selectWitnessGenes {MineICA} | R Documentation |
This function selects a gene per component.
selectWitnessGenes(icaSet, params, level = c("genes", "features"), maxNbOcc = 1, selectionByComp = NULL)
icaSet |
An object of class |
params |
An object of class
|
level |
The attribute of |
maxNbOcc |
The maximum number of components where
the genes can have an absolute projection value higher
than |
selectionByComp |
The list of components already restricted to the contributing genes |
Selects as feature/gene witness, for each component, the
first gene whose absolute projection is greater than a
given threshold in at the most maxNbOcc
components. These witnesses can then be used as
representatives of the expression behavior of the
contributing genes of the components.
When a feature/gene respecting the given constraints is
not found, maxNbOcc
is incremented of one until a
gene is found.
This function returns a vector of IDs.
Anne Biton
## load an example of IcaSet data(icaSetCarbayo) ## define parameters: features or genes are considered to be contributor # when their absolute projection value exceeds a threshold of 4. params <- buildMineICAParams(resPath="carbayo/", selCutoff=4) ## selection, as gene witnesses, of the genes whose absolute projection is greater than 4 # in at the most one component. I.e, a gene is selected as a gene witness of a component # if he has a large projection on this component only. selectWitnessGenes(icaSet=icaSetCarbayo, params=params, level="genes", maxNbOcc=1) ## selection, as gene witnesses, of the genes whose absolute projection is greater than 4 # in at the most two components. # I.e, a gene is selected as a gene witness of a given component if he has a large projection # in this component and at the most another. selectWitnessGenes(icaSet=icaSetCarbayo, params=params, level="genes", maxNbOcc=2)