addDiffexp: Calculate differential expression

In mariavica/mircomb: An R package for analyzing miRNA-mRNA interactions

Description

Calculate miRNA or mRNA differential expression from a corObject

Usage

addDiffexp(obj, dataset, classes, method.dif = "t.test", method.adj = "BH",
  var.t.test = FALSE, trend = FALSE, norm = NULL)

Arguments

obj	a corObject
dataset	"miRNA" or "mRNA"
classes	column name of the pheno.miRNA or pheno.mRNA encoding the contrast to make. The column must contain "0" (reference) and "1" (case). Missing values are also allowed. More levels are accepted for "anova" method.
method.dif	method used for differential expression: "t.test", "wilcoxon", "limma", "rankprod", "anova" or "only.fc"
method.adj	multiple testing correction method used (only for method.dif = "t.test", "wilcoxon", "limma" or "anova". One of "holm", "hochberg", "hommel", "bonferroni", "BH", "BY", "fdr", "none" (see \link{p.adjust}).
var.t.test	TRUE or FALSE (default). Only for method.dif = "t.test". If TRUE, considers equality of variances in a T-test.
trend	if TRUE use "limma-trend" method (only for method.dif = "t.test"). Recommended for log-normalised counts)
norm	NULL (default) (in implementation, this feature will help to integrate and normalise NGS data for the correlation)

Value

a corObject with a diffexp.miRNA or diffexp.mRNA slot added, where the rownames of the data.frame are the names of the miRNAs or mRNAs. This slot is a data.frame with the following columns:

• FC: foldchange

• logratio: logratio

• meanExp: mean value log2-expression for the probe across all samples.

• pval: p values. In RankProd the minimum of both p values is reported.

• adj.pval: p values corrected for multiple testing. In the RankProd the pfp (estimated percentage of false positives, which are, in theory, equivalent to False Discovery Rate) are added.

* If method "anova" is selected, then the proportion of SS related to factor is showed instead of FC and logratio. ** If method "only.fc" is selected, then only FC and logratio are computed (useful if there is no enough samples to perform a statistical test).

See Also

corObject-class,package:limma, package:RankProd, voom, t.test, wilcoxon, aov, p.adjust

Examples

data(miRNA)
data(mRNA)
data(pheno.miRNA)
data(pheno.mRNA)

minimal<-new("corObject",dat.miRNA=miRNA,dat.mRNA=mRNA,
	pheno.miRNA=pheno.miRNA,pheno.mRNA=pheno.mRNA)

minimal.diffexp<-addDiffexp(minimal, "miRNA", classes="DvH",
 method.dif="limma")
head([e-mail])

minimal.diffexp<-addDiffexp(minimal.diffexp, "mRNA", classes="DvH",
 method.dif="limma")
head([e-mail])

mariavica/mircomb documentation built on Feb. 3, 2020, 2:28 a.m.