generateS: generateS

In Caleb-Huo/ClustersSimulation: Clusters Simulation for IS Kmeans

Description

generate S for ISKmeans

Usage

generateS(seed = 15213, S = 2, k = 3, meanSamplesPerK, nModule,
  meanGenesPerModule = 20, Gmean = 9, Gsd = 2, sigma1, sigma2, sigma3, G0,
  nconfounder, nrModule, rMeanSubtypes, diffmu, fold = rep(1, S), rho = 0.5,
  df.prior = 60, groupProb = 1)

Arguments

seed	random seed
S	number of studies
k	number of clusters
meanSamplesPerK	mean samples per cluster
nModule	number of modules. A module is a group of genes.
meanGenesPerModule	number of genes per module
Gmean	gene expression template follows N(Gmean,Gsd^2)
Gsd	gene expression template follows N(Gmean,Gsd^2)
sigma1	noise 1
sigma2	noise 2
sigma3	noise 3
G0	number of noise genes
nconfounder	number of confounders
nrModule	number of modules for confounding variables
rMeanSubtypes	number of subtypes defined by confounding variables
diffmu	effect size difference for subtype predictive genes
fold	how to vary subtype predictive gene signal. 1: original. 0: no signal.
rho	para for inverse Wishart distribution.
df.prior	para for inverse Wishart distribution.
groupProb	subtype predictive genes have prior group information. By prob 1-groupProb, the information will be altered.

Details

generate S for ISKmeans

Value

alist

Author(s)

Caleb

Examples

Sdata =generateS(seed=15213,S=2,k=3,meanSamplesPerK=c(40,40,30),nModule=30,meanGenesPerModule=30,
                 sigma1=1,sigma2=1,sigma3=1,G0=5000,
                  nconfounder=4,nrModule=20,rMeanSubtypes=3,diffmu=1,fold=c(1,1),
				  rho = 0.5,df.prior = 100,groupProb=1)
#
gapStat = gapStatistics(Sdata$d, K = 3,alpha = 0.5, group = Sdata$groupUnion)
iRes <- ISKmeans(Sdata$d,K=3,group=Sdata$groupUnion,gamma=0.15,alpha=0.05)

iRes$Cs
iRes$ws
Sdata$subPredictGeneUnion
table(Sdata$subPredictGeneUnion, iRes$ws!=0)
sum(Sdata$subPredictGeneUnion)
sum(iRes$ws!=0)

Caleb-Huo/ClustersSimulation documentation built on May 6, 2019, 9:27 a.m.