ModelSingleGene_Biv: The ModelSingleGene_Biv function
In mpru/BIMEGA: BIMEGA: BIvariate Methylation and Expression GAussian mixture model

Description Usage Arguments Details Value

Internal. For a given gene, this function fits the mixture model, selects the number of components and defines the respective methylation states.

ModelSingleGene_Biv(GeneName, METdataVector, MAdataVector, METdataNormalVector,
  MAdataNormalVector = NULL, NoNormalMode = FALSE, test = c("ttest",
  "wilcoxon"), maxComp = 3, PvalueThreshold = 0.01,
  METDiffThreshold = 0.1, minSamplesPerGroup = -1)

`GeneName`	character string with the name of the gene to model
`METdataVector`	vector with methylation data for cancer samples.
`MAdataVector`	vector with gene expression data for cancer samples
`METdataNormalVector`	vector with methylation data for normal samples.
`MAdataNormalVector`	vector with gene expression data for normal samples (optional).
`NoNormalMode`	logical, if TRUE no comparison to normal samples is performed. Defaults to FALSE.
`test`	how to do the comparison between cancer samples in one state and the normal samples. Default is "ttest", the other option is "wilcoxon".
`maxComp`	maximum number of mixture components admitted in the model (3 by default).
`PvalueThreshold`	threshold to consider results significant.
`METDiffThreshold`	threshold in beta value scale from which two methylation means are considered different.
`minSamplesPerGroup`	minimum number of samples required to belong to a new mixture component in order to accept it. Defaults to 5% of all cancer samples.

test, maxComp, PvalueThreshold, METDiffThreshold, minSamplesPerGroup are arguments for this function but are fixed in their default values for the user because they are not available in the main BIMEGA function, to keep it simple. It would be easy to make them available to the user if we want to.

NrComponents number of components identified.

Models an object of class 'Mclust' with the output model

MethylationStates vector with DM values for each sample.

MixtureStates vector with DMvalues for each component.

Classifications a vector indicating to which component each sample was assigned.

mpru/BIMEGA documentation built on May 23, 2019, 6:34 a.m.