IsoDataSet-core: GLM fitting and hypothesis testing

In NBSplice: Negative Binomial Models to detect Differential Splicing

Description

buildData is an R function to build the data structure useful to fit a negative binomial model at the gene level. fitModel is an R function to fit a negative binomial generalized linear model in order to evaluate changes in the isoform ratio. Each isoform is considered as a factor. The model incorporates one factor related to one experimental condition with two levels.

Usage

buildData(isoCounts, geneCounts, geneIso, gene, designMatrix, colName)

fitModel(myData, gene, formula, colName, test = c("F", "Chisq"), contrast)

Arguments

isoCounts	Matrix having the expression counts at the isoform level. Isoforms must be in rows and samples in columns. Rownames and colnames must be defined with isoform and samples names, respectively.
geneCounts	Matrix having the expression counts at the gene level. Genes must be in rows and samples in columns. Rownames and colnames must be defined with gene and sample names, respectively.
geneIso	Data.frame containing the relationship between isoforms and genes. It must contain two columns, named as 'gene_id' and 'isoform_id'. Its isoforms should be the same specified in the isoCounts matrix.
gene	Character indicating the name of the gene to be analyzed
designMatrix	Data.frame specifying metadata related to the experiment. Its rows must be the samples and experimental factors should be arranged on its columns.
colName	Character indicating the name of the column in the design matrix to be considered for mean expression calculations per experimental condition and differential expression test.
myData	Data.frame containing the expression matrix at isoform levels, it means, isoform in rows and samples in columns. It is obtained using the buildData NBSplice method.
formula	Object with the formula of the GLM.
test	Character indicating the name of the distribution to assume for linear hypothesis statistic. Could be "F" or "chisq".
contrast	Character vector with the names of the two levels of the experimental factor to be contrasted.

Value

A data.frame ready to use by the model fitting function

A data.frame summarizing gene results.

Note

see full example in IsoDataSet-class

Author(s)

Gabriela A. Merino [e-mail] and Elmer A. Fernandez [e-mail]

Gabriela A. Merino [e-mail] and Elmer A. Fernandez [e-mail] ## Data loading data(myIsoDataSet, package="NBSplice")

## Identificating Low expressed Isoforms myIsoDataSet<-buildLowExpIdx(myIsoDataSet)

##Arguments definition colName<-"condition" test<-"F"

## Differential splicing test myDSResults<-NBTest(myIsoDataSet, colName, test)

See Also

IsoDataSet

Other IsoDataSet: IsoDataSet-class, IsoDataSet, NBTest, buildLowExpIdx, designMatrix, geneIso, initialize, isoCountsData, myIsoDataSet

Examples

## Data loading
data(myIsoDataSet, package="NBSplice")

## Identificating Low expressed Isoforms
myIsoDataSet<-buildLowExpIdx(myIsoDataSet)

##Arguments definition
colName<-"condition"
test<-"F"

## Differential splicing test
myDSResults<-NBTest(myIsoDataSet, colName, test)

NBSplice documentation built on Nov. 8, 2020, 8:07 p.m.