compare.two.runs {conStruct}

Compare two conStruct runs

Description

compare.two.runs makes figures comparing the output from two conStruct analyses.

Usage

compare.two.runs(
  conStruct.results1,
  data.block1,
  conStruct.results2,
  data.block2,
  prefix,
  layer.colors = NULL
)

Arguments

conStruct.results1

The list output by a conStruct run.

data.block1

A data.block list saved during a conStruct run.

conStruct.results2

The list output by a second conStruct run.

data.block2

A data.block list saved during a second conStruct run.

prefix

A character vector to be prepended to all figures.

layer.colors

A vector of colors to be used in plotting results for different layers. Users must specify one color per layer. If NULL, plots will use a pre-specified vector of colors.

Details

This function takes the outputs from two conStruct analyses and generates a number of plots for comparing results and diagnosing MCMC performance.

This function produces a variety of plots that can be useful for comparing results from two conStruct analyses. The runs must have the same number of independent MCMC chains, but may have different values of K. The spatial and nonspatial models can be compared. If the runs were executed with different values of K, the run with the smaller value of K should be specified in the first set of arguments (conStruct.results1 and data.block1).

The plots made are by no means an exhaustive, and users are encouraged to make further plots, or customize these plots as they see fit. For each plot, one file is generated for each MCMC chain in each analysis (specified with the n.chains argument in the function conStruct. For clarity, the layers in the second are matched to those in the first using the function match.clusters.x.runs The plots generated (as .pdf files) are:

• Structure plot - STRUCTURE-style plot, where each sample is represented as a stacked bar plot, and the length of the bar plot segments of each color represent that sample's admixture proportion in that layer. Described further in the help page for make.structure.plot.

• Admixture pie plot - A map of samples in which each sample's location is denoted with a pie chart, and the proportion of a pie chart of each color represents that sample's admixture in each layer. Described further in the help page for make.admix.pie.plot

• model.fit.CIs - A plot of the sample allelic covariance shown with the 95% credible interval of the parametric covariance for each entry in the matrix.

• layer.covariances - A plot of the layer-specific covariances overlain unto the sample allelic covariance.

• Trace plots - Plots of parameter values over the MCMC.

  • lpd - A plot of the log posterior probability over the MCMC.

  • nuggets - A plot of estimates of the nugget parameters over the MCMC.

  • gamma - A plot of estimates of the gamma parameter over the MCMC.

  • layer.cov.params - Plots of estimates of the layer-specific parameters over the MCMC.

  • admix.props - A plot of estimates of the admixture proportions over the MCMC.

Value

This function has only invisible return values.