sim.mcmc | R Documentation |
Run multiple simulations of a backcross, analyzed using several model selection techniques, including MCMC.
sim.mcmc(n.sim=1000, max.steps=28, bic.mult=2.5, n.steps=1000,
n.ind=100, n.mar=rep(11,9), mar.sp=rep(10,10*9),
qtl.chr=c(1,1,2,2,3,4,5), qtl.mar=c(4,8,4,8,6,4,1),
qtl.dist=rep(0,7), qtl.eff=c(1,1,1,-1,1,1,1)/2,
sigma=1, only.two=FALSE)
n.sim |
Number of simulation replicates. |
max.steps |
Maximum number of steps in forward/backward selection. |
bic.mult |
Multiplier for BIC. |
n.steps |
Number of steps in MCMC. |
n.ind |
Number of backcross individuals. |
n.mar |
Vector indicating the number of markers on each chromosome. |
mar.sp |
Vector of length sum(n.mar)-n.chr-1, giving the inter-marker spacings in cM. |
qtl.chr |
Chromosomes on which qtls are sitting. |
qtl.mar |
Markers to the left of each qtl, numbered 1, 2, ... |
qtl.dist |
Distance between qtl and marker to left, in cM. |
qtl.eff |
QTL effects. |
sigma |
Residual (environmental) standard deviation. |
only.two |
If TRUE, only do forward selection and MCMC |
A list of three components. The first is a list of length 5, whose components correspond to the results of forward and backward selection, the same followed by branch-and-bound, with BIC, and MCMC. (Each component of this list is itself a list of length n.sim, with each component giving the chromosome numbers and marker numbers of the identified QTLs.) The second component of the main list gives the BIC values corresponding to the identified models. The last component gives the step at which the final model from MCMC was first observed, plus the number of distinct models observed.
Karl W Broman, broman@wisc.edu
Broman, K. W. (1997) Identifying quantitative trait loci in experimental crosses. PhD dissertation, Department of Statistics, University of California, Berkeley.
simbc
, anal.mcmc
,
anal.leaps
,
anal.multi
## Not run: results <- sim.mcmc(n.sim=10)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.