addLong: Time analysis

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

Description

Time analysis of a corObject

Usage

addLong(obj, dataset, classes, method.dif = "time.point", method.adj = "BH",
 var.t.test = FALSE)

Arguments

obj	corObject
dataset	"miRNA" or "mRNA"
classes	column name of the pheno.miRNA or pheno.mRNA encoding codification of time variable. The column must contain "0" (t_init) and "1" (t_fin), or the specific times. Missing values are also allowed.
method.dif	"time.point" or "linear.regression"
method.adj	Multiple testing correction method used. One of "holm", "hochberg", "hommel", "bonferroni", "BH", "BY", "fdr", "none" (see p.adjust).
var.t.test	TRUE or FALSE (default). If TRUE, considers equality of variances in a T-test.

Details

If method.dif="time.point", both times are compared using a t-test, time variable should be encoded as kbd"0" (t_init) and "1" (t_fin).

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. Depending on the selected method, the diffexp is a data.frame with the following columns:

If method.dif="time.point":

• FC: foldchange between the two selected times.

• logratio: logratio between the two selected times.

• 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 values (estimated percentage of false positives, which are, in theory, equivalent to False Discovery Rate) are added.

If method.dif="linear.regression":

• slope: slope of the linear regression.

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

• pval: p values associated to the slope of the linear regression.

• adj.pval: p values corrected for multiple testing.

See Also

addDiffexp

Examples

################################
### data special preparation ###
################################

data(miRNA)
data(mRNA)

### simulated phenotypical: 3 samples in 4 time-points, one treatment.

pheno.miRNA<-data.frame(sample=rep(1:3, each=4),
	time=rep(c(1,2,5,9),3),
	time_alt=rep(c(0,NA,NA,1),3))
rownames(pheno.miRNA)<-paste(pheno.miRNA$sample,pheno.miRNA$time,sep="_")

pheno.mRNA<-pheno.miRNA	#same sample distribution

## modify sample names from miRNA and mRNA dataset
colnames(miRNA)<-rownames(pheno.miRNA)
colnames(mRNA)<-rownames(pheno.mRNA)

############################
### start of the example ###
############################

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

#### comparing times 1 (t_init) and 9 (t_fin):
minimal.diffexp<-addLong(minimal, "miRNA", classes="time_alt",
 method.dif="time.point")
head([e-mail])

#### treated as linear regression:
minimal.diffexp<-addLong(minimal, "miRNA", classes="time",
 method.dif="linear.regression")
head([e-mail])

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