Drug Response Estimation from single-cell Expression Profiles (DREEP)

Description

DREEP (Drug Response Estimation from single-cell Expression Profiles) published BMC medicine and available at Pellecchia, Viscido et al. 2023.

DREEP is a bioinformatics tool that leverages results from large-scale cell-line viability screens and enrichment analysis to predict drug vulnerability from the transcriptional state of a cell. It only requires a pre-defined collection of Genomic Profiles of Drug Sensitivity (GPDS) signatures that are ranked lists of genes reflecting their importance in predicting the effect of a small molecule.

Installation

DREEP require the installation of gficf v2 package first. Follow the instruction below to install it on Ubuntu. For other OS althoug not officially supported you can give a look here

1. Installation of GFICF on Ubuntu/Debian

You need gsl dev library to successfully install RcppGSL library. On Ubuntu/Debian systems this can be accomplished by running the following command sudo apt-get install libgsl-dev libcurl4-openssl-dev libssl-dev libxml2-dev from the terminal.

Then exec in R terminal the following commands to install gficf

# Install required bioconductor packages
if (!requireNamespace("BiocManager", quietly = TRUE)) {install.packages("BiocManager")}

BiocManager::install(setdiff(c("sva","edgeR", "fgsea"),rownames(installed.packages())),update = F)

# We rquire RcppML package from github (not the cran version)
if("RcppML" %in% rownames(installed.packages())) {remove.packages("RcppML")}
devtools::install_github("zdebruine/RcppML")

if(!require(devtools)){ install.packages("devtools")}
devtools::install_github("gambalab/gficf")

2. Installation of DREEP

From the R terminal exec the following commands

if(!require(devtools)){ install.packages("devtools")}
devtools::install_github("gambalab/DREEP")

3. Example of use

library(gficf)
library(DREEP)
library(ggplot2)

data(small_BC_atlas)
data <- gficf(M=small_BC_atlas,verbose = T)

# Run DREEP on all the cell of the atlas using only CTRP2 and GDSC drug datasets
dreep.data <- DREEP::runDREEP(M = data$gficf,
                              n.markers = 500,
                              gsea = "simple",
                              gpds.signatures = c("CTRP2","GDSC"))

# DREEP predictions are into dreep.data$df data frame
# Each row is a drug and the column sens contains
# the percentage of cells predicted sensitive to the drug.
# Drugs in this data frame are sorted according to the median ES of the drug on the cell population.
# This is the best measure to determine if your cells are sensible or not to a drug.
# So please use these values instead of the percentage of cells.
head(dreep.data$df)


# Run cell reduction using drug response profile estimated from DREEP
dreep.data <- DREEP::runDrugReduction(dreep.data,
                                      verbose = T,
                                      cellDistAbsolute = F,
                                      reduction = "umap",
                                      storeCellDist = T)

# UMAP coordinate are stored into dreep.data$embedding data frame
head(dreep.data$embedding)

# Let's add cell line names and plot the recontructed cell embedding space
dreep.data$embedding$ccl <- sapply(strsplit(x = rownames(dreep.data$es.mtx),split = "_",fixed = T),function(x) x[1])
ggplot(data = dreep.data$embedding,aes(x=X,y=Y,color=ccl)) + geom_point(size=.5) + theme_bw() + xlab("UMAP 1") + ylab("UMAP 2")

# We can cluster cells using the respone profiles estimated by DREEP
dreep.data <- DREEP::clusterCells(dreep.data,verbose = T,resolution = 0.01)
ggplot(data = dreep.data$embedding,aes(x=X,y=Y,color=cluster)) + geom_point(size=.5) + theme_bw() + xlab("UMAP 1") + ylab("UMAP 2")