Help for package PCLassoReg

Title:

Group Regression Models for Risk Protein Complex Identification

Version:

1.0.0

Description:

Two protein complex-based group regression models (PCLasso and PCLasso2) for risk protein complex identification. PCLasso is a prognostic model that identifies risk protein complexes associated with survival. PCLasso2 is a classification model that identifies risk protein complexes associated with classes. For more information, see Wang and Liu (2021) <doi:10.1093/bib/bbab212>.

Depends:

R (≥ 3.5.0)

License:

GPL-3

Encoding:

UTF-8

LazyData:

true

RoxygenNote:

7.1.1

Imports:

survival, grpreg

URL:

https://github.com/weiliu123/PCLassoReg

BugReports:

https://github.com/weiliu123/PCLassoReg/issues

Suggests:

rmarkdown, knitr

VignetteBuilder:

knitr

NeedsCompilation:

Packaged:

2021-10-25 11:34:16 UTC; liuwei

Author:

Wei Liu

[aut, cre]

Maintainer:

Wei Liu <freelw@qq.com>

Repository:

CRAN

Date/Publication:

2021-10-26 14:50:05 UTC

Protein complexes.

Description

A dataset containing the protein complexes

Usage

PCGroups

Format

A data frame with 3512 rows and 6 variables:

ComplexID: ID of the protein complex
ComplexName: name of the protein complex
Organism: organism
UniprotID: Uniprot IDs of the proteins in the protein complex
EntrezID: Entrez IDs of the proteins in the protein complex
GeneSymbol: gene symbols of the proteins in the protein complex

Source

https://mips.helmholtz-muenchen.de/corum/

Protein complex-based group lasso-Cox model

Description

Construct a PCLasso model based on a gene/protein expression matrix, survival data, and protein complexes.

Usage

PCLasso(
  x,
  y,
  group,
  penalty = c("grLasso", "grMCP", "grSCAD"),
  standardize = TRUE,
  ...
)

Arguments

x

A n x p matrix of gene/protein expression measurements with n samples and p genes/proteins.

y

The time-to-event outcome, as a two-column matrix or Surv object. The first column should be time on study (follow up time); the second column should be a binary variable with 1 indicating that the event has occurred and 0 indicating (right) censoring.

group

A list of groups. The feature (gene/protein) names in group should be consistent with the feature (gene/protein) names in x.

penalty

The penalty to be applied to the model. For group selection, one of grLasso, grMCP, or grSCAD. See grpsurv in the R package grpreg for details.

standardize

Logical flag for x standardization, prior to fitting the model. Default is TRUE.

...

Arguments to be passed to grpsurv in the R package grpreg.

Details

The function PCLasso implements the PCLasso model when the parameter penalty is set to "grLasso". The PCLasso model is a prognostic model which selects important predictors at the protein complex level to achieve accurate prognosis and identify risk protein complexes. The PCLasso model has three inputs: a gene expression matrix, survival data, and protein complexes. It estimates the correlation between gene expression in protein complexes and survival data at the level of protein complexes. Similar to the traditional Lasso-Cox model, PCLasso is based on the Cox PH model and estimates the Cox regression coefficients by maximizing partial likelihood with regularization penalty. The difference is that PCLasso selects features at the level of protein complexes rather than individual genes. Considering that genes usually function by forming protein complexes, PCLasso regards genes belonging to the same protein complex as a group, and constructs a l1/l2 penalty based on the sum (i.e., l1 norm) of the l2 norms of the regression coefficients of the group members to perform the selection of features at the group level. Since a gene may belong to multiple protein complexes, that is, there is overlap between protein complexes, the classical group Lasso-Cox model for non-overlapping groups may lead to false sparse solutions. The PCLasso model deals with the overlapping problem of protein complexes by constructing a latent group Lasso-Cox model. And by reconstructing the gene expression matrix of the protein complexes, the latent group Lasso-Cox model is transformed into a non-overlapping group Lasso-Cox model in an expanded space, which can be directly solved using the classical group Lasso method. Through the final sparse solution, we can predict the patient's risk score based on a small set of protein complexes and identify risk protein complexes that are frequently selected to construct prognostic models. The penalty parameters grSCAD and grMCP can also be used to identify survival-related risk protein complexes. Their penalty for large coefficients is smaller than grLasso, so they tend to choose less risk protein complexes.

Value

An object with S3 class \code{PCLasso} containing:

fit

An object of class grpsurv

complexes.dt

Complexes with features (genes/proteins) not included in x being filtered out.

References

PCLasso: a protein complex-based, group lasso-Cox model for accurate prognosis and risk protein complex discovery. Brief Bioinform, 2021.

Park, H., Niida, A., Miyano, S. and Imoto, S. (2015) Sparse overlapping group lasso for integrative multi-omics analysis. Journal of computational biology: a journal of computational molecular cell biology, 22, 73-84.

Examples

# load data
data(survivalData)
data(PCGroups)

x = survivalData$Exp
y = survivalData$survData

PC.Human <- getPCGroups(Groups = PCGroups, Organism = "Human",
Type = "EntrezID")

# fit PCLasso model
fit.PCLasso <- PCLasso(x, y, group = PC.Human, penalty = "grLasso")

# fit PCSCAD model
fit.PCSCAD <- PCLasso(x, y, group = PC.Human, penalty = "grSCAD")

# fit PCMCP model
fit.PCMCP <- PCLasso(x, y, group = PC.Human, penalty = "grMCP")

Protein complex-based group Lasso-logistic model

Description

Protein complex-based group Lasso-logistic model

Usage

PCLasso2(
  x,
  y,
  group,
  penalty = c("grLasso", "grMCP", "grSCAD"),
  family = c("binomial", "gaussian", "poisson"),
  gamma = 8,
  standardize = TRUE,
  ...
)

Arguments

x

A n x p matrix of gene/protein expression measurements with n samples and p genes/proteins.

y

The response vector.

group

A list of groups. The feature (gene/protein) names in group should be consistent with the feature (gene/protein) names in x.

penalty

The penalty to be applied to the model. For group selection, one of grLasso, grMCP, or grSCAD. See grpreg in the R package grpreg for details.

family

Either "binomial" or "gaussian", depending on the response.

gamma

Tuning parameter of the grSCAD/grMCP penalty. Default is 8.

standardize

Logical flag for x standardization, prior to fitting the model. Default is TRUE.

...

Arguments to be passed to grpreg in the R package grpreg.

Details

The PCLasso2 model is a classification model that selects important predictors at the protein complex level to achieve accurate classification and identify risk protein complexes. The PCLasso2 model has three inputs: a protein expression matrix, a vector of binary response variables, and a number of known protein complexes. It estimates the correlation between protein expression and response variable at the level of protein complexes. Similar to traditional Lasso-logistic model, PCLasso2 is based on the logistic regression model and estimates the logistic regression coefficients by maximizing likelihood function with regularization penalty. The difference is that PCLasso2 selects features at the level of protein complexes rather than individual proteins. Considering that proteins usually function by forming protein complexes, PCLasso2 regards proteins belonging to the same protein complex as a group and constructs a group Lasso penalty (l1/l2 penalty) based on the sum (i.e. l1 norm) of the l2 norms of the regression coefficients of the group members to perform the selection of features at the group level. With the group Lasso penalty, PCLasso2 trains the logistic regression model and obtains a sparse solution at the protein complex level, that is, the proteins belonging to a protein complex are either wholly included or wholly excluded from the model. PCLasso2 outputs a prediction model and a small set of protein complexes included in the model, which are referred to as risk protein complexes. The PCSCAD and PCMCP are performed by setting the penalty parameter penalty as grSCAD and grMCP, respectively.

Value

An object with S3 class PCLasso2 containing:

fit

An object of class grpreg

Complexes.dt

Complexes with features (genes/proteins) not included in x being filtered out.

References

PCLasso2: a protein complex-based, group Lasso-logistic model for risk protein complex discovery. To be published.

PCLasso: a protein complex-based, group lasso-Cox model for accurate prognosis and risk protein complex discovery. Brief Bioinform, 2021.

Examples

# load data
data(classData)
data(PCGroups)

x = classData$Exp
y = classData$Label

PC.Human <- getPCGroups(Groups = PCGroups, Organism = "Human",
Type = "GeneSymbol")

# fit PCLasso2 model
fit.PCLasso2 <- PCLasso2(x, y, group = PC.Human, penalty = "grLasso",
family = "binomial")

# fit PCSCAD model
fit.PCSCAD <- PCLasso2(x, y, group = PC.Human, penalty = "grSCAD",
family = "binomial", gamma = 10)

# fit PCMCP model
fit.PCMCP <- PCLasso2(x, y, group = PC.Human, penalty = "grMCP",
family = "binomial", gamma = 9)

Group Regression Models for Risk Protein Complex Identification

Description

Details

The PCLasso model accepts a protein expression matrix, survival data, and protein complexes for training the prognostic model, and makes predictions for new samples and identifies risk protein complexes associated with survival.

The PCLasso2 model accepts a protein expression matrix, a response vector, and protein complexes for training the classification model, and makes predictions for new samples and identifies risk protein complexes associated with classes.

Both PCLasso and PCLasso2 use grLasso as the penalty function. The other two penalties grSCAD and grMCP can also be used for model construction and risk protein complex identification. The package also provides methods for plotting coefficient paths and cross-validation curves.

References

PCLasso2: a protein complex-based, group Lasso-logistic model for risk protein complex discovery. To be published.

PCLasso: a protein complex-based group lasso-Cox model for accurate prognosis and risk protein complex discovery. Brief Bioinform, 2021.

A dataset for classification

Description

A dataset for classification

Usage

classData

Format

A list containing a protein expression matrix and a response vector

Exp: a protein expression matrix
Label: a response vector

Cross-validation for `PCLasso`

Description

Perform k-fold cross validations for the PCLasso model with grouped covariates over a grid of values for the regularization parameter lambda.

Usage

cv.PCLasso(
  x,
  y,
  group,
  penalty = c("grLasso", "grMCP", "grSCAD"),
  nfolds = 5,
  standardize = TRUE,
  ...
)

Arguments

x

A n x p design matrix of gene/protein expression measurements with n samples and p genes/proteins, as in PCLasso.

y

The time-to-event outcome, as a two-column matrix or Surv object, as in PCLasso. The first column should be time on study (follow up time); the second column should be a binary variable with 1 indicating that the event has occurred and 0 indicating (right) censoring.

group

A list of groups as in PCLasso. The feature (gene/protein) names in group should be consistent with the feature (gene/protein) names in x.

penalty

The penalty to be applied to the model. For group selection, one of grLasso, grMCP, or grSCAD. For bi-level selection, one of gel or cMCP. See grpsurv in the R package grpreg for details.

nfolds

The number of cross-validation folds. Default is 5.

standardize

Logical flag for x standardization, prior to fitting the model. Default is TRUE.

...

Arguments to be passed to cv.grpsurv in the R package grpreg.

Details

The function calls PCLasso nfolds times, each time leaving out 1/nfolds of the data. The cross-validation error is based on the deviance. The numbers for censored samples are balanced across the folds. cv.PCLasso uses the approach of calculating the full Cox partial likelihood using the cross-validated set of linear predictors. See cv.grpsurv in the R package grpreg for details.

Value

An object with S3 class "cv.PCLasso" containing:

cv.fit

An object of class "cv.grpsurv".

complexes.dt

Complexes with features (genes/proteins) not included in x being filtered out.

Author(s)

Wei Liu

References

PCLasso: a protein complex-based, group lasso-Cox model for accurate prognosis and risk protein complex discovery. Brief Bioinform, 2021.

Examples

# load data
data(survivalData)
data(PCGroups)

x = survivalData$Exp
y = survivalData$survData

PC.Human <- getPCGroups(Groups = PCGroups, Organism = "Human",
Type = "EntrezID")

# fit model
cv.fit1 <- cv.PCLasso(x, y, group = PC.Human, penalty = "grLasso",
nfolds = 10)

Cross-validation for `PCLasso2`

Description

Perform k-fold cross validations for the PCLasso2 model with grouped covariates over a grid of values for the regularization parameter lambda.

Usage

cv.PCLasso2(
  x,
  y,
  group,
  penalty = c("grLasso", "grMCP", "grSCAD"),
  family = c("binomial", "gaussian", "poisson"),
  nfolds = 5,
  gamma = 8,
  standardize = TRUE,
  ...
)

Arguments

x

A n x p design matrix of gene/protein expression measurements with n samples and p genes/proteins, as in PCLasso2.

y

The response vector.

group

A list of groups as in PCLasso. The feature (gene/protein) names in group should be consistent with the feature (gene/protein) names in x.

penalty

The penalty to be applied to the model. For group selection, one of grLasso, grMCP, or grSCAD. See grpreg in the R package grpreg for details.

family

Either "binomial" or "gaussian", depending on the response.

nfolds

The number of cross-validation folds. Default is 5.

gamma

Tuning parameter of the grSCAD/grMCP penalty. Default is 8.

standardize

Logical flag for x standardization, prior to fitting the model. Default is TRUE.

...

Arguments to be passed to cv.grpreg in the R package grpreg.

Details

The function calls PCLasso2 nfolds times, each time leaving out 1/nfolds of the data. The cross-validation error is based on the deviance. The numbers for each class are balanced across the folds; i.e., the number of outcomes in which y is equal to 1 is the same for each fold, or possibly off by 1 if the numbers do not divide evenly. See cv.grpreg in the R package grpreg for details.

Value

An object with S3 class "cv.PCLasso2" containing:

cv.fit

An object of class "cv.grpreg".

complexes.dt

Complexes with features (genes/proteins) not included in x being filtered out.

Author(s)

Wei Liu

References

PCLasso2: a protein complex-based, group Lasso-logistic model for risk protein complex discovery. To be published.

Examples

# load data
data(classData)
data(PCGroups)

x = classData$Exp
y = classData$Label

PC.Human <- getPCGroups(Groups = PCGroups, Organism = "Human",
Type = "GeneSymbol")

# fit model
cv.fit1 <- cv.PCLasso2(x, y, group = PC.Human, penalty = "grLasso",
family = "binomial", nfolds = 5)


cv.fit1 <- cv.PCLasso2(x, y, group = PC.Human, penalty = "grSCAD",
family = "binomial", nfolds = 5, gamma = 10)
cv.fit1 <- cv.PCLasso2(x, y, group = PC.Human, penalty = "grMCP",
family = "binomial", nfolds = 5, gamma = 15)

get protein complexes

Description

get protein complexes

Usage

getPCGroups(
  Groups,
  Organism = c("Human", "Mouse", "Rat", "Mammalia", "Bovine", "Dog", "Rabbit"),
  Type = c("GeneSymbol", "EntrezID", "UniprotID")
)

Arguments

Groups

A data frame containing the protein complexes

Organism

Organism. one of Human, Mouse, Rat, Mammalia, Bovine, Dog, or Rabbit.

Type

The name type of the proteins in the protein complexes. One of GeneSymbol, EntrezID, UniprotID

Value

A list of protein complexes

Examples

data(PCGroups)
PC.Human <- getPCGroups(Groups = PCGroups, Organism = "Human",
Type = "GeneSymbol")

Plot coefficients from a PCLasso object

Description

Produces a plot of the coefficient paths for a fitted PCLasso object.

Usage

## S3 method for class 'PCLasso'
plot(x, norm = TRUE, ...)

Arguments

x

Fitted PCLasso model.

norm

If TRUE, plot the norm of each group, rather than the individual coefficients.

...

Other graphical parameters to plot.

Value

No return value, called for plotting of PCLasso objects.

Examples

# load data
data(survivalData)
data(PCGroups)

x = survivalData$Exp
y = survivalData$survData

PC.Human <- getPCGroups(Groups = PCGroups, Organism = "Human",
Type = "EntrezID")

# fit PCLasso model
fit.PCLasso <- PCLasso(x, y, group = PC.Human, penalty = "grLasso")

# plot the norm of each group
plot(fit.PCLasso, norm = TRUE)

# plot the individual coefficients
plot(fit.PCLasso, norm = FALSE)

Plot coefficients from a PCLasso2 object

Description

Produces a plot of the coefficient paths for a fitted PCLasso2 object.

Usage

## S3 method for class 'PCLasso2'
plot(x, norm = TRUE, ...)

Arguments

x

Fitted PCLasso2 model.

norm

If TRUE, plot the norm of each group, rather than the individual coefficients.

...

Other graphical parameters to plot.

Value

No return value, called for plotting of PCLasso2 objects.

Examples

# load data
data(classData)
data(PCGroups)

x = classData$Exp
y = classData$Label

PC.Human <- getPCGroups(Groups = PCGroups, Organism = "Human",
Type = "GeneSymbol")

# fit PCLasso2 model
fit.PCLasso2 <- PCLasso2(x, y, group = PC.Human, penalty = "grLasso")

# plot the norm of each group
plot(fit.PCLasso2, norm = TRUE)

# plot the individual coefficients
plot(fit.PCLasso2, norm = FALSE)

Plot the cross-validation curve from a `cv.PCLasso` object

Description

Plot the cross-validation curve from a cv.PCLasso object, along with standard error bars.

Usage

## S3 method for class 'cv.PCLasso'
plot(x, type = c("cve", "rsq", "snr", "all"), norm = NULL, ...)

Arguments

x

Fitted cv.PCLasso model.

type

What to plot on the vertical axis. "cve" plots the cross-validation error (deviance); "rsq" plots an estimate of the fraction of the deviance explained by the model (R-squared); "snr" plots an estimate of the signal-to-noise ratio; "all" produces all of the above.

norm

If TRUE, plot the norm of each group, rather than the individual coefficients.

...

Other graphical parameters to plot

Details

Error bars representing approximate +/- 1 SE (68% confidence intervals) are plotted along with the estimates at value of lambda. See plot.cv.grpreg in the R package grpreg for details.

Value

No return value, called for plotting of cv.PCLasso objects.

Examples

# load data
data(survivalData)
data(PCGroups)

x = survivalData$Exp
y = survivalData$survData

PC.Human <- getPCGroups(Groups = PCGroups, Organism = "Human",
Type = "EntrezID")

# fit model
cv.fit1 <- cv.PCLasso(x, y, group = PC.Human, penalty = "grLasso",
nfolds = 10)

# plot the norm of each group
plot(cv.fit1, norm = TRUE)

# plot the individual coefficients
plot(cv.fit1, norm = FALSE)

# plot the cross-validation error (deviance)
plot(cv.fit1, type = "cve")

Plot the cross-validation curve from a `cv.PCLasso2` object

Description

Plot the cross-validation curve from a cv.PCLasso2 object, along with standard error bars.

Usage

## S3 method for class 'cv.PCLasso2'
plot(x, type = c("cve", "rsq", "snr", "all"), norm = NULL, ...)

Arguments

x

Fitted cv.PCLasso2 model.

type

norm

If TRUE, plot the norm of each group, rather than the individual coefficients.

...

Other graphical parameters to plot

Details

Error bars representing approximate +/- 1 SE (68% confidence intervals) are plotted along with the estimates at value of lambda. See plot.cv.grpreg in the R package grpreg for details.

Value

No return value, called for plotting of cv.PCLasso2 objects.

Examples

# load data
data(classData)
data(PCGroups)

x = classData$Exp
y = classData$Label

PC.Human <- getPCGroups(Groups = PCGroups, Organism = "Human",
Type = "GeneSymbol")

# fit model
cv.fit1 <- cv.PCLasso2(x, y, group = PC.Human, penalty = "grLasso",
family = "binomial", nfolds = 10)

# plot the norm of each group
plot(cv.fit1, norm = TRUE)

# plot the individual coefficients
plot(cv.fit1, norm = FALSE)

# plot the cross-validation error (deviance)
plot(cv.fit1, type = "cve")

Make predictions from a PCLasso model

Description

Similar to other predict methods, this function returns predictions from a fitted PCLasso object.

Usage

## S3 method for class 'PCLasso'
predict(
  object,
  x = NULL,
  type = c("link", "response", "survival", "median", "norm", "coefficients", "vars",
    "nvars", "vars.unique", "nvars.unique", "groups", "ngroups"),
  lambda,
  ...
)

Arguments

object

Fitted PCLasso model object.

x

Matrix of values at which predictions are to be made. The features (genes/proteins) contained in x should be consistent with those contained in x in the PCLasso function. Not used for type="coefficients" or for some of the type settings in predict.

type

Type of prediction: "link" returns the linear predictors; "response" gives the risk (i.e., exp(link)); "vars" returns the indices for the nonzero coefficients; "vars.unique" returns unique features (genes/proteins) with nonzero coefficients (If a feature belongs to multiple groups and multiple groups are selected, the feature will be repeatedly selected. Compared with "var", "var.unique" will filter out repeated features.); "groups" returns the groups with at least one nonzero coefficient; "nvars" returns the number of nonzero coefficients; "nvars.unique" returns the number of unique features (genes/proteins) with nonzero coefficients; "ngroups" returns the number of groups with at least one nonzero coefficient; "norm" returns the L2 norm of the coefficients in each group."survival" returns the estimated survival function; "median" estimates median survival times.

lambda

Values of the regularization parameter lambda at which predictions are requested. For values of lambda not in the sequence of fitted models, linear interpolation is used.

...

Arguments to be passed to predict.grpsurv in the R package grpreg.

Details

See predict.grpsurv in the R package grpreg for details.

Value

The object returned depends on type.

Examples

# load data
data(survivalData)
data(PCGroups)

x <- survivalData$Exp
y <- survivalData$survData
PC.Human <- getPCGroups(Groups = PCGroups, Organism = "Human",
Type = "EntrezID")

set.seed(20150122)
idx.train <- sample(nrow(x), round(nrow(x)*2/3))
x.train <- x[idx.train,]
y.train <- y[idx.train,]
x.test <- x[-idx.train,]
y.test <- y[-idx.train,]

# fit PCLasso model
fit.PCLasso <- PCLasso(x = x.train, y = y.train, group = PC.Human,
                  penalty = "grLasso")

# predict risk scores of samples in x.test
s <- predict(object = fit.PCLasso, x = x.test, type="link",
lambda=fit.PCLasso$fit$lambda)

s <- predict(object = fit.PCLasso, x = x.test, type="link",
lambda=fit.PCLasso$fit$lambda[10])

# Nonzero coefficients
sel.groups <- predict(object = fit.PCLasso, type="groups",
                      lambda = fit.PCLasso$fit$lambda)
sel.ngroups <- predict(object = fit.PCLasso, type="ngroups",
                       lambda = fit.PCLasso$fit$lambda)
sel.vars.unique <- predict(object = fit.PCLasso, type="vars.unique",
                          lambda = fit.PCLasso$fit$lambda)
sel.nvars.unique <- predict(object = fit.PCLasso, type="nvars.unique",
                            lambda = fit.PCLasso$fit$lambda)
sel.vars <- predict(object = fit.PCLasso, type="vars",
                    lambda=fit.PCLasso$fit$lambda)
sel.nvars <- predict(object = fit.PCLasso, type="nvars",
                     lambda=fit.PCLasso$fit$lambda)

# For values of lambda not in the sequence of fitted models,
# linear interpolation is used.
sel.groups <- predict(object = fit.PCLasso, type="groups",
                      lambda = c(0.1, 0.05))
sel.ngroups <- predict(object = fit.PCLasso, type="ngroups",
                       lambda = c(0.1, 0.05))
sel.vars.unique <- predict(object = fit.PCLasso, type="vars.unique",
                           lambda = c(0.1, 0.05))
sel.nvars.unique <- predict(object = fit.PCLasso, type="nvars.unique",
                            lambda = c(0.1, 0.05))
sel.vars <- predict(object = fit.PCLasso, type="vars",
                    lambda=c(0.1, 0.05))
sel.nvars <- predict(object = fit.PCLasso, type="nvars",
                     lambda=c(0.1, 0.05))

Make predictions from a PCLasso2 model

Description

Similar to other predict methods, this function returns predictions from a fitted PCLasso2 object.

Usage

## S3 method for class 'PCLasso2'
predict(
  object,
  x = NULL,
  type = c("link", "response", "class", "norm", "coefficients", "vars", "nvars",
    "vars.unique", "nvars.unique", "groups", "ngroups"),
  lambda,
  ...
)

Arguments

object

Fitted PCLasso2 model object.

x

Matrix of values at which predictions are to be made. The features (genes/proteins) contained in x should be consistent with those contained in x in the PCLasso2 function. Not used for type="coefficients" or for some of the type settings in predict.

type

Type of prediction: "link" returns the linear predictors; "response" gives the risk (i.e., exp(link)); "class" returns the binomial outcome with the highest probability; "vars" returns the indices for the nonzero coefficients; "vars.unique" returns unique features (genes/proteins) with nonzero coefficients (If a feature belongs to multiple groups and multiple groups are selected, the feature will be repeatedly selected. Compared with "var", "var.unique" will filter out repeated features.); "groups" returns the groups with at least one nonzero coefficient; "nvars" returns the number of nonzero coefficients; "nvars.unique" returns the number of unique features (genes/proteins) with nonzero coefficients; "ngroups" returns the number of groups with at least one nonzero coefficient; "norm" returns the L2 norm of the coefficients in each group.

lambda

Values of the regularization parameter lambda at which predictions are requested. For values of lambda not in the sequence of fitted models, linear interpolation is used.

...

Arguments to be passed to predict.grpreg in the R package grpreg.

Details

See predict.grpreg in the R package grpreg for details.

Value

The object returned depends on type.

Examples

# load data
data(classData)
data(PCGroups)

x <- classData$Exp
y <- classData$Label
PC.Human <- getPCGroups(Groups = PCGroups, Organism = "Human",
Type = "GeneSymbol")

set.seed(20150122)
idx.train <- sample(nrow(x), round(nrow(x)*2/3))
x.train <- x[idx.train,]
y.train <- y[idx.train]
x.test <- x[-idx.train,]
y.test <- y[-idx.train]

# fit PCLasso2 model
fit.PCLasso2 <- PCLasso2(x = x.train, y = y.train, group = PC.Human,
                  penalty = "grLasso", family = "binomial")

# predict risk scores of samples in x.test
s <- predict(object = fit.PCLasso2, x = x.test, type="link",
lambda=fit.PCLasso2$fit$lambda)

# predict classes of samples in x.test
s <- predict(object = fit.PCLasso2, x = x.test, type="class",
lambda=fit.PCLasso2$fit$lambda[10])

# Nonzero coefficients
sel.groups <- predict(object = fit.PCLasso2, type="groups",
                      lambda = fit.PCLasso2$fit$lambda)
sel.ngroups <- predict(object = fit.PCLasso2, type="ngroups",
                       lambda = fit.PCLasso2$fit$lambda)
sel.vars.unique <- predict(object = fit.PCLasso2, type="vars.unique",
                          lambda = fit.PCLasso2$fit$lambda)
sel.nvars.unique <- predict(object = fit.PCLasso2, type="nvars.unique",
                            lambda = fit.PCLasso2$fit$lambda)
sel.vars <- predict(object = fit.PCLasso2, type="vars",
                    lambda=fit.PCLasso2$fit$lambda)
sel.nvars <- predict(object = fit.PCLasso2, type="nvars",
                     lambda=fit.PCLasso2$fit$lambda)

# For values of lambda not in the sequence of fitted models,
# linear interpolation is used.
sel.groups <- predict(object = fit.PCLasso2, type="groups",
                      lambda = c(0.1, 0.05))
sel.ngroups <- predict(object = fit.PCLasso2, type="ngroups",
                       lambda = c(0.1, 0.05))
sel.vars.unique <- predict(object = fit.PCLasso2, type="vars.unique",
                           lambda = c(0.1, 0.05))
sel.nvars.unique <- predict(object = fit.PCLasso2, type="nvars.unique",
                            lambda = c(0.1, 0.05))
sel.vars <- predict(object = fit.PCLasso2, type="vars",
                    lambda=c(0.1, 0.05))
sel.nvars <- predict(object = fit.PCLasso2, type="nvars",
                     lambda=c(0.1, 0.05))

Make predictions from a cross-validated PCLasso model

Description

Similar to other predict methods, this function returns predictions from a fitted cv.PCLasso object, using the optimal value chosen for lambda.

Usage

## S3 method for class 'cv.PCLasso'
predict(
  object,
  x = NULL,
  type = c("link", "response", "survival", "median", "norm", "coefficients", "vars",
    "nvars", "vars.unique", "nvars.unique", "groups", "ngroups"),
  lambda,
  ...
)

Arguments

object

Fitted cv.PCLasso model object.

x

Matrix of values at which predictions are to be made. The features (genes/proteins) contained in x should be consistent with those contained in x in the cv.PCLasso function. Not used for type="coefficients" or for some of the type settings in predict.

type

Type of prediction: "link" returns the linear predictors; "response" gives the risk (i.e., exp(link)); "vars" returns the indices for the nonzero coefficients; "vars.unique" returns unique features (genes/proteins) with nonzero coefficients (If a feature belongs to multiple groups and multiple groups are selected, the feature will be repeatedly selected. Compared with "var", "var.unique" will filter out repeated features.); "groups" returns the groups with at least one nonzero coefficient; "nvars" returns the number of nonzero coefficients; "nvars.unique" returens the number of unique features (genes/proteins) with nonzero coefficients; "ngroups" returns the number of groups with at least one nonzero coefficient; "norm" returns the L2 norm of the coefficients in each group."survival" returns the estimated survival function; "median" estimates median survival times.

lambda

Values of the regularization parameter lambda at which predictions are requested. For values of lambda not in the sequence of fitted models, linear interpolation is used.

...

Arguments to be passed to predict.cv.grpsurv in the R package grpreg.

Value

The object returned depends on type.

Examples

# load data
data(survivalData)
data(PCGroups)

x <- survivalData$Exp
y <- survivalData$survData
PC.Human <- getPCGroups(Groups = PCGroups, Organism = "Human",
Type = "EntrezID")

set.seed(20150122)
idx.train <- sample(nrow(x), round(nrow(x)*2/3))
x.train <- x[idx.train,]
y.train <- y[idx.train,]
x.test <- x[-idx.train,]
y.test <- y[-idx.train,]

# fit cv.PCLasso model
cv.fit1 <- cv.PCLasso(x = x.train,
                      y = y.train,
                      group = PC.Human,
                      nfolds = 5)

# predict risk scores of samples in x.test
s <- predict(object = cv.fit1, x = x.test, type="link",
             lambda=cv.fit1$cv.fit$lambda.min)

# Nonzero coefficients
sel.groups <- predict(object = cv.fit1, type="groups",
                      lambda = cv.fit1$cv.fit$lambda.min)
sel.ngroups <- predict(object = cv.fit1, type="ngroups",
                       lambda = cv.fit1$cv.fit$lambda.min)
sel.vars.unique <- predict(object = cv.fit1, type="vars.unique",
                           lambda = cv.fit1$cv.fit$lambda.min)
sel.nvars.unique <- predict(object = cv.fit1, type="nvars.unique",
                            lambda = cv.fit1$cv.fit$lambda.min)
sel.vars <- predict(object = cv.fit1, type="vars",
                    lambda=cv.fit1$cv.fit$lambda.min)
sel.nvars <- predict(object = cv.fit1, type="nvars",
                     lambda=cv.fit1$cv.fit$lambda.min)

Make predictions from a cross-validated PCLasso2 model

Description

Similar to other predict methods, this function returns predictions from a fitted cv.PCLasso2 object, using the optimal value chosen for lambda.

Usage

## S3 method for class 'cv.PCLasso2'
predict(
  object,
  x = NULL,
  type = c("link", "response", "class", "norm", "coefficients", "vars", "nvars",
    "vars.unique", "nvars.unique", "groups", "ngroups"),
  lambda,
  ...
)

Arguments

object

Fitted cv.PCLasso2 model object.

x

Matrix of values at which predictions are to be made. The features (genes/proteins) contained in x should be consistent with those contained in x in the cv.PCLasso2 function. Not used for type="coefficients" or for some of the type settings in predict.

type

lambda

Values of the regularization parameter lambda at which predictions are requested. For values of lambda not in the sequence of fitted models, linear interpolation is used.

...

Arguments to be passed to predict.cv.grpreg in the R package grpreg.

Value

The object returned depends on type.

Examples

# load data
data(classData)
data(PCGroups)

x = classData$Exp
y = classData$Label

PC.Human <- getPCGroups(Groups = PCGroups, Organism = "Human",
Type = "GeneSymbol")

#' set.seed(20150122)
idx.train <- sample(nrow(x), round(nrow(x)*2/3))
x.train <- x[idx.train,]
y.train <- y[idx.train]
x.test <- x[-idx.train,]
y.test <- y[-idx.train]

# fit model
cv.fit1 <- cv.PCLasso2(x = x.train, y = y.train, group = PC.Human,
penalty = "grLasso", family = "binomial", nfolds = 10)

# predict risk scores of samples in x.test
s <- predict(object = cv.fit1, x = x.test, type="link",
             lambda=cv.fit1$cv.fit$lambda.min)

# predict classes of samples in x.test
s <- predict(object = cv.fit1, x = x.test, type="class",
             lambda=cv.fit1$cv.fit$lambda.min)

# Nonzero coefficients
sel.groups <- predict(object = cv.fit1, type="groups",
                      lambda = cv.fit1$cv.fit$lambda.min)
sel.ngroups <- predict(object = cv.fit1, type="ngroups",
                       lambda = cv.fit1$cv.fit$lambda.min)
sel.vars.unique <- predict(object = cv.fit1, type="vars.unique",
                           lambda = cv.fit1$cv.fit$lambda.min)
sel.nvars.unique <- predict(object = cv.fit1, type="nvars.unique",
                            lambda = cv.fit1$cv.fit$lambda.min)
sel.vars <- predict(object = cv.fit1, type="vars",
                    lambda=cv.fit1$cv.fit$lambda.min)
sel.nvars <- predict(object = cv.fit1, type="nvars",
                     lambda=cv.fit1$cv.fit$lambda.min)

A dataset for prognostic model

Description

A dataset for prognostic model

Usage

survivalData

Format

A list containing a protein expression matrix and survival data

Exp: a protein expression matrix
survData: Survival data. The first column is the time on study (follow up time); the second column is a binary variable with 1 indicating that the event has occurred and 0 indicating right censoring.

Protein complexes.

Description

Usage

Format

Source

Protein complex-based group lasso-Cox model

Description

Usage

Arguments

Details

Value

References

See Also

Examples

Protein complex-based group Lasso-logistic model

Description

Usage

Arguments

Details

Value

References

See Also

Examples

Group Regression Models for Risk Protein Complex Identification

Description

Details

References

A dataset for classification

Description

Usage

Format

Cross-validation for PCLasso

Description

Usage

Arguments

Details

Value

Author(s)

References

See Also

Examples

Cross-validation for PCLasso2

Description

Usage

Arguments

Details

Value

Author(s)

References

See Also

Examples

get protein complexes

Description

Usage

Arguments

Value

Examples

Plot coefficients from a PCLasso object

Description

Usage

Arguments

Value

See Also

Examples

Plot coefficients from a PCLasso2 object

Description

Usage

Arguments

Value

See Also

Examples

Plot the cross-validation curve from a cv.PCLasso object

Description

Usage

Arguments

Details

Value

See Also

Examples

Plot the cross-validation curve from a cv.PCLasso2 object

Cross-validation for `PCLasso`

Cross-validation for `PCLasso2`

Plot the cross-validation curve from a `cv.PCLasso` object

Plot the cross-validation curve from a `cv.PCLasso2` object