Terminology and units

Terminology

genealogy: the true ancestry of a set of sampled genes at some location in the genome. In a simulation framework the genealogy is known, but in the real world it is an unobservable variable. In the presence of recombination, a single genome represents a mosaic of many ancestors, and consequently, the genealogical history for a set of sampled genomes differs spatially across the genome as different regions share different ancestors. The relationships among nearby (linked) genealogies which share some ancestors, versus unlinked genealogies that do not, is a source of phylogenetic information that can be easily explored in ipcoal.

gene tree: the inferred ancestry of a set of sampled genes at some location in the genome based on a sequence alignment. In practice gene trees rarely match the true genealogy since there is often insufficient information (substitutions) within a small genomic region. Over deeper evolutionary time scales the size of a non-recombined locus becomes very small such that most loci represent substitutions evolved across several linked genealogies.

species tree: a model describing the topology (ancestral relationships) and demographic parameters for a number of lineages. This includes divergence times, effective population sizes, and admixture events.

substitution model: a parameterized Markov model for probabilistically modeling the evolution of sequences along the edges of a genealogical tree. This uses edge lengths described in units of generation times (g), a mutation rate in units of mutations per site per generation (u), and additional optional parameters of the Markov model (e.g., state frequencies, transition-transversion ratio).

loci: a region of length “nsites” spanning some interval of a chromosome. The ancestry for some number of samples may represent multiple distinct genealogies across the length of a locus, if recombination has occurred within it. ipcoal includes functions to infer gene trees for individual loci, or in subwindows across loci (e.g., sliding windows across a locus/chromosome), or to concatenate many loci to infer a concatenated gene tree.

sites: a single position in the genome, at which any individual might have an A, C, G, or T.

Units

genealogical branch lengths: defined in number of generations.

gene tree branch lengths: Depends on the inference method. If using the default ML implementation in raxml then gene tree branch lengths are estimated parameters representing the expected number of substitutions per site.

species tree branch lengths: defined in number of generations.

generation: the length of time from birth until reproduction. Under the Wright-Fisher process, which the coalescent approximates, generations do not overlap.

coalescent units: units of species tree branch lengths that describe the probability that n samples coalesce over a length of a branch. It is calculated as (time in generations) / 2Ne.

mutation rate: the expected number of mutations per base per generation.

recombination rate: the expected number of recombinations per base per generation.

admixture time: a float between 0.0 and 1.0 defining the moment within the overlapping interval of two branches that an admixture event occurred. (For example, if admixture time=0.5, the event happens at the midpoint of the overlap between the two branches)

admixture proportion: the proportion of the source population that migrates to the destination branch.