Abstract
Motivation: Simulation-based inference can bypass the limitations of statistical methods based on analytical approximations, but software allowing simulation of structured population genetic data without the classical n-coalescent approximations (such as those following from assuming large population size) are scarce or slow.Results: We present GSpace, a simulator for genomic data, based on a generation-by-generation coalescence algorithm taking into account small population size, recombination, and isolation by distance.Availability and implementation: Freely available at site web INRAe (http://www1.montpellier.inra.fr/CBGP/software/gspace/download.html).