gcSim {gcExplorer} | R Documentation |
Functionality to create artificial time course gene cluster data.
gcSim(sim=c("arima","norm","pattern","noise","outlier"), time=10, sd=0.1, sd.ri=0, size=50, n=10, ar=NULL, o=NULL, cent) gcData(...)
sim |
simulation method used |
time |
number of time points |
sd |
standard deviation of the expression profiles |
sd.ri |
standard deviation of the random intercept or gene specific shift |
size |
cluster size, either one value for all clusters or a vector of cluster sizes of length n |
n |
number of clusters |
ar |
any value between -1 and 1 |
o |
the degree of differencing |
cent |
a data matrix giving expression profiles in rows, only used if sim="pattern" or sim="outlier" |
... |
Several "gcSim" objects can be combined using function gcData . |
gcSim
is a unifying function to call different data simulators.
arima
generates expression patterns that come from an integrated AR-process with AR order 1 that can be controlled via ar
and the degree of differencing o
. sim="norm"
and sim="noise"
generate normally distributed expression patterns where sim="noise"
is used to form a noise set of genes.
sim="pattern"
and sim="outlier"
can be used to generate clusters based on a set of cluster centers which are passed to the functions using the argument cent
. sim="outlier"
can be used to test Jackknife distance measures.
gcData
can be used to combine different artificial data generators.
a data matrix
Theresa Scharl
## generate 10 clusters with normally distributed expression patterns: data <- gcSim(sim="norm", time=16, sd=0.1, sd.ri=0.5, size=50, n=10) matplot(t(data),type="l",pch=1) ## combine expression patterns that follow an ARIMA process and a null cluster: data <- gcData(gcSim(sim="arima", time=16, sd=0.1, sd.ri=0.5, size=c(20,50,100,100), n=4), gcSim(sim="noise",time=16, size=100, sd=0)) matplot(t(data),type="l")