added scripts for generating plots found in paper

+ minor bug fix in plot.synergyscreen

Author: ltronneb

R/plot.synergyscreen.R

@@ -111,6 +111,7 @@ plot.synergyscreen <- function(x, groupbyExperimentID = T, save_plots = FALSE, p
   lower_breaks[1] = -l
   upper_breaks[4] = l
   breaks = as.numeric(c(lower_breaks,upper_breaks))
+  breaks = sort(jitter(breaks))
   Delta_col_palette <- inlmisc::GetColors(scheme = "BuRd")
 
 
@@ -216,7 +217,7 @@ plot.synergyscreen <- function(x, groupbyExperimentID = T, save_plots = FALSE, p
                       breaks = round(breaks,2),
                       values = scales::rescale(breaks),
                       limits = round(max(abs(synscores3$mean_int))*c(-1,1),2),
-                      guide = guide_coloursteps(even.steps = F,
+                      guide = guide_coloursteps(even.steps = T,
                                                 show.limits = T,
                                                 title = "Interaction",
                                                 barheight = unit(3.3, "in"))) + 
@@ -312,6 +313,7 @@ plot.synergyscreen <- function(x, groupbyExperimentID = T, save_plots = FALSE, p
         lower_breaks[1] = -l
         upper_breaks[4] = l
         breaks = as.numeric(c(lower_breaks,upper_breaks))
+        breaks = sort(jitter(breaks))
 
 
         # Center
@@ -368,7 +370,7 @@ plot.synergyscreen <- function(x, groupbyExperimentID = T, save_plots = FALSE, p
                             breaks = round(breaks,2),
                             values = scales::rescale(breaks),
                             limits = round(max(abs(synscores3$mean_int))*c(-1,1),2),
-                            guide = guide_coloursteps(even.steps = F,
+                            guide = guide_coloursteps(even.steps = T,
                                                       show.limits = T,
                                                       title = "Interaction",
                                                       barheight = unit(3.3, "in"))) +

doc/paperFigures/generate.R

@@ -0,0 +1,63 @@
+###### Script for generating plots for the bayesynergy paper
+
+# The dataset is from [1] and has been preprocessed according to the function preprocess.R
+
+# For Figure 4, we include a subset of this dataset, consisting of 3498 experiments
+# on six breast cancer cell lines. This dataset has been run through the synergyscreen
+# function on default settings, and the file "screen.RData" contains the resulting object.
+
+# For Figure 5 and Table 1, the data must be generated according to the instructions in 
+# preprocess.R
+
+########################################################################################
+############################    FIG 4   ################################################
+###### Drug combination screen on breast cancer cell line     ##########################
+########################################################################################
+library(bayesynergy)
+
+# Data for figure 4 is included in the folder
+load("screen.RData")
+
+plot(fit_screen,save_plots=T,plotdevice="png", width=7,height=7,units="in",res=600, 
+     groupbyExperimentID=F)
+
+
+########################################################################################
+############################    FIG 5   ################################################
+###### Gemcitabine + MK-8776 on the OCUBM cell line     ################################
+########################################################################################
+# The file "OCUBM : Gemcitabine + MK-8776.csv" can be generated by following instructions 
+# in preprocess.R
+
+data = read.table("OCUBM : Gemcitabine + MK-8776.csv",sep=",",header=T)
+
+x = as.matrix(data[,c(4,6)])
+y = data$viability
+drug_names = c(data$drugA_name[1],data$drugB_name[1])
+experiment_ID = data$cell_line[1]
+units = c("μM","μM")
+
+fit = bayesynergy(y=y, x=x, drug_names=drug_names, experiment_ID = experiment_ID, units=units)
+
+plot(fit,save_plots=T, plotdevice="png", width=7,height=7,units="in",res=600, plot3D=F)
+
+
+
+########################################################################################
+############################    Table 1   ##############################################
+###### Gemcitabine + MK-8776 on the OCUBM cell line     ################################
+########################################################################################
+summary(fit)
+
+
+
+
+
+# References:
+# [1]
+# Jennifer O'Neil, Yair Benita, Igor Feldman, Melissa Chenard, Brian Roberts, Yaping Liu, Jing Li, Astrid Kral, Serguei Lejnine, Andrey Loboda, William Arthur, Razvan Cristescu, Brian B. Haines, Christopher Winter, Theresa Zhang, Andrew Bloecher and Stuart D. Shumway
+# Mol Cancer Ther June 1 2016 (15) (6) 1155-1162; DOI: 10.1158/1535-7163.MCT-15-0843
+
+
+
+

doc/paperFigures/preprocess.R

@@ -0,0 +1,62 @@
+##### Script for pre-processing the dataset from [1]
+# Used for generating data for Figure 5 and Table 1
+
+# Input: 
+#         156849_1_supp_0_w2lh45.xlsx (single agent viabilities)
+#         156849_1_supp_1_w2lrww.xls  (combination viabilities) 
+#         N.B. This file needs to be opened and resaved as xlsx
+#         Both of these available as supplementary material for [1]
+
+# Output:
+#         OCUBM : Gemcitabine + MK-8776.csv
+#         .csv file containing the single example experiment used for Figure 5 and Table 1
+
+# set path to the location of these files
+
+#path = "path to raw input files"
+
+# Use dplyr and tidyr for preprocessing, readxl for input
+library(tidyr)
+library(dplyr)
+library(readxl)
+
+# Read in files
+mono = read_excel(paste0(path,"/156849_1_supp_0_w2lh45.xlsx"))
+combo = read_excel(paste0(path,"/156849_1_supp_1_w2lrww.xlsx"))
+
+# Select only OCUBM cell line and the drugs c("Gemcitabine","MK-8776")
+drugs =  c("Gemcitabine","MK-8776")
+cellLine = "OCUBM"
+
+mono = mono %>% filter(`cell_line` == cellLine) %>% filter(`drug_name` %in% drugs)
+combo = combo %>% filter(`cell_line` == cellLine) %>% filter(`drugA_name` == drugs[1] & `drugB_name` == drugs[2] |
+                                                             `drugA_name` == drugs[2] & `drugB_name` == drugs[1])
+
+# Now work everything into the correct format, viabilities are numeric
+mono = mono %>% mutate(across(.cols=starts_with("viability"),as.numeric))
+combo = combo %>% mutate(across(.cols=starts_with("viability"),as.numeric))
+
+# Convert to long format and remove some columns
+mono  = pivot_longer(mono,cols=starts_with("viability"),values_to = "viability") %>% 
+  select(cell_line, drug_name, `Drug_concentration (µM)`,viability)
+combo = pivot_longer(combo, cols=starts_with("viability"),values_to="viability") %>% 
+  select("cell_line", "drugA_name", "drugA Conc (µM)", "drugB_name","drugB Conc (µM)","viability")
+
+# Now get mono into the same format as the combo
+mono = mono %>% mutate("drugA_name" = combo$drugA_name[1],
+                "drugB_name" = combo$drugB_name[1]) %>%
+  mutate(`drugA Conc (µM)` = ifelse(drug_name == drugA_name,`Drug_concentration (µM)`,0)) %>%
+  mutate(`drugB Conc (µM)` = ifelse(drug_name == drugB_name,`Drug_concentration (µM)`,0)) %>%
+  select("cell_line", "drugA_name", "drugA Conc (µM)", "drugB_name","drugB Conc (µM)","viability")
+# Combine the two sets
+final = bind_rows(mono,combo)
+
+
+# Write csv
+final %>% write.csv(paste0(path,"/OCUBM : Gemcitabine + MK-8776.csv"))
+
+
+# References:
+# [1]
+# Jennifer O'Neil, Yair Benita, Igor Feldman, Melissa Chenard, Brian Roberts, Yaping Liu, Jing Li, Astrid Kral, Serguei Lejnine, Andrey Loboda, William Arthur, Razvan Cristescu, Brian B. Haines, Christopher Winter, Theresa Zhang, Andrew Bloecher and Stuart D. Shumway
+# Mol Cancer Ther June 1 2016 (15) (6) 1155-1162; DOI: 10.1158/1535-7163.MCT-15-0843