data_examples: Example data under simple and complex models

In audreyqyfu/mrpc: PC Algorithm with the Principle of Mendelian Randomization

Example data under simple and complex models

Description

Example data under the simple and complex graphs. Data may be continuous or discrete.

Usage

data(data_examples)

Details

For each model, the graph and a simulated data matrix are available for both continuous and discrete data.

For continuous data with genetic information: 1000 samples in row and 6 variables in column. First two columns are the genetic variants and remaning columns are gene expression.

Continuous data without genetic information: 1000 samples in row and 8 variables in column.

Discrete data with genetic information: 1000 samples in row and 6 variables in column. First column is the genetic variant and remaning columns are the gene expression.

Discrete data without genetic information: 1000 samples in row and 5 variables in column.

Continuous data with genetic information for complex model: 1000 samples in row and 22 variables in column. First 14 column is the genetic variants and remaning columns are the genes expression.

Value

A list that containing the numeric data matrix and components of a graph.

• simple: Simple model.

• complex: Complex model.

• cont: Continuous.

• disc: Discrete.

• withGV: With genetic information.

• withoutGV: Without genetic information.

• data: Data matrix.

• graph: Components of a graph.

Author(s)

Md Bahadur Badsha ([e-mail])

Examples

## Not run: 
# Continuous data with genetic varitant (GV)
# load the data
data("data_examples")  
data <- data_examples$simple$cont$withGV$data
# Extract the sample size
n <- nrow(data)  

# Extract the node/column names
V <- colnames(data)  

# Calculate Pearson correlation
suffStat_C <- list(C = cor(data),
                   n = n)

# Infer the graph by MRPC
data.mrpc.cont.withGV <- MRPC(data = data,
                              suffStat = suffStat_C, 
                              GV = 2,
                              FDR = 0.05, 
                              indepTest = 'gaussCItest',
                              labels = V,
                              FDRcontrol = 'LOND',
                              verbose = FALSE)
# Plot the results
par(mfrow = c(1, 2))
# plot the true graph
plot(data_examples$simple$cont$withGV$graph,
     main = "truth") 
# plot the inferred graph
plot(data.mrpc.cont.withGV,
     main = "inferred")    


# Continuous data without genetic information
# load the data
data("data_examples")    
data <- data_examples$simple$cont$withoutGV$data
  
# Extract the sample size
n <- nrow(data) 

# Extract the node/column names
V <- colnames(data)

# Calculate Pearson correlation
suffStat_C <- list(C = cor(data),
                   n = n)

# Infer the graph by MRPC
data.mrpc.cont.withoutGV <- MRPC(data = data,
                                 suffStat = suffStat_C, 
                                 GV = 0,
                                 FDR = 0.05,
                                 indepTest = 'gaussCItest',
                                 labels = V, 
                                 FDRcontrol = 'LOND',
                                 verbose = FALSE)
# Plot the results
par(mfrow = c(1, 2))
# plot the true graph
plot(data_examples$simple$cont$withoutGV$graph,
     main = "truth") 
# plot the inferred graph
plot(data.mrpc.cont.withoutGV,
     main = "inferred") 

# Discrete data with genetic information
# load the data
data("data_examples")     
data <- data_examples$simple$disc$withGV$data
  
# Extract the sample size
n <- nrow(data)  

# Extract the node/column names
V <- colnames(data)

suffStat_C <- list (dm = data, adaptDF = FALSE, n.min = 1000)

# Infer the graph by MRPC
data.mrpc.disc.withGV <- MRPC(data = data,
                              suffStat = suffStat_C, 
                              GV = 1,
                              FDR = 0.05, 
                              indepTest = 'disCItest',
                              labels = V, 
                              FDRcontrol = 'LOND',
                              verbose = FALSE)
# Plot the results
par (mfrow = c(1, 2))
# plot the true graph
plot(data_examples$simple$disc$withGV$graph,
     main = "truth") 
# Plot the inferred causal graph
plot(data.mrpc.disc.withGV,
     main = "inferred") 

# Discrete data without genetic information
# load the data
data("data_examples")    
data <- data_examples$simple$disc$withoutGV$data
  
# Extract the sample size
n <- nrow (data)  

# Extract the node/column names
V <- colnames(data)   

suffStat_C <- list (dm = data, adaptDF = FALSE, n.min = 1000)

# Infer the graph by MRPC
data.mrpc.disc.withoutGV <- MRPC(data = data,
                                 suffStat = suffStat_C, 
                                 GV = 1,
                                 FDR = 0.05, 
                                 indepTest = 'disCItest',
                                 labels = V, 
                                 FDRcontrol = 'LOND',
                                 verbose = FALSE)
# Plot the results
par(mfrow = c(1, 2))
# plot the true graph
plot(data_examples$simple$disc$withoutGV$graph,
     main = "truth") 
# plot the inferred graph
plot(data.mrpc.disc.withoutGV,
     main = "inferred") 

# Continuous data with genetic information for complex model
# load the data
data("data_examples")   

# Graph without clustering
plot(data_examples$complex$cont$withGV$graph) 

# Adjacency matrix from directed example graph
Adj_directed <- as(data_examples$complex$cont$withGV$graph,
                   "matrix")

# Plot of dendrogram with modules colors of nodes
PlotDendrogramObj <- PlotDendrogram(Adj_directed,
                                    minModuleSize = 5)
                  
# Visualization of inferred graph with modules colors
PlotGraphWithModulesObj <- PlotGraphWithModules(Adj_directed,
                                                PlotDendrogramObj,
                                                GV = 14,
                                                node.size = 8,
                                                arrow.size = 5,
                                                label.size = 3,
                                                alpha = 1) 
# plot 
plot(PlotGraphWithModulesObj)

# Run MRPC on the complex data set with ADDIS as the FDR control method.
data <- data_examples$complex$cont$withGV$data
n <- nrow (data)        # Number of rows
V <- colnames(data)     # Column names

# Calculate Pearson correlation
suffStat_C <- list(C = cor(data),
                   n = n)

# Infer the graph by MRPC
MRPC.addis <- MRPC(data,
                   suffStat = suffStat_C,
                   GV = 14,
                   FDR = 0.05,
                   indepTest = 'gaussCItest',
                   labels = V,
                   FDRcontrol = 'ADDIS',
                   tau = 0.5,
                   lambda = 0.25,
                   verbose = FALSE)
                   
# Plot the true and inferred graphs.
par(mfrow = c(1, 2))
plot(data_examples$complex$cont$withGV$graph,
     main = 'True graph')
plot(MRPC.addis,
     main = 'Inferred graph')

## End(Not run)
  

audreyqyfu/mrpc documentation built on April 17, 2022, 7:35 a.m.