genesDetails-methods: Details of the genes in the network.

In geNetClassifier: Classify diseases and build associated gene networks using gene expression profiles

Description

Information of the genes in the ranking (table format).

Arguments

object	a GenesRanking
nGenes	integer. Number of genes to show per class
numDecimals	integer. Number of decimals to show in the numeric values
classes	character. Classes of the genes to show
genes	character. Genes to show

Value

A list containing a dataframe with the details of the genes of each class. For each gene, the following information is provided:

ranking	Ranking of the gene.
gERankMean	Mean rank the gene obtained in the cross-validation loops. Only available if geNetClassifier() was called with option estimateGError=TRUE (False by default).
class	Class the gene was chosen for (the class the gene differentiates from the other classes).
postProb	Posterior probability which the gene was assigned by the expectation-maximization algorithm (emfit). Tied values are ranked based on the higher absolute value of exprsMeanDiff. Values are rounded. Several genes may look tied at posterior probability '1' but may actually be i.e. 0.999998 and 0.999997.
exprsMeanDiff	Difference betwen the mean expression of the gene within its class and its mean expression in the other classes.
exprsUpDw	Gene repressed (DOWN) or over-expressed(UP) for the current class (compared to the other classes).
discriminantPower	Measure calculated based on the coordinates of the support vectors. Represents the weight that the classifier gives to each gene to separate the classes.
discrPwClass	Class for which the Discriminant Power was calculated for.
isRedundant	Does the gene have a high correlation or mutual information with other genes in the list? The threshold to consider a gene redundant can be set through the arguments (by default: correlationsThreshold=0.8 and interactionsThreshold=0.5).

Methods

genesDetails(object, nGenes=NULL, numDecimals=4, classes=NULL, genes=NULL)

See Also

Main package function and classifier training: geNetClassifier
This method's class (GenesRanking) help page.

Examples

data(leukemiasClassifier) # Sample geNetClassifier() return
options(width=200) # Optional, use in case the table rows are wrapped

genesDetails(leukemiasClassifier@classificationGenes)$CML
genesDetails(leukemiasClassifier@genesRanking, nGenes=5, numDecimals=2, 
classes="AML")
genesDetails(leukemiasClassifier@genesRanking, genes=c("ENSG00000096006", 
"ENSG00000168081","ENSG00000105699"))$CLL

geNetClassifier documentation built on Nov. 8, 2020, 4:53 p.m.