Package 'netprioR' reference manual

Title:	A model for network-based prioritisation of genes
Description:	A model for semi-supervised prioritisation of genes integrating network data, phenotypes and additional prior knowledge about TP and TN gene labels from the literature or experts.
Authors:	Fabian Schmich
Maintainer:	Fabian Schmich <[email protected]>
License:	GPL-3
Version:	1.31.0
Built:	2024-07-15 05:00:43 UTC
Source:	https://github.com/bioc/netprioR

Package: netprioR

Description

This package provides a model for semi-supervised prioritisation of genes integrating network data, phenotypes and additional prior knowledge about TP and TN gene labels.

Author(s)

Fabian Schmich | Computational Biology Group, ETH Zurich | [email protected]

References

Fabian Schmich et. al (2016).

Class Mass Normalization (CMN) from Zhu et al., 2003

Description

Class Mass Normalization (CMN) from Zhu et al., 2003

Usage

cmn(yhat, l, u)
cmn(yhat, l, u)

Arguments

`yhat`	Response for labeled (l) and unlabeld (u) genes
`l`	Indices of labeled genes
`u`	Indices of unlabeled genes

Value

Class normalized yhat

Author(s)

Fabian Schmich

Conjugate Gradient Solver

Description

Solves linear equation systems iteratively

Usage

conjugate_gradient(A, b, x0 = rep(0, ncol(A)), threshold = 1e-15,
  verbose = FALSE)
conjugate_gradient(A, b, x0 = rep(0, ncol(A)), threshold = 1e-15,
  verbose = FALSE)

Arguments

`A`	Matrix
`b`	Coefficients
`x0`	Starting solution
`threshold`	Termination threshold
`verbose`	Show iterative progress

Value

Solution for equation system

Author(s)

Fabian Schmich

Fit `netprioR` model

Description

Fit netprioR model

Usage

fit(object, ...)

## S4 method for signature 'netprioR'
fit(object, refit = FALSE, ...)
fit(object, ...)

## S4 method for signature 'netprioR'
fit(object, refit = FALSE, ...)

Arguments

`object`	A `netprioR` object
`...`	Additional arguments
`refit`	Flag whether to overwrite existing fit

Value

A netprioR object with fitted model

Author(s)

Fabian Schmich

Examples

data(simulation)
np <- netprioR(networks = simulation$networks,
               phenotypes = simulation$phenotypes,
               labels = simulation$labels.obs,
               model.fit = FALSE)
summary(np)
np <- fit(np, nrestarts = 1, verbose = FALSE)
summary(np)
data(simulation)
np <- netprioR(networks = simulation$networks,
               phenotypes = simulation$phenotypes,
               labels = simulation$labels.obs,
               model.fit = FALSE)
summary(np)
np <- fit(np, nrestarts = 1, verbose = FALSE)
summary(np)

Fit netprioR model

Description

Infer parameters and hidden data using the EM algorithm of netprioR

Usage

learn(Yobs, X, G, l, u, a = 0.1, b = 0.1, sigma2 = 1, tau2 = 10,
  eps = 1e-11, max.iter = 500, thresh = 0.001, use.cg = TRUE,
  thresh.cg = 1e-05, nrestarts = 5, max.cores = detectCores(),
  verbose = FALSE)
learn(Yobs, X, G, l, u, a = 0.1, b = 0.1, sigma2 = 1, tau2 = 10,
  eps = 1e-11, max.iter = 500, thresh = 0.001, use.cg = TRUE,
  thresh.cg = 1e-05, nrestarts = 5, max.cores = detectCores(),
  verbose = FALSE)

Arguments

`Yobs`	Observed labels (NA, if not observed)
`X`	Phenotypes
`G`	Graph Laplacians
`l`	Indices of labelled instances
`u`	Indices of unlabelled instances
`a`	Shape parameter of Gamma prior for W
`b`	Scale parameter of Gamma prior for W
`sigma2`	Cariance for Gaussian labels
`tau2`	Variance for Gaussian prior for beta
`eps`	Small value added to diagonal of Q in order to make it non-singular
`max.iter`	Maximum number of iterations for EM
`thresh`	Threshold for termination of EM with respect to change in parameters
`use.cg`	Flag whether to use conjugate gradient instead of exact computation of expectations
`thresh.cg`	Threshold for the termination of the conjugate gradient solver
`nrestarts`	Number of restarts for EM
`max.cores`	Maximum number of cores to use for parallel computation
`verbose`	Print verbose output

Value

List containing: Predicted labels Yhat and inferred parameters W and beta

Author(s)

Fabian Schmich

netprioR

Description

Class that represents a netprioR model.

Usage

netprioR(networks, phenotypes, labels, ...)

## S4 method for signature 'list,matrix,factor'
netprioR(networks, phenotypes, labels,
  fit.model = FALSE, a = 0.1, b = 0.1, sigma2 = 0.1, tau2 = 100,
  eps = 1e-10, max.iter = 500, thresh = 1e-06, use.cg = FALSE,
  thresh.cg = 1e-06, nrestarts = 5, max.cores = detectCores(),
  verbose = TRUE, ...)
netprioR(networks, phenotypes, labels, ...)

## S4 method for signature 'list,matrix,factor'
netprioR(networks, phenotypes, labels,
  fit.model = FALSE, a = 0.1, b = 0.1, sigma2 = 0.1, tau2 = 100,
  eps = 1e-10, max.iter = 500, thresh = 1e-06, use.cg = FALSE,
  thresh.cg = 1e-06, nrestarts = 5, max.cores = detectCores(),
  verbose = TRUE, ...)

Arguments

`networks`	List of NxN adjacency matrices of gene-gene similarities
`phenotypes`	Matrix of dimension NxP containing covariates
`labels`	Vector of Nx1 labels for all genes (NA if no label available)
`...`	Additional arguments
`fit.model`	Indicator whether to fit the model
`a`	Shape parameter of Gamma prior for W
`b`	Scale parameter of Gamma prior for W
`sigma2`	Cariance for Gaussian labels
`tau2`	Variance for Gaussian prior for beta
`eps`	Small value added to diagonal of Q in order to make it non-singular
`max.iter`	Maximum number of iterations for EM
`thresh`	Threshold for termination of EM with respect to change in parameters
`use.cg`	Flag whether to use conjugate gradient instead of exact computation of expectations
`thresh.cg`	Threshold for the termination of the conjugate gradient solver
`nrestarts`	Number of restarts for EM
`max.cores`	Maximum number of cores to use for parallel computation
`verbose`	Print verbose output

Value

A netprioR object

Slots

networks: List of NxN adjacency matrices of gene-gene similarities
phenotypes: Matrix of dimension NxP containing covariates
labels: Vector of Nx1 labels for all genes. NA if no label available.
is.fitted: Flag indicating if model is fitted
model: List containing estimated parameters and imputed missing data

Author(s)

Fabian Schmich

Examples

 # runs long-ish
data(simulation)
np <- netprioR(networks = simulation$networks,
               phenotypes = simulation$phenotypes,
               labels = simulation$labels.obs,
               fit.model = TRUE)
summary(np)          

# runs long-ish
data(simulation)
np <- netprioR(networks = simulation$networks,
               phenotypes = simulation$phenotypes,
               labels = simulation$labels.obs,
               fit.model = TRUE)
summary(np)

Plot method for `netprioR` objects

Description

Plot method for netprioR objects

Usage

## S3 method for class 'netprioR'
plot(x, which = c("all", "weights", "lik", "scores"), ...)
## S3 method for class 'netprioR'
plot(x, which = c("all", "weights", "lik", "scores"), ...)

Arguments

`x`	A `netprioR` object
`which`	Flag for which plot should be shown, options: weights, lik, scores, all
`...`	Additional paramters for plot

Value

Plot of the weights, likelihood, ranks, or all three

Author(s)

Fabian Schmich

Examples

data(simulation)
plot(simulation$model)
data(simulation)
plot(simulation$model)

Retrieve ranked prioritisation list

Description

Retrieve ranked prioritisation list

Usage

ranks(object)

## S4 method for signature 'netprioR'
ranks(object)
ranks(object)

## S4 method for signature 'netprioR'
ranks(object)

Arguments

object

A netprioR object

Value

Ranked list of prioritised genes

Author(s)

Fabian Schmich

Examples

data(simulation)
ranks(simulation$model)
data(simulation)
ranks(simulation$model)

Compute ROC curve from netprioR model and true labels

Description

Compute ROC curve from netprioR model and true labels

Usage

ROC(object, ...)

## S4 method for signature 'netprioR'
ROC(object, true.labels, plot = FALSE, ...)
ROC(object, ...)

## S4 method for signature 'netprioR'
ROC(object, true.labels, plot = FALSE, ...)

Arguments

`object`	A `netprioR` object
`...`	Additional arguments
`true.labels`	True full set of underlying labels
`plot`	Flag whether to plot the AUC curve

Value

ROC curve with AUC

Author(s)

Fabian Schmich

Examples

data(simulation)
ROC(simulation$model, true.labels = simulation$labels.true)
data(simulation)
ROC(simulation$model, true.labels = simulation$labels.true)

Simulate labels

Description

Simulate labels

Usage

simulate_labels(values, sizes, nobs)
simulate_labels(values, sizes, nobs)

Arguments

`values`	Vector of labels for groups
`sizes`	Vector of group sizes
`nobs`	Vector of number of observed labels per group

Value

List of Y, Yobs and indices for labeled instances

Author(s)

Fabian Schmich

Examples

labels <- simulate_labels(values = c("Positive", "Negative"), 
sizes = c(10, 10), 
nobs = c(5, 5))
labels <- simulate_labels(values = c("Positive", "Negative"), 
sizes = c(10, 10), 
nobs = c(5, 5))

Simulate random networks with predefined number of members for each of the two groups and the number of neighbours for each node

Description

Simulate random networks with predefined number of members for each of the two groups and the number of neighbours for each node

Usage

simulate_network_random(nmemb, nnei = 1)
simulate_network_random(nmemb, nnei = 1)

Arguments

`nmemb`	Vector of number of members for each group
`nnei`	Number of neighbours for each node

Value

Adjacency matrix of graph

Author(s)

Fabian Schmich

Examples

network <- simulate_network_random(nmemb = c(10, 10), nnei = 1)
network <- simulate_network_random(nmemb = c(10, 10), nnei = 1)

Simulate scalefree networks

Description

Simulate scale free networks for predefined number of members for each of two groups and a parameter pclus that determines how strictly distinct the groups are

Usage

simulate_network_scalefree(nmemb, pclus = 1)
simulate_network_scalefree(nmemb, pclus = 1)

Arguments

`nmemb`	Vector of numbers of members per group
`pclus`	Scalar in [0, 1] determining how strictly distinct groups are

Value

Adjacency matrix

Author(s)

Fabian Schmich

Examples

network <- simulate_network_scalefree(nmemb = c(10, 10), pclus = 0.8)
network <- simulate_network_scalefree(nmemb = c(10, 10), pclus = 0.8)

Simulate phenotypes correlated to labels pivoted into two groups

Description

Simulate phenotypes correlated to labels pivoted into two groups

Usage

simulate_phenotype(labels.true, meandiff, sd)
simulate_phenotype(labels.true, meandiff, sd)

Arguments

`labels.true`	Vector of labels
`meandiff`	difference of means between positive and negative groups
`sd`	Standard deviation of the phenotype

Value

Simulated phenotype

Author(s)

Fabian Schmich

Examples

data(simulation)
phenotypes <- simulate_phenotype(labels.true = simulation$labels.true, meandiff = 0.5, sd = 1)
data(simulation)
phenotypes <- simulate_phenotype(labels.true = simulation$labels.true, meandiff = 0.5, sd = 1)

Example data: Simulated networks, phenotypes and labels for N = 1000 genes

Description

The data set contains simulated data for N = 1000 genes and P = 1 (univariate) phenotypes. The list of networks contains 2 low noise networks and two high noise networks. The class labels are "Positive" and "Negative".

Usage

data(simulation)
data(simulation)

Details

The code used to simluate the data can be found in system.file("example", "data_simulation.R", package = "netprioR")

Value

List of simulated networks, phenotypes and labels for 1000 genes

Retrieve network weights

Description

Retrieve network weights

Usage

weights(object, ...)

## S4 method for signature 'netprioR'
weights(object)
weights(object, ...)

## S4 method for signature 'netprioR'
weights(object)

Arguments

`object`	A `netprioR` object
`...`	Additional arguments

Value

Estimated network weights

Author(s)

Fabian Schmich

Examples

data(simulation)
weights(simulation$model)
data(simulation)
weights(simulation$model)

`x`	Adjacency matrix
`norm`	Type of normalisation

Package 'netprioR'

Help Index