condiGRN: Build conditional GRN

In SCBIT-YYLab/DysRegSig: Gene dysregulation analysis and construction of mechanistic signature in cancer

Description

Build conditional gene regulatory network (GRN) with expression data and prior reference network by using feature selection algorithm.

Usage

condiGRN(exp.data, tf2tar, 
         method = 'Boruta', pValue = 0.01, 
         threshold = NULL, verbose = TRUE, ...)

Arguments

exp.data	Expression matrix. Columns correspond to genes, rows correspond to experiments. The matrix is expected to be already normalized.
tf2tar	The prior reference GRN containing TF-target relationships.
method	The method used to build conditional GRN, such as 'Boruta', or 'RGBM'.
pValue	Confidence level used in Boruta. Default value should be used.
threshold	The threshould for weight in RGBM.
verbose	A logical value indicating whether display the computating progress.
...	Other parameters passed to Boruta or RGBM.

Details

While using method Boruta, the predifined pValue could be used as the threshold to filter out nonsignificant regulatory relationships. While using method RGBM, users need to explore the threshould of weight based on the output of RGBM to filter out nonsignificant regulatory relationships before following analysis.

Value

Conditional GRN.

References

Kursa M B, Rudnicki W R. Feature Selection with the Boruta Package. Journal of Statistical Software. 2010, 36(11): 13.

Mall R, Cerulo L, Garofano L, et al. RGBM: regularized gradient boosting machines for identification of the transcriptional regulators of discrete glioma subtypes. Nucleic Acids Res. 2018, 46 (7), e39.

Examples

# Build a conditional GRN based on a reference GRN.
data(ExpData)
ExpData[1:5,1:5]

data(tf2tar)
head(tf2tar)

data(ClinData)

group.1 <- ClinData$sample[which(ClinData$binaryResponse == 'CR/PR')]
exp.1 <- ExpData[,colnames(ExpData) %in% group.1]

## using method Boruta
net.1 <- condiGRN(exp.data = exp.1, tf2tar = tf2tar, method = 'Boruta', pValue = 0.01)

SCBIT-YYLab/DysRegSig documentation built on July 19, 2021, 4:38 a.m.