findCluster {GLAD} | R Documentation |
The function findCluster
allows the clustering of regions.
findCluster.profileChr(profileChr, region="Region", genome=TRUE, lambda=10, nmin=1, nmax=10, type="tricubic", param=c(d=6), verbose=FALSE, ...)
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 . |
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 ). |
Philippe Hupé, Philippe.Hupe@curie.fr.