mcmc.converge.check: MCMC diagnostic plots

In MCMC.qpcr: Bayesian Analysis of qRT-PCR Data

Description

A wrapper function for plot.MCMCglmm to plot diagnostic convergence plots for selected fixed effects

Usage

mcmc.converge.check(model, factors, ...)

Arguments

model	output of mcmc.qpcr (or any MCMCglmm class object)
factors	A vector of names of fixed effects of interest; see details in HPDplot help page.
...	other options to pass to plot.MCMCglmm

Value

A series of plots for each gene-specific fixed effect.

The MCMC trace plot is on the left, to see if there is convergence (lack of systematic trend) and no autocorrelation (no low-frequency waves). If lack of convergence is suspected, try increasing number of iterations and burnin by specifying, for example, nitt=50000, burnin=5000, as additional options for mcmc.qpcr. If autocorrelation is present, increase thinning interval by specifying thin=20 in mcmc.qpcr (you might wish to increase the number of iterations, nitt, to keep the size of MCMC sample the same)

The right plot is posterior density distribution.

Author(s)

Mikhail V. Matz, UT Austin

References

Matz MV, Wright RM, Scott JG (2013) No Control Genes Required: Bayesian Analysis of qRT-PCR Data. PLoS ONE 8(8): e71448. doi:10.1371/journal.pone.0071448

MCMC.qpcr documentation built on March 31, 2020, 5:22 p.m.