runQgen: Run Q-Gen (generalized QuSAGE) algorithm using gene level...

Description Usage Arguments Details Value Examples

View source: R/geneSetTests.R

Description

Run Q-Gen (generalized QuSAGE) algorithm using gene level statistics

Usage

1
runQgen(model.results, gene.sets, annotations = NULL)

Arguments

model.results

object returned by genModelResults.

gene.sets

list of gene sets. See genModelResults for more formatting details.

annotations

A data frame of additional annotations for the gene sets. See genModelResults for more formatting details.

Details

This function takes the gene level comparison estimates and test statistics contained in the object returned by genModelResults and runs the Q-Gen algorithm across all of the comparisons. The VIFs are estimated using the raw residuals, which are also contained in the output of genModelResults.

Value

qusage.results Tall formatted matrix of results

lower.ci Matrix of gene level lower 95% confidence intervals

upper.ci Matrix gene level upper 95% confidence intervals

gene.sets List of gene sets provided to gene.sets

annotations data frame of gene set annotations. Default is NULL.

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
# Example data
data(tb.expr)
data(tb.design)

# Use first 100 probes to demonstrate
dat <- tb.expr[1:100,]

# Create desInfo object
meta.data <- metaData(y = dat, design = tb.design, data.type = "microarray", 
                    columnname = "columnname", long = TRUE, subject.id = "monkey_id",
                    baseline.var = "timepoint", baseline.val = 0, time.var = "timepoint", 
                    sample.id = "sample_id")

# Generate lmFit and eBayes (limma) objects needed for genModelResults
tb.design$Group <- paste(tb.design$clinical_status,tb.design$timepoint, sep = "")
grp <- factor(tb.design$Group)
design2 <- model.matrix(~0+grp)
colnames(design2) <- levels(grp)
dupcor <- limma::duplicateCorrelation(dat, design2, block = tb.design$monkey_id)
fit <- limma::lmFit(dat, design2, block = tb.design$monkey_id, 
                    correlation = dupcor$consensus.correlation)
contrasts <- limma::makeContrasts(A_20vsPre = Active20-Active0, A_42vsPre = Active42-Active0, 
                                  levels=design2)
fit2 <- limma::contrasts.fit(fit, contrasts)
fit2 <- limma::eBayes(fit2, trend = FALSE)

# Create model results object for runQgen
model.results <- genModelResults(y = dat, data.type = "microarray", object = fit2, lm.Fit = fit, 
                                 method = "limma")
                               
# Run Q-Gen on baylor modules                             
data(modules)
qus.results <- runQgen(model.results, modules)

genBart documentation built on May 2, 2019, 9:13 a.m.