findCluster {GLAD}R Documentation

Clustering of genomic regions

Description

The function findCluster allows the clustering of regions.

Usage

findCluster.profileChr(profileChr, region="Region", genome=TRUE,
                       lambda=10, nmin=1, nmax=10, type="tricubic",
                       param=c(d=6), verbose=FALSE, ...)

Arguments

profileChr Object of class profileChr.
region Label which indicated where outliers are detected.
genome If TRUE, the "MSHR clustering by chromosome" clustering step is run otherwise it is the "HCSR clustering throughout the genome" clustering step.
lambda Penalty λ*.
nmin Minimal number of cluster to find.
nmax Maximal number of cluster to find.
type Type of kernel to use in the penalty term (this arguments is passed to the kernelpen function).
param Parameters of the kernel (this arguments is passed to the kernelpen function).
verbose If TRUE some information are printed
... Risk alpha used in detectOutliers.

Value

An object of class profileChr with the following element added to the data.frame profileValues:
either:

ZoneChr indicates the regions which belongs to the same HCSR (if genome=FALSE).
ZoneGen indicates HCSR which belongs to the same cluster (if genome=TRUE).

Author(s)

Philippe Hupé, Philippe.Hupe@curie.fr.

See Also

glad, detectOutliers


[Package GLAD version 1.0.1 Index]