Estimation of parameters and simulation of population data

Accueil
Les personnes
Personnels permanents
SCOTTI Ivan
COLLOQUE ESEB 2015
Estimation of parameters and simulation of population data

Estimation of parameters and simulation of population data

The scenario shown in the figure was used for both parameter estimation and simulations of 2D-SFS's conditional on the most-likely set of demographic parameter values under a coalescent model using Fastsimcoal2.51. Current effective population sizes, population splitting times, timing and size of changes of effective population sizes and migration rates were all estimated from the model and left to vary within reasonable constraints (e.g. migration rates between regions weaker than within site).

SimulationofPopulation

Parameters were estimated as follows:

Current population sizes (left to right in the figure): 477, 436, 54, 408

Migration rates: within Régina, 0.004; within Laussat, 0.014; between sites, 0.0002

Time of population size changes (Size of population size changes, backward in time, in parentheses): for current populations, left to right: 107 (0.16), 1 (0.10), 17 (2.4), 41 (4.2); for ancestors of current populations: Régina, 190 (1.4), Laussat, 744 (2.9); regional pool, 4511 (5.5).

Merger times: within Régina, 173; within Laussat, 502; between sites, 4123.

It has to be noted that Log-Likelihood of max-likelihood simulated data (-54230) under the best estimated parameters was still far from likelihood of experimental data (-41667), so the model is not optimal anyway.

Date de modification : 22 juin 2023 | Date de création : 16 juillet 2015 | Rédaction : Ivan Scotti