Package 'RLassoCox' reference manual

Title:	A reweighted Lasso-Cox by integrating gene interaction information
Description:	RLassoCox is a package that implements the RLasso-Cox model proposed by Wei Liu. The RLasso-Cox model integrates gene interaction information into the Lasso-Cox model for accurate survival prediction and survival biomarker discovery. It is based on the hypothesis that topologically important genes in the gene interaction network tend to have stable expression changes. The RLasso-Cox model uses random walk to evaluate the topological weight of genes, and then highlights topologically important genes to improve the generalization ability of the Lasso-Cox model. The RLasso-Cox model has the advantage of identifying small gene sets with high prognostic performance on independent datasets, which may play an important role in identifying robust survival biomarkers for various cancer types.
Authors:	Wei Liu [cre, aut]
Maintainer:	Wei Liu <[email protected]>
License:	Artistic-2.0
Version:	1.15.0
Built:	2024-11-19 04:10:39 UTC
Source:	https://github.com/bioc/RLassoCox

A reweighted Lasso-Cox by integrating gene interaction information

Description

RLassoCox is a package that implements the RLasso-Cox model proposed by Wei Liu. The RLasso-Cox model integrates gene interaction information into the Lasso-Cox model for accurate survival prediction and survival biomarker discovery. It is based on the hypothesis that topologically important genes in the gene interaction network tend to have stable expression changes. The RLasso-Cox model uses random walk to evaluate the topological weight of genes, and then highlights topologically important genes to improve the generalization ability of the Lasso-Cox model. The RLasso-Cox model has the advantage of identifying small gene sets with high prognostic performance on independent datasets, which may play an important role in identifying robust survival biomarkers for various cancer types.

Details

The DESCRIPTION file:

Package:	RLassoCox
Type:	Package
Title:	A reweighted Lasso-Cox by integrating gene interaction information
Version:	1.15.0
Date:	2020-10-21
Authors@R:	c(person(given = "Wei", family = "Liu", email = "[email protected]", role = c("cre", "aut"),comment = c(ORCID = "0000-0002-5496-3641")))
Depends:	R (>= 4.1), glmnet
Imports:	Matrix, igraph, survival, stats
Description:	RLassoCox is a package that implements the RLasso-Cox model proposed by Wei Liu. The RLasso-Cox model integrates gene interaction information into the Lasso-Cox model for accurate survival prediction and survival biomarker discovery. It is based on the hypothesis that topologically important genes in the gene interaction network tend to have stable expression changes. The RLasso-Cox model uses random walk to evaluate the topological weight of genes, and then highlights topologically important genes to improve the generalization ability of the Lasso-Cox model. The RLasso-Cox model has the advantage of identifying small gene sets with high prognostic performance on independent datasets, which may play an important role in identifying robust survival biomarkers for various cancer types.
License:	Artistic-2.0
biocViews:	Survival, Regression, GeneExpression, GenePrediction, Network
BugReports:	https://github.com/weiliu123/RLassoCox/issues
BiocType:	Software
Suggests:	knitr
VignetteBuilder:	knitr
Config/pak/sysreqs:	libglpk-dev libxml2-dev
Repository:	https://bioc.r-universe.dev
RemoteUrl:	https://github.com/bioc/RLassoCox
RemoteRef:	HEAD
RemoteSha:	9f0003eb903429f42429632ae254f922eaa28252
Author:	Wei Liu [cre, aut] (<https://orcid.org/0000-0002-5496-3641>)
Maintainer:	Wei Liu <[email protected]>

Index of help topics:

RLassoCox               Reweighted Lasso-Cox model
RLassoCox-package       A reweighted Lasso-Cox by integrating gene
                        interaction information
cvRLassoCox             Cross-validation for the RLasso-Cox model
dGMMirGraph             The KEGG network
mRNA_matrix             The expression data
predict.RLassoCox       Make predictions from a RLasso-Cox model
predict.cvRLassoCox     Make predictions from a cross-validated
                        RLasso-Cox model
rw                      Directed Random Walk
survData                Survival data

Very simple to use. Accepts x,y data for the RLasso-Cox model, and makes predictions for new samples.

RLassoCox A rewighted Lasso-Cox model for survival prediction and biomarker discovery. predict.RLassoCox This function predicts the risk of new samples from a fitted RLasso-Cox model. cvRLassoCox Does k-fold cross-validation for the RLasso-Cox model, produces a plot, and returns a value for lambda predict.cvRLassoCox This function makes predictions from a cross-validated RLasso-Cox model, using the optimal value chosen for lambda.

Author(s)

Wei Liu [cre, aut] (<https://orcid.org/0000-0002-5496-3641>)

Maintainer: Wei Liu <[email protected]>

References

Integration of gene interaction information into a reweighted Lasso-Cox model for accurate survival prediction. To be published.

Examples

library("survival")
library("igraph")
library("glmnet")
library("Matrix")

data(dGMMirGraph)
data(mRNA_matrix)
data(survData)

trainSmpl.Idx <- sample(1:dim(mRNA_matrix)[1], floor(2/3*dim(mRNA_matrix)[1]))
testSmpl.Idx <- setdiff(1:dim(mRNA_matrix)[1], trainSmpl.Idx)
trainSmpl <- mRNA_matrix[trainSmpl.Idx ,]
testSmpl <- mRNA_matrix[testSmpl.Idx ,]

res <- RLassoCox(x=trainSmpl, y=survData[trainSmpl.Idx ,], 
                globalGraph=dGMMirGraph)
lp <- predict(object = res, newx = testSmpl)

cv.res <- cvRLassoCox(x=trainSmpl, y=survData[trainSmpl.Idx ,], 
                        globalGraph=dGMMirGraph, nfolds = 5)
cv.lp <- predict(object = cv.res, newx = testSmpl, 
                            s = "lambda.min")
library("survival")
library("igraph")
library("glmnet")
library("Matrix")

data(dGMMirGraph)
data(mRNA_matrix)
data(survData)

trainSmpl.Idx <- sample(1:dim(mRNA_matrix)[1], floor(2/3*dim(mRNA_matrix)[1]))
testSmpl.Idx <- setdiff(1:dim(mRNA_matrix)[1], trainSmpl.Idx)
trainSmpl <- mRNA_matrix[trainSmpl.Idx ,]
testSmpl <- mRNA_matrix[testSmpl.Idx ,]

res <- RLassoCox(x=trainSmpl, y=survData[trainSmpl.Idx ,], 
                globalGraph=dGMMirGraph)
lp <- predict(object = res, newx = testSmpl)

cv.res <- cvRLassoCox(x=trainSmpl, y=survData[trainSmpl.Idx ,], 
                        globalGraph=dGMMirGraph, nfolds = 5)
cv.lp <- predict(object = cv.res, newx = testSmpl, 
                            s = "lambda.min")

Cross-validation for the RLasso-Cox model

Description

Does k-fold cross-validation for the RLasso-Cox model, produces a plot, and returns a value for lambda

Usage

cvRLassoCox(x, y, globalGraph = NULL, nfolds = 10, Gamma = 0.3, 
            DEBUG = TRUE, standardize = TRUE, ...)
cvRLassoCox(x, y, globalGraph = NULL, nfolds = 10, Gamma = 0.3, 
            DEBUG = TRUE, standardize = TRUE, ...)

Arguments

`x`	a n x p matrix of gene expression measurements with n samples and p genes.
`y`	a n x 2 matrix of survival data. The two columns represent disease status 'status' and survival time 'time' respectively.
`globalGraph`	An `igraph` R object containing the interaction network.
`nfolds`	number of folds - default is 10.
`Gamma`	A numeric value. The restart probability in directed random walk. Default is `Gamma = 0.3`.
`DEBUG`	Logical. Should debugging information be shown.
`standardize`	Logical flag for `x` standardization, prior to fitting the model. Default is `TRUE`.
`...`	Arguments to be passed to `cv.glmnet` in R package `glmnet`.

Value

`glmnetRes`	An object of class "cv.glmnet"
`PT`	The topological weights of genes

Author(s)

Wei Liu

References

Integration of gene interaction information into a reweighted Lasso-Cox model for accurate survival prediction. To be published.

Examples

library("survival")
library("igraph")
library("glmnet")
library("Matrix")

data(dGMMirGraph)
data(mRNA_matrix)
data(survData)

trainSmpl.Idx <- sample(1:dim(mRNA_matrix)[1], floor(2/3*dim(mRNA_matrix)[1]))
testSmpl.Idx <- setdiff(1:dim(mRNA_matrix)[1], trainSmpl.Idx)
trainSmpl <- mRNA_matrix[trainSmpl.Idx ,]
testSmpl <- mRNA_matrix[testSmpl.Idx ,]

cv.res <- cvRLassoCox(x=trainSmpl, y=survData[trainSmpl.Idx ,], 
                        globalGraph=dGMMirGraph, nfolds = 5)

library("survival")
library("igraph")
library("glmnet")
library("Matrix")

data(dGMMirGraph)
data(mRNA_matrix)
data(survData)

trainSmpl.Idx <- sample(1:dim(mRNA_matrix)[1], floor(2/3*dim(mRNA_matrix)[1]))
testSmpl.Idx <- setdiff(1:dim(mRNA_matrix)[1], trainSmpl.Idx)
trainSmpl <- mRNA_matrix[trainSmpl.Idx ,]
testSmpl <- mRNA_matrix[testSmpl.Idx ,]

cv.res <- cvRLassoCox(x=trainSmpl, y=survData[trainSmpl.Idx ,], 
                        globalGraph=dGMMirGraph, nfolds = 5)

The KEGG network

Description

The KEGG network constructed by the R package iSubpathwayMiner.

Usage

data("dGMMirGraph")data("dGMMirGraph")

Format

An igraph R object.

Details

There are 7159 nodes and 39930 edges in dGMMirGraph. Each node in the graph represents a gene/miRNA/metabolite. The KEGG network is used to evaluate the topological importance of genes by the random walk method.

Examples

data(dGMMirGraph)
data(dGMMirGraph)

The expression data

Description

An example of GBM expression data. We acknowledge the TCGA Research Network for generating the GBM datasets.

Usage

data("mRNA_matrix")data("mRNA_matrix")

Format

The format is: num [1:314, 1:4853] 0.562167 0.022435 -0.000102 -0.719444 0.620269 ... - attr(*, "dimnames")=List of 2 ..$ : chr [1:314] "TCGA-02-0001" "TCGA-02-0003" "TCGA-02-0006" ..$ : chr [1:4853] "90993" "4313" "26248" "57680" ...

Examples

data(mRNA_matrix)
data(mRNA_matrix)

Make predictions from a cross-validated RLasso-Cox model

Description

This function makes predictions from a cross-validated RLasso-Cox model, using the optimal value chosen for lambda.

Usage

## S3 method for class 'cvRLassoCox'
predict(object, newx, ...)
## S3 method for class 'cvRLassoCox'
predict(object, newx, ...)

Arguments

`object`	cross-validated RLasso-Cox model
`newx`	A matrix with new samples to predict.
`...`	Arguments to be passed to `predict.cv.glmnet` in R package `glmnet`.

Value

Predicted results of new patients in newx.

Examples

library("survival")
library("igraph")
library("glmnet")
library("Matrix")

data(dGMMirGraph)
data(mRNA_matrix)
data(survData)

trainSmpl.Idx <- sample(1:dim(mRNA_matrix)[1], floor(2/3*dim(mRNA_matrix)[1]))
testSmpl.Idx <- setdiff(1:dim(mRNA_matrix)[1], trainSmpl.Idx)
trainSmpl <- mRNA_matrix[trainSmpl.Idx ,]
testSmpl <- mRNA_matrix[testSmpl.Idx ,]

cv.res <- cvRLassoCox(x=trainSmpl, y=survData[trainSmpl.Idx ,], 
                        globalGraph=dGMMirGraph, nfolds = 5)
lp <- predict(object = cv.res, newx = testSmpl, 
                            s = "lambda.min")
library("survival")
library("igraph")
library("glmnet")
library("Matrix")

data(dGMMirGraph)
data(mRNA_matrix)
data(survData)

trainSmpl.Idx <- sample(1:dim(mRNA_matrix)[1], floor(2/3*dim(mRNA_matrix)[1]))
testSmpl.Idx <- setdiff(1:dim(mRNA_matrix)[1], trainSmpl.Idx)
trainSmpl <- mRNA_matrix[trainSmpl.Idx ,]
testSmpl <- mRNA_matrix[testSmpl.Idx ,]

cv.res <- cvRLassoCox(x=trainSmpl, y=survData[trainSmpl.Idx ,], 
                        globalGraph=dGMMirGraph, nfolds = 5)
lp <- predict(object = cv.res, newx = testSmpl, 
                            s = "lambda.min")

Make predictions from a RLasso-Cox model

Description

This function predicts the risk of new samples from a fitted RLasso-Cox model.

Usage

## S3 method for class 'RLassoCox'
predict(object, newx, ...)
## S3 method for class 'RLassoCox'
predict(object, newx, ...)

Arguments

`object`	Fitted `"RLassoCox"` model object.
`newx`	A matrix with new samples to predict.
`...`	Arguments to be passed to `predict.glmnet` in R package `glmnet`.

Value

Predicted results of new patients in newx.

Author(s)

Wei Liu

Examples

library("survival")
library("igraph")
library("glmnet")
library("Matrix")

data(dGMMirGraph)
data(mRNA_matrix)
data(survData)

trainSmpl.Idx <- sample(1:dim(mRNA_matrix)[1], floor(2/3*dim(mRNA_matrix)[1]))
testSmpl.Idx <- setdiff(1:dim(mRNA_matrix)[1], trainSmpl.Idx)
trainSmpl <- mRNA_matrix[trainSmpl.Idx ,]
testSmpl <- mRNA_matrix[testSmpl.Idx ,]

res <- RLassoCox(x=trainSmpl, y=survData[trainSmpl.Idx ,],
                globalGraph=dGMMirGraph)
lp <- predict(object = res, newx = testSmpl)
library("survival")
library("igraph")
library("glmnet")
library("Matrix")

data(dGMMirGraph)
data(mRNA_matrix)
data(survData)

trainSmpl.Idx <- sample(1:dim(mRNA_matrix)[1], floor(2/3*dim(mRNA_matrix)[1]))
testSmpl.Idx <- setdiff(1:dim(mRNA_matrix)[1], trainSmpl.Idx)
trainSmpl <- mRNA_matrix[trainSmpl.Idx ,]
testSmpl <- mRNA_matrix[testSmpl.Idx ,]

res <- RLassoCox(x=trainSmpl, y=survData[trainSmpl.Idx ,],
                globalGraph=dGMMirGraph)
lp <- predict(object = res, newx = testSmpl)

Reweighted Lasso-Cox model

Description

A rewighted Lasso-Cox model for survival prediction and biomarker discovery.

Usage

RLassoCox(x, y, globalGraph = NULL, Gamma = 0.3, DEBUG = TRUE, 
            standardize = TRUE, ...)
RLassoCox(x, y, globalGraph = NULL, Gamma = 0.3, DEBUG = TRUE, 
            standardize = TRUE, ...)

Arguments

`x`	a n x p matrix of gene expression measurements with n samples and p genes.
`y`	a n x 2 matrix of survival data. The two columns represent disease status 'status' and survival time 'time' respectively.
`globalGraph`	An `igraph` R object containing the interaction network.
`Gamma`	A numeric value. The restart probability in directed random walk. Default is `Gamma = 0.3`.
`DEBUG`	Logical. Should debugging information be shown.
`standardize`	Logical flag for `x` standardization, prior to fitting the model. Default is `TRUE`.
`...`	Arguments to be passed to `glmnet` in R package `glmnet`.

Details

RLassoCox integrates gene interaction information into the Lasso-Cox model for accurate survial prediction and biomarker discovery.

Value

`glmnetRes`	An object of class "glmnet"
`PT`	The topological weights of genes

Author(s)

Wei Liu

References

Integration of gene interaction information into a reweighted Lasso-Cox model for accurate survival prediction. To be published.

Examples

library("survival")
library("igraph")
library("glmnet")
library("Matrix")

data(dGMMirGraph)
data(mRNA_matrix)
data(survData)

trainSmpl.Idx <- sample(1:dim(mRNA_matrix)[1], floor(2/3*dim(mRNA_matrix)[1]))
testSmpl.Idx <- setdiff(1:dim(mRNA_matrix)[1], trainSmpl.Idx)
trainSmpl <- mRNA_matrix[trainSmpl.Idx ,]
testSmpl <- mRNA_matrix[testSmpl.Idx ,]

res <- RLassoCox(x=trainSmpl, y=survData[trainSmpl.Idx ,], 
                globalGraph=dGMMirGraph)
library("survival")
library("igraph")
library("glmnet")
library("Matrix")

data(dGMMirGraph)
data(mRNA_matrix)
data(survData)

trainSmpl.Idx <- sample(1:dim(mRNA_matrix)[1], floor(2/3*dim(mRNA_matrix)[1]))
testSmpl.Idx <- setdiff(1:dim(mRNA_matrix)[1], trainSmpl.Idx)
trainSmpl <- mRNA_matrix[trainSmpl.Idx ,]
testSmpl <- mRNA_matrix[testSmpl.Idx ,]

res <- RLassoCox(x=trainSmpl, y=survData[trainSmpl.Idx ,], 
                globalGraph=dGMMirGraph)

Directed Random Walk

Description

The directed random walk algorithm proposed by Liu et al (2013).

Usage

rw(W, p0, gamma)
rw(W, p0, gamma)

Arguments

`W`	The adjacency matrix of the gene interaction network.
`p0`	A vector containing the initial weights of genes in the gene interaction network.
`gamma`	A numeric value. The restart probability in directed random walk.

Details

This function implements the directed random walk algorithm proposed by Liu et al (2013). It evaluates the topological weight of each gene according to its topological importance in the gene interaction network. The genes that close to many other genes that have large weights will receive larger weights. The final weights reflect the topological importances of genes in the gene interaction network.

Value

A matrix containing the topological weights of nodes in igraphM.

Author(s)

Wei Liu <[email protected]>

References

Liu, W., et al., Topologically inferring risk-active pathways toward precise cancer classification by directed random walk. Bioinformatics, 2013. 29(17): p. 2169-77.

Survival data

Description

The survival data of patients in mRNA_matrix.

Usage

data("survData")data("survData")

Format

A data frame with 314 observations on the following 2 variables.

status: a logical vector
time: a numeric vector

Examples

data(survData)
data(survData)

Package 'RLassoCox'

Help Index

A reweighted Lasso-Cox by integrating gene interaction information

Description

Details

Author(s)

References

Examples

Cross-validation for the RLasso-Cox model

Description

Usage

Arguments

Value

Author(s)

References

Examples

The KEGG network

Description

Usage

Format

Details

Examples

The expression data

Description

Usage

Format

Examples

Make predictions from a cross-validated RLasso-Cox model

Description

Usage

Arguments

Value

Examples

Make predictions from a RLasso-Cox model

Description

Usage

Arguments

Value

Author(s)

Examples

Reweighted Lasso-Cox model

Description

Usage

Arguments

Details

Value

Author(s)

References

See Also

Examples

Directed Random Walk

Description

Usage

Arguments

Details

Value

Author(s)

References

Survival data

Description

Usage

Format

Examples