Self-contained Gene Set Tests for Digital Gene Expression Data
Rotation gene set testing for Negative Binomial generalized linear models.
## S3 method for class 'DGEList'
fry(y, index = NULL, design = NULL, contrast = ncol(design), geneid = NULL,
sort = "directional", ...)
## S3 method for class 'DGEList'
roast(y, index = NULL, design = NULL, contrast = ncol(design), geneid = NULL,
set.statistic = "mean", gene.weights = NULL, nrot = 1999, ...)
## S3 method for class 'DGEList'
mroast(y, index = NULL, design = NULL, contrast = ncol(design), geneid = NULL,
set.statistic = "mean", gene.weights = NULL, nrot = 1999,
adjust.method = "BH", midp = TRUE, sort = "directional", ...)y |
|
index |
index vector specifying which rows (probes) of |
design |
the design matrix. Defaults to |
contrast |
contrast for which the test is required.
Can be an integer specifying a column of |
geneid |
gene identifiers corresponding to the rows of |
set.statistic |
summary set statistic. Possibilities are |
gene.weights |
numeric vector of directional (positive or negative) genewise weights.
For |
nrot |
number of rotations used to compute the p-values. |
adjust.method |
method used to adjust the p-values for multiple testing. See |
midp |
logical, should mid-p-values be used in instead of ordinary p-values when adjusting for multiple testing? |
sort |
character, whether to sort output table by directional p-value ( |
... |
other arguments are currently ignored. |
These functions perform self-contained gene set tests against the null hypothesis that none of the genes in the set are differentially expressed.
fry is the recommended function in the edgeR context.
The roast gene set test was proposed by Wu et al (2010) for microarray data and the roast and mroast methods documented here extend the test to digital gene expression data.
The roast method uses residual space rotations instead of permutations to obtain p-values, a technique that take advantage of the full generality of linear models.
The negative binomial count data is converted to approximate normal deviates by computing mid-p quantile residuals (Dunn and Smyth, 1996; Routledge, 1994) under the null hypothesis that the contrast is zero, and the normal deviates are then passed to the limma roast function.
See roast for more description of the test and for a complete list of possible arguments.
mroast is similar but performs roast tests for multiple of gene sets instead of just one.
The fry method documented here similarly generalizes the fry gene set test for microarray data.
fry is recommended over roast or mroast for count data because, in this context, it is equivalent to mroast but with an infinite number of rotations.
mroast and fry produce a data.frame. See mroast for details.
Yunshun Chen and Gordon Smyth
Dunn, PK, and Smyth, GK (1996). Randomized quantile residuals. J. Comput. Graph. Statist., 5, 236-244. http://www.statsci.org/smyth/pubs/residual.html
Routledge, RD (1994). Practicing safe statistics with the mid-p. Canadian Journal of Statistics 22, 103-110.
Wu, D, Lim, E, Francois Vaillant, F, Asselin-Labat, M-L, Visvader, JE, and Smyth, GK (2010). ROAST: rotation gene set tests for complex microarray experiments. Bioinformatics 26, 2176-2182. http://bioinformatics.oxfordjournals.org/content/26/17/2176
mu <- matrix(10, 100, 4) group <- factor(c(0,0,1,1)) design <- model.matrix(~group) # First set of 10 genes that are genuinely differentially expressed iset1 <- 1:10 mu[iset1,3:4] <- mu[iset1,3:4]+10 # Second set of 10 genes are not DE iset2 <- 11:20 # Generate counts and create a DGEList object y <- matrix(rnbinom(100*4, mu=mu, size=10),100,4) y <- DGEList(counts=y, group=group) # Estimate dispersions y <- estimateDisp(y, design) roast(y, iset1, design, contrast=2) mroast(y, iset1, design, contrast=2) mroast(y, list(set1=iset1, set2=iset2), design, contrast=2)
Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.