Spatial Classification by EM algorithm

Description

The function nem computes spatial classification by EM algorithm.

Usage

## Default S3 method:
nem(LogRatio, Col, Row, nk=nk, beta=1, iters=2000, ...)
## S3 method for class 'arrayCGH'
nem(arrayCGH, variable, nk=5, beta=1, iters=2000, ...)

Arguments

Value

Either a data frame with the following added elements:

or an object of class arrayCGH with the following elements added to the data.frame attribute arrayValues:

Note

People interested in tools for array-CGH analysis can visit our web-page: http://bioinfo.curie.fr.

Author(s)

Philippe Hupé, manor@curie.fr.

References

C. Ambroise, Approche probabiliste en classification automatique et contraintes de voisinage, Ph.D. thesis, Université de Technologie de Compiègne, 1996.

C. Ambroise, M. Dang, and G. Govaert, Clustering of spatial data by the em algorithm in Geostatistics for Environmental Applications, A. Soares, J. Gomez-Hernandez, and R. Froidevaux, Eds., pp. 493-504. Kluwer Academic Publisher, 1997.

P. Neuvial, P. Hupé, I. Brito, S. Liva, E. Manié, C. Brennetot, A. Aurias, F. Radvanyi, and E. Barillot. Spatial normalization of array-CGH data. BMC Bioinformatics, 7(1):264. May 2006.

Examples

data(spatial)  ## arrays with local spatial effects

## Plot of LogRatio measured on the array CGH
## Not run: 
arrayPlot(edge,"LogRatio", main="Log2-Ratio measured on the array
CGH", zlim=c(-1,1), bar="v", mediancenter=TRUE) 

## End(Not run)

## Spatial trend of the scaled log-ratios (the variable "ScaledLogRatio"
## equals to the log-ratio minus the median value of the corresponding chromosome arm) 
edgeTrend <- arrayTrend(edge, variable="ScaledLogRatio",
span=0.03, degree=1, iterations=3, family="symmetric")

## Not run: 
arrayPlot(edgeTrend, variable="Trend", main="Spatial trend of the array CGH", bar="v") 

## End(Not run)

## Classification with spatial constraint of the spatial trend
edgeNem <- nem(edgeTrend, variable="Trend")
## Not run: 
arrayPlot(edgeNem, variable="ZoneNem", main="Spatial zones identified by nem", bar="v")

## End(Not run)

Results

R version 3.3.1 (2016-06-21) -- "Bug in Your Hair"
Copyright (C) 2016 The R Foundation for Statistical Computing
Platform: x86_64-pc-linux-gnu (64-bit)

R is free software and comes with ABSOLUTELY NO WARRANTY.
You are welcome to redistribute it under certain conditions.
Type 'license()' or 'licence()' for distribution details.

R is a collaborative project with many contributors.
Type 'contributors()' for more information and
'citation()' on how to cite R or R packages in publications.

Type 'demo()' for some demos, 'help()' for on-line help, or
'help.start()' for an HTML browser interface to help.
Type 'q()' to quit R.

> library(MANOR)
Loading required package: GLAD

######################################################################################

Have fun with GLAD

For smoothing it is possible to use either
the AWS algorithm (Polzehl and Spokoiny, 2002,
or the HaarSeg algorithm (Ben-Yaacov and Eldar, Bioinformatics,  2008,

If you use the package with AWS, please cite:
Hupe et al. (Bioinformatics, 2004, and Polzehl and Spokoiny (2002,

If you use the package with HaarSeg, please cite:
Hupe et al. (Bioinformatics, 2004, and (Ben-Yaacov and Eldar, Bioinformatics, 2008,

For fast computation it is recommanded to use
the daglad function with smoothfunc=haarseg

######################################################################################

New options are available in daglad: see help for details.


Attaching package: 'MANOR'

The following object is masked from 'package:base':

    norm

> png(filename="/home/ddbj/snapshot/RGM3/R_BC/result/MANOR/nem.Rd_%03d_medium.png", width=480, height=480)
> ### Name: nem
> ### Title: Spatial Classification by EM algorithm
> ### Aliases: nem nem.default nem.arrayCGH
> ### Keywords: spatial models
> 
> ### ** Examples
> 
> data(spatial)  ## arrays with local spatial effects
> 
> ## Plot of LogRatio measured on the array CGH
> ## Not run: 
> ##D arrayPlot(edge,"LogRatio", main="Log2-Ratio measured on the array
> ##D CGH", zlim=c(-1,1), bar="v", mediancenter=TRUE) 
> ## End(Not run)
> 
> ## Spatial trend of the scaled log-ratios (the variable "ScaledLogRatio"
> ## equals to the log-ratio minus the median value of the corresponding chromosome arm) 
> edgeTrend <- arrayTrend(edge, variable="ScaledLogRatio",
+ span=0.03, degree=1, iterations=3, family="symmetric")
> 
> ## Not run: 
> ##D arrayPlot(edgeTrend, variable="Trend", main="Spatial trend of the array CGH", bar="v") 
> ## End(Not run)
> 
> ## Classification with spatial constraint of the spatial trend
> edgeNem <- nem(edgeTrend, variable="Trend")

************************************************
*** Spatial Classification with EM algorithm ***
************************************************


Data :   nb points   =       7392
  grid size =    88 rows,   84 columns

Neighborhood system :
  max neighb =           4
  Default 1st-order neighbors (horizontal and vertical)


NEM parameters :
  beta       =        1.00   |   nk                    =   5

Computing initial partition (sort variable 1) ...

  criterion NEM = 22269.992 / Ps-Like = 7296.387 / Lmix = 9738.891
  NEM converged after 443 iterations

> ## Not run: 
> ##D arrayPlot(edgeNem, variable="ZoneNem", main="Spatial zones identified by nem", bar="v")
> ## End(Not run)
> 
> 
> 
> 
> 
> dev.off()
null device 
          1 
>