Testing for differentially expressed genes

Share:

Description

This function can be used to test for genes that are differentially expressed between levels of an ordinal factor, such as dose levels or ordinal phenotypes.

Usage

1
2
ordGene(xpr, lvs, type = c("RLRT", "LRT"), nsim = 1e6,
null.sample=NULL, progressBar = TRUE, ...)

Arguments

xpr

a matrix or data frame of gene expression data with Probe IDs as row names.

lvs

a numeric vector containing the factor levels (e.g., dose levels) corresponding to the columns of xpr.

type

the type of test to carry out: likelihood ratio ("LRT") or restricted likelihood ratio ("RLRT").

nsim

number of values to simulate from the null distribution.

null.sample

a vector containing values already simulated from the null distribution (overrides nsim)

progressBar

enable or disable the progress bar; default is TRUE, set it as FALSE if problems with the tcltk package occur.

...

additional arguments to LRTSim and RLRTSim, respectively.

Details

For each gene in the dataset, ordAOV is applied to test for differences between levels given in lvs. See ordAOV for further information on the testing procedure. Simulation studies by Gertheiss (2014) suggest that a restricted likelihood test (RLRT) should rather be used than a likelihood ratio test (LRT).

In addition to (R)LRT, results of usual one-way ANOVA (not taking the factor's ordinal scale level into account) and a t-test assuming a linear trend across factor levels are reported. Note that the t-test does not assume linearity in the doses (such as 0, 0.5, 2.0, 5.0, ...), if given, but in the levels, i.e., 1, 2, 3, etc.

Value

A matrix containig the raw p-values for each gene (rows) when using (R)LRT, ANOVA or a t-test (columns).

Author(s)

Jan Gertheiss

References

Crainiceanu, C. and D. Ruppert (2004). Likelihood ratio tests in linear mixed models with one variance component, Journal of the Royal Statistical Society B, 66, 165-185.

Gertheiss, J. (2014). ANOVA for factors with ordered levels, Journal of Agricultural, Biological and Environmental Statistics, 19, 258-277.

Gertheiss, J. and F. Oehrlein (2011). Testing relevance and linearity of ordinal predictors, Electronic Journal of Statistics, 5, 1935-1959.

Sweeney, E., C. Crainiceanu, and J. Gertheiss (2015). Testing differentially expressed genes in dose-response studies and with ordinal phenotypes, Preprint (available on request).

See Also

ordAOV

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
## Not run: 
# use dopamine data from package IsoGene
require(IsoGene)
require(Biobase)
data(dopamine)
xpr <- data.frame(exprs(dopamine))
dose <- unlist(pData(dopamine))
plot(dose,xpr[83,], col=as.factor(dose), lwd=2, ylab="expression")

# calculate p-values
pvals <- ordGene(xpr = xpr, lvs = dose, nsim=1e6)

# compare distribution of (small) p-values
plot(ecdf(pvals[,1]), xlim=c(0,0.05), ylim=c(0,0.12),
main="dopamine", xlab="p-value", ylab="F(p-value)")
plot(ecdf(pvals[,2]), xlim=c(0,0.05), add=TRUE, col=2)
plot(ecdf(pvals[,3]), xlim=c(0,0.05), add=TRUE, col=3)
legend('topleft', colnames(pvals), col=1:3, lwd=2, lty=1)

## End(Not run)