DIGREscore: Drug Induced Genomic Residual Effect (DIGRE) model

In Minzhe/DIGREsyn: Drug Induced Genomic Residual Effect (DIGRE) model for predicting drug pair synergistic score

Description

This function estimate the compound pair synergism/antagonism. It takes drug treated gene expression data and drug dose response curve data as input, and caculate the synergistic score using Drug Induced Genomic Residual Effect (DIGRE) model.

Usage

DIGREscore(geneExpDiff, doseRes, pathway = "KEGG", geneNet, fold = 0.6)

Arguments

geneExpDiff	a matrix of drug treated gene expression data. Each column represent one drug, each row represent one gene. The value represent the fold change of the expression of a particular gene after drug treated compared to negative control.
doseRes	a matrix contains drug dose response data. Each column represent one drug, two rows of two different drug dose response curve. See 'doseRes.demo' for example.
pathway	pathway information used in DIGRE model. User would specify either "KEGG" to use the KEGG pathway information or "GeneNet" to use the gene network information.
geneNet	optional parameter. If pathway parameter is "GeneNet", then specify this parameter to use your own gene network data. If pathway parameter is "KEGG", then do not set this parameter.
fold	a value between 0 and 1. Gene expression fold change above this value would be considered as upregulated, below the opposite number would be considered as downregulated, otherwise would be considered as no effect. The default value is 0.6.

Details

This function takes drug treated gene expression data, dose-response curve data and pathway information as inputs, and calculate pair synergistic score of all the possible combination of the compound you provided, and their rank from the most synergistic to the most antagonist. Larger score indicates high possibility of the pair to have synergistic effect, and vice versa. (Notice that this algorithm focus more on predicting the relative rank of your compound pairs not the exact synergistic strength. If you want to do that, maybe you should involve positive control in your experiment. And also the score calculated by two pathway information is not comparable.)

Value

a list contains two matrices. One gives the drug pair synergistic score and their rank, the other contains the raw data to calculate the score.

Author(s)

Jichen Yang, Sangin Lee, Minzhe Zhang([e-mail])

See Also

Bansal M, Yang J, Karan C, et al. A community computational challenge to predict the activity of pairs of compounds[J]. Nature biotechnology, 2014, 32(12): 1213-1222.

Yang J, Tang H, Li Y, et al. DIGRE: Drug Induced Genomic Residual Effect Model for Successful Prediction of Multidrug Effects[J]. CPT: pharmacometrics & systems pharmacology, 2015, 4(2): 91-97.

Examples

### profile gene expression data
geneExpDiff <- profileGeneExp(geneExp = geneExp.demo)
### DIGRE prediction
res.KEGG <- DIGREscore(geneExpDiff = geneExpDiff, doseRes = doseRes.demo, pathway = "KEGG", fold = 0.6) # KEGG pathway 
res.geneNet <- DIGREscore(geneExpDiff = geneExpDiff, doseRes = doseRes.demo, pathway = "GeneNet", geneNet = geneNetLymph.mat) # Gene network

Minzhe/DIGREsyn documentation built on May 7, 2019, 4:58 p.m.