Team 3

Use-cases and power-testing for integrating simulation with popgen analysis

Members:

• Allen Strand (team lead)
• Michelle DePrenger-Levin
• members go here

Goals

• Game plan

1. Gather the information needed for the project and make some decisions what simulators are available in R? ..1. Forward time: rmetasim, MetaPopGen, ..*coalescence: coalsim, sim.coalescent ..*what simulators do we want that are not present in R? ..**forward time: ..**coalescent: ms (already an R wrapper called microsimr), fastsimcoal (already some wrappers in R, sounds like strataG is a good start) What statistic(s) do we want to highlight for power analyses performance analysis Should we make explicit links to group #1 to help decide what statistics might be suitable? What demographies do we want to investigate? Seems like we need to limit the parameter space here. Do we want to put these analyses in the context of an existing paper or project? How do we feel about including ABC? What additional popgen estimators should be included? Eric has written a bunch that might be included directly from strataG. There are also estimators in all sorts of other packages (including adegenet) Tasks Writing document. Depending on choices above start working on vignette describing why simulation approaches are important (well outlined in Hoban et al 2012 and Hoban 2014). Why choose a forward-time simulator? When would a coalscent simulator be a better choice? describe the use cases and the utility of these cases There has got to be more, but I’m tired at the moment Writing code how to represent data umbrella function to set up common parameters. Individual functions for each simulator to set up parameters specific to that software. The idea is that one of the general parameters would be “which type of simulation” and the answer would cause the user to be confronted with a new set of questions possibly use idea to parameterize from existing data to populate some of the fields for the general umbrella function simulation wrapper. A function that runs a simulation X number of times with the parameters defined above analysis wrappers. Functions that take output of simulation and apply an analysis function. Creating reference data sets and parameter choices for simulations to enable validation. Estimate what “tractable” settings are for simulations to allow them to run on a reasonably powerful desktop computer.

Products

Status: Put status here.

• products go here