MethylMix_ModelSingleGene: The MethylMix_ModelSingleGene function

In gevaertlab/MethylMix: MethylMix: Identifying methylation driven cancer genes

Description

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

Usage

MethylMix_ModelSingleGene(GeneName, METdataVector, METdataNormalVector = NULL,
  NoNormalMode = FALSE, maxComp = 3, PvalueThreshold = 0.01,
  MeanDifferenceTreshold = 0.1, minSamplesPerGroup = 1)

Arguments

GeneName	character string with the name of the gene to model
METdataVector	vector with methylation data for cancer samples.
METdataNormalVector	vector with methylation data for normal samples. It can be NULL and then no normal mode will be used.
NoNormalMode	logical, if TRUE no comparison to normal samples is performed. Defaults to FALSE.
maxComp	maximum number of mixture components admitted in the model (3 by default).
PvalueThreshold	threshold to consider results significant.
MeanDifferenceTreshold	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. Defaul is 1 (not used). If -1, each component has to have at least 5% of all cancer samples.

Details

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 MethylMix function, to keep it simple. It would be easy to make them available to the user if we want to.

Value

NrComponents number of components identified.

Models an object with the parameters of the model fitted.

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.

FlipOverState FlipOverState

gevaertlab/MethylMix documentation built on May 13, 2019, 11:53 p.m.