plot.bim.mcmc {bim} | R Documentation |
Plot MCMC burnin and iteration sequences for number of QTL, LOD score, environmental SD, and heritability.
plot.bim.mcmc(x, element = c("burnin","iter"), xlab = c("burnin sequence","mcmc sequence"), items = <<see below>>, ylabs = items, types = <<see below>>, ... )
x |
object of class bim |
element |
elements of x to plot |
xlab |
x-axis labels for elements |
items |
diagnostic item names to be plotted as time series
(default is all names in x$iter |
ylabs |
labels for items |
types |
line types (default is both for first item–usually number of QTL–and lines for the rest) |
... |
graphical parameters can be given as arguments to
plot |
Plots show grey line for saved sequences, usually nearly vertical due to mixing, plus a smoothed curve (blue). Burnin plots should show initial jump in number of QTL during pre-burnin sequence, ideally settling into a well-behaved Markov chain by the end of the burn-in. That is, look for visual evidence of stationarity as flat blue line and even spread of grey. Trend or cyclic pattern indicates Markov chain is not mixing well and needs to be run much longer. No formal assessment is provided.
Brian S. Yandell, yandell@stat.wisc.edu
http://www.stat.wisc.edu/~yandell/qtl/software/Bmapqtl
data( verngeo.bim ) plot.bim.mcmc( verngeo.bim )