different gencalogies, which gives us much more information, but also greatly complicates population genetic analysis. We review the close relation between the distribution of genealogies and the classic theory of identity by descent in spatially structured populations, and develop a simple diffusion approximation to the distribution of coalescence times in a homogeneous two-dimensional habitat. This shows that when neighbourhood size is large (as in most populations) only a small fraction of pairs of genes are closely related, and only this fraction gives information about current rates of gene flow. The increase of spatial dispersion with lineage age is thus a poor estimator of gene flow. The bulk of the genealogy depends on the long-term history of the population; we discuss ways of inferring this history from the concordance between genealogies across loci.