Component: CN2GECollection

CN2GECollection

Encapsulates three algorithms with similar inputs for copy number to RNA expression integration. All inputs must be in the same order and have matching IDs.

In pint (R Bioconductor), canonical correlation analysis using pSimCCA to detect integrative effects of CNA and gene expression. Inputs must be matched i.e., dim(exprMatrix) = dim(cnaMatrix) and rows (samples) and columns (genes/transcripts/exons) match.

In DRI (R CRAN), either a standard correlation coefficient between the two data or a supervised learning method is employed to detect similarily directed abnormalities.

In edira, equally directed deviations from reference samples are detected by combined modified correlation coefficient followed by a Wilcoxon test.

Version	1.0
Bundle	microarray
Categories	Integration Copy Number Analysis
Authors	Riku Louhimo (Riku.Louhimo@Helsinki.FI)
Issue tracker	View/Report issues
Requires	edira (R-package) ; mixOmics (R-package) ; pint (R-bioconductor) ; PREDA (R-package) ; SIM (R-bioconductor) ; iCluster (R-bioconductor) ; tilingArray (R-bioconductor)
Source files	component.xml CN2GECollection.r
Usage	Example with default values CN2GECollection( exprMatrix = LogMatrix, cnaMatrix = CSV, exprAnnotation = AnnotationTable, //exprRefMatrix = LogMatrix, //cnaRefMatrix = CSV, //labels = CSV, DRImethod = "pearson", SGRthold = 0.1, algorithm = <string>, ediraMaxseg = 50, ediraSegment = false, endCol = "", fdrLimit = 0.1, iClusterK = 3, iClusterOpt = false, locusCol = "", perm = 1000, pintArm = "", pintMethod = "pSimCCA", scoreLimit = 0.1, transformType = "raw", windowSize = 10 )

Version

1.0

Bundle

microarray

Categories

Integration Copy Number Analysis

Authors

Riku Louhimo (Riku.Louhimo@Helsinki.FI)

Issue tracker

View/Report issues

Requires

edira (R-package) ; mixOmics (R-package) ; pint (R-bioconductor) ; PREDA (R-package) ; SIM (R-bioconductor) ; iCluster (R-bioconductor) ; tilingArray (R-bioconductor)

Source files

component.xml CN2GECollection.r

Usage

Example with default values

Inputs

Name	Type	Mandatory	Description
exprMatrix	LogMatrix	Mandatory	Matrix of expression values.
cnaMatrix	CSV	Mandatory	Matrix of copy number values.
exprAnnotation	AnnotationTable	Mandatory	Annotations for rows. Shall include id, chromosome and locus.
exprRefMatrix	LogMatrix	Optional	Matrix of normalized expression values from reference samples. Only used if algorithm=edira.
cnaRefMatrix	CSV	Optional	Matrix of normalized copy number values from reference samples. Only used if algorithm=edira.
labels	CSV	Optional	Sample groupings for DRImethod=SAM.

Name

Type

Mandatory

Description

exprMatrix

LogMatrix

Mandatory

Matrix of expression values.

cnaMatrix

CSV

Mandatory

Matrix of copy number values.

exprAnnotation

AnnotationTable

Mandatory

Annotations for rows. Shall include id, chromosome and locus.

exprRefMatrix

LogMatrix

Optional

Matrix of normalized expression values from reference samples. Only used if algorithm=edira.

cnaRefMatrix

CSV

Optional

Matrix of normalized copy number values from reference samples. Only used if algorithm=edira.

labels

CSV

Optional

Sample groupings for DRImethod=SAM.

Outputs

Name	Type	Description
concomitantGenes	CSV	Genes that show concomitant expression and copy number aberration.
concomitantGenes2	CSV	Genes that show concomitant (down) expression and copy number aberration. Output is reserved for SODEGIR output.
plot	Latex	Clustering plot for iCluster algorithm.

Name

Type

Description

concomitantGenes

CSV

Genes that show concomitant expression and copy number aberration.

concomitantGenes2

CSV

Genes that show concomitant (down) expression and copy number aberration. Output is reserved for SODEGIR output.

plot

Latex

Clustering plot for iCluster algorithm.

Parameters

Name	Type	Default	Description
DRImethod	string	"pearson"	Specify the integration method for DRI. One of "pearson", "spearman", "ttest" or "SAM", where SAM stands for supervised learing algorithm, and pearson and spearman are variants of the correlation coefficient.
SGRthold	float	0.1	Thresholds for test statistic distribution for SODEGIR.
algorithm	string	(no default)	Define the algorithm for integration. Must be one of "pint", "edira", 'SODEGIR', 'SIM', 'iCluster', 'integrOmics' or "DRI".
ediraMaxseg	int	50	Maximum number of segments for edira segmentation. Only used if algorithm='edira' and ediraSegment='true'.
ediraSegment	boolean	false	Define whether CNA data is segmentated prior to integration. Only used if algorithm=edira.
endCol	string	""	Name of the column with probe end locus information.
fdrLimit	float	0.1	Result fdr cutoff limit. Only used if algorithm=DRI, SODEGIR or SIM.
iClusterK	int	3	Number of K means clusters for iClust.
iClusterOpt	boolean	false	Optimize K for iClust. See 'iClusterK' parameter also. Enabling this slows down processing.
locusCol	string	""	Name of the column with locus information.
perm	int	1000	Number of permutations for null distribution. Only used if algorithm=DRI,SODEGIR.
pintArm	string	""	Arm for which scores are calculated. Must be one of "p", "q" or empty. Default value indicates both arms to be analyzed. Only used if algorithm=pint.
pintMethod	string	"pSimCCA"	Specify the dependency model: "pCCA" probabilistic canonical correlation analysis (Bach and Jordan 2005) "pPCA" probabilistic principal component analysis (Tipping and Bishop 1999) "pFA" probabilistic factor analysis (Rubin and Thayer 1982) "pSimCCA" probabilistic similarity constrained canonical correlation analysis (Lahti et al. 2009) "TPriorpSimCCA" probabilistic similarity constarined canonical correlation analysis with possibility to tune T prior (Lahti et al. 2009) This parameter is only used if algorithm=pint.
scoreLimit	float	0.1	Result cutoff limit. Only used if algorithm=pint.
transformType	string	"raw"	Type of transformation which is applied when executing drsam algorithm. Must be one of "raw", "rank" or "standardize". Only used if algorithm=DRI and DRImethod=SAM.
windowSize	int	10	Window size for pint algorithm sliding window. Only used if algorithm=pint.

Name

Type

Default

Description

DRImethod

string

"pearson"

Specify the integration method for DRI. One of "pearson", "spearman", "ttest" or "SAM", where SAM stands for supervised learing algorithm, and pearson and spearman are variants of the correlation coefficient.

SGRthold

float

0.1

Thresholds for test statistic distribution for SODEGIR.

algorithm

string

(no default)

Define the algorithm for integration. Must be one of "pint", "edira", 'SODEGIR', 'SIM', 'iCluster', 'integrOmics' or "DRI".

ediraMaxseg

int

Maximum number of segments for edira segmentation. Only used if algorithm='edira' and ediraSegment='true'.

ediraSegment

boolean

false

Define whether CNA data is segmentated prior to integration. Only used if algorithm=edira.

endCol

string

Name of the column with probe end locus information.

fdrLimit

float

0.1

Result fdr cutoff limit. Only used if algorithm=DRI, SODEGIR or SIM.

iClusterK

int

Number of K means clusters for iClust.

iClusterOpt

boolean

false

Optimize K for iClust. See 'iClusterK' parameter also. Enabling this slows down processing.

locusCol

string

Name of the column with locus information.

perm

int

1000

Number of permutations for null distribution. Only used if algorithm=DRI,SODEGIR.

pintArm

string

Arm for which scores are calculated. Must be one of "p", "q" or empty. Default value indicates both arms to be analyzed. Only used if algorithm=pint.

pintMethod

string

"pSimCCA"

Specify the dependency model: "pCCA" probabilistic canonical correlation analysis (Bach and Jordan 2005) "pPCA" probabilistic principal component analysis (Tipping and Bishop 1999) "pFA" probabilistic factor analysis (Rubin and Thayer 1982) "pSimCCA" probabilistic similarity constrained canonical correlation analysis (Lahti et al. 2009) "TPriorpSimCCA" probabilistic similarity constarined canonical correlation analysis with possibility to tune T prior (Lahti et al. 2009) This parameter is only used if algorithm=pint.

scoreLimit

float

0.1

Result cutoff limit. Only used if algorithm=pint.

transformType

string

"raw"

Type of transformation which is applied when executing drsam algorithm. Must be one of "raw", "rank" or "standardize". Only used if algorithm=DRI and DRImethod=SAM.

windowSize

int

Window size for pint algorithm sliding window. Only used if algorithm=pint.

Test cases

Test case	Parameters▼	IN exprMatrix	IN cnaMatrix	IN exprAnnotation	IN exprRefMatrix	IN cnaRefMatrix	IN labels	OUT concomitantGenes	OUT concomitantGenes2	OUT plot
case1	properties	exprMatrix	cnaMatrix	exprAnnotation	(missing)	(missing)	(missing)	concomitantGenes	(missing)	(missing)
	algorithm=pint, locusCol=loc
case3_edira	properties	exprMatrix	cnaMatrix	exprAnnotation	exprRefMatrix	cnaRefMatrix	(missing)	concomitantGenes	(missing)	(missing)
	algorithm=edira, locusCol=loc
case4	properties	exprMatrix	cnaMatrix	exprAnnotation	(missing)	(missing)	(missing)	concomitantGenes	(missing)	(missing)
	algorithm=pint, windowSize=15, fdrLimit=0.5, pintArm=q, locusCol=loc,
case5_SIM	properties	exprMatrix	cnaMatrix	exprAnnotation	(missing)	(missing)	(missing)	concomitantGenes	(missing)	(missing)
	algorithm=SIM, locusCol=loc
case6_SODEGIR	properties	exprMatrix	cnaMatrix	exprAnnotation	(missing)	(missing)	(missing)	(missing)	(missing)	(missing)
	algorithm=SODEGIR, perm = 10, fdrLimit=1.0, SGRthold=0.1, locusCol=loc, endCol=loc
case7_iCluster	properties	exprMatrix	cnaMatrix	exprAnnotation	(missing)	(missing)	(missing)	concomitantGenes	(missing)	plot
	algorithm=iCluster, locusCol=loc, iClusterK=5
case8_int	properties	exprMatrix	cnaMatrix	exprAnnotation	(missing)	(missing)	(missing)	(missing)	(missing)	(missing)
	algorithm=integrOmics, locusCol=loc

Test case

Parameters▼

IN
exprMatrix

IN
cnaMatrix

IN
exprAnnotation

IN
exprRefMatrix

IN
cnaRefMatrix

IN
labels

OUT
concomitantGenes

OUT
concomitantGenes2

OUT
plot

case1