rogge_12231_data_compilation.R 73 KB
Newer Older
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
1
2
3
4
5
6
7
8
9
10
11
12

################################ Initialization

erase.objects <- TRUE # write TRUE to erase all the existing objects in R before starting the algorithm and FALSE otherwise. Beginners should use TRUE
if(erase.objects == TRUE){
    rm(list=ls())
    erase.objects = TRUE
}
erase.graphs <- TRUE # write TRUE to erase all the graphic windows in R before starting the algorithm and FALSE otherwise

################################ End Initialization

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
13
sink(stdout(), type = "message") # diverts R output to a connection
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
14
script <- commandArgs(trailingOnly = FALSE)[4]  # recover script name, e.g., r_341_conf $check_lod_gael_conf. 1) .exe R path, 2) --slave, 3) --no-restore, 4) --file and 5) --args
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
15
args <- commandArgs(trailingOnly = TRUE)  # recover arguments written after the call of the Rscript, ie after r_341_conf $check_lod_gael_conf 
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
16
tempo.arg.names <- c("path.lib", "path.in", "path.out", "path.function1", "project.name", "label.size", "optional.text", "slurm.loop.nb", "analysis.kind") # objects names exactly in the same order as in the bash code and recovered in args
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
if(length(args) != length(tempo.arg.names)){
    tempo.cat <- paste0("\n\n================\n\nERROR: THE NUMBER OF ELEMENTS IN args (", length(args),") IS DIFFERENT FROM THE NUMBER OF ELEMENTS IN tempo.arg.names (", length(tempo.arg.names),")\nargs:", paste0(args, collapse = ","), "\ntempo.arg.names:", paste0(tempo.arg.names, collapse = ","), "\n\n================\n\n")
    stop(tempo.cat)
}
for(i in 1:length(tempo.arg.names)){
    assign(tempo.arg.names[i], args[i])
    if(is.character(get(tempo.arg.names[i])) == TRUE & grepl(x = get(tempo.arg.names[i]), pattern = "^[1234567890]+$")){  # convert into numeric
        assign(tempo.arg.names[i], as.integer(get(tempo.arg.names[i])))
    }else if(is.character(get(tempo.arg.names[i])) == TRUE & grepl(x = get(tempo.arg.names[i]), pattern = "^[1234567890.+e-]+$")){
        assign(tempo.arg.names[i], as.numeric(get(tempo.arg.names[i])))
    }else if(is.character(get(tempo.arg.names[i])) == TRUE & grepl(x = get(tempo.arg.names[i]), pattern = "^(TRUE|FALSE)$")){  # convert into logical
        assign(tempo.arg.names[i], as.logical(get(tempo.arg.names[i])))
    }
}

################################ Recording of the initial parameters
param.list <- c(
    "script", 
    "args",
    "tempo.arg.names",
    tempo.arg.names
)
if(any(duplicated(param.list))){
    stop(paste0("\n\n================\n\nERROR: THE param.list OBJECT CONTAINS DUPLICATED ELEMENTS\n\n================\n\n")) # message for developers
}
char.length <- nchar(param.list)
space.add <- max(char.length) - char.length + 5
param.ini.settings <- NULL
for(i in 1:length(param.list)){
    param.ini.settings <- c(param.ini.settings, paste0("\n", param.list[i], paste0(rep(" ", space.add[i]), collapse = ""), paste0(get(param.list[i]), collapse = ",")))
}
################################ End Recording of the initial parameters

################################ DEBUG


debug2 <- '
rm(list = ls())
erase.objects <- TRUE
erase.graphs <- TRUE
script <- "code ini v1.0.0"
project.name <-"rogge12231"
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
59
path.lib <- "none" # absolute path of the library folder. Write "none" if not required
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
60
path.in <- "Z:/rogge12231/rogge_12231_1550514780/" # absolute path of the data folder
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
61
path.out <- "C:/Users/Gael/Desktop/" # absolute path of the output folder
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
62
63
path.function1 <- "C:/Users/Gael/Documents/Git_versions_to_use/cute_little_R_functions-v4.5.0/cute_little_R_functions.R" # Define the absolute pathway of the folder containing functions created by Gael Millot
project.name <- "rogge_project"
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
64
label.size <-6
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
65
optional.text <- ""
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
66
slurm.loop.nb <- 100
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
67
analysis.kind <- "valid_boot"
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
68
69
70

activate.pdf <- FALSE # graph file parameter
cut.off.freq.for.selected.genes <- 0.01 # graph file parameter
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
71
'
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
72
# eval(parse(text = debug2)) ; cat(paste0("\n\n================\n\nERROR: ACTIVE DEBUG VALUES\n\n================\n\n")) ; stop()
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
73
74
75

# data.frame(PARAM = tempo.arg.names, ARG = args) # for debug mode

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
76
activate.pdf <- TRUE ; cat(paste0("\n\n================\n\nERROR: ACTIVE GRAPH FILE PARAMETERS\n\n================\n\n")) # graph file parameter
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
77
78
79



Gael  MILLOT's avatar
interm  
Gael MILLOT committed
80
81
82
################################ End DEBUG

################################ Packages verification and import
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
83

Gael  MILLOT's avatar
interm  
Gael MILLOT committed
84
85
# packages are imported even if functions are used using package.name::function()
req.package.list <- c(
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
86
    "plyr",
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
87
88
    "pheatmap",
    "corrplot",
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
89
    "ggbeeswarm",
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
90
91
    "ggplot2",
    "gridExtra",
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
92
93
    "lubridate",
    "RCurl" # for url.exists
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
94
)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
95
if(all(path.lib == "none")){
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
96
    path.lib <- .libPaths() #
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
97
}else{
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
98
99
    # .libPaths(new = ) add path to default path
  .libPaths(new = sub(x = path.lib, pattern = "/$|\\\\$", replacement = "")) # .libPaths() does not support / at the end of a submitted path. Thus check and replace last / or \\ in path
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
100
101
102
103
104
105
106
107
}
for(i0 in 1:length(req.package.list)){
    if( ! req.package.list[i0] %in% rownames(installed.packages(lib.loc =  path.lib))){
        stop(paste0("\n\n================\n\nERROR: PACKAGE ", req.package.list[i0], " MUST BE INSTALLED IN THE MENTIONNED DIRECTORY:\n", paste(path.lib, collapse = "\n"), "\n CHECK ALSO IN :\n", paste(.libPaths(), collapse = "\n"), "\n\n================\n\n"))
    }else{
        suppressPackageStartupMessages(library(req.package.list[i0], lib.loc = path.lib, quietly = TRUE, character.only = TRUE))
    }
}
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
108

Gael  MILLOT's avatar
interm  
Gael MILLOT committed
109
110
111
################################ End Packages verification and import

################################ Functions
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
112
113
114

if(length(path.function1) != 1){
    stop(paste0("\n\n============\n\nERROR: path.function1 PARAMETER MUST BE LENGTH 1: ", paste(path.function1, collapse = " "), "\n\n============\n\n"))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
115
116
117
118
119
120
121
122
123
124
125
126
}else if(grepl(x = path.function1, pattern = "^http")){
    if( ! RCurl::url.exists(path.function1)){
        stop(paste0("\n\n============\n\nERROR: HTTP INDICATED IN THE path.function1 PARAMETER DOES NOT EXISTS: ", path.function1, "\n\n============\n\n"))
    }else{
        source(path.function1) # source the fun_ functions used below
    }
}else if( ! grepl(x = path.function1, pattern = "^http")){
    if( ! file.exists(path.function1)){
        stop(paste0("\n\n============\n\nERROR: FILE INDICATED IN THE path.function1 PARAMETER DOES NOT EXISTS: ", path.function1, "\n\n============\n\n"))
    }else{
        source(path.function1) # source the fun_ functions used below
    }
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
127
}
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
128

Gael  MILLOT's avatar
interm  
Gael MILLOT committed
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
################################ End Functions

################################ Main code

################ Pre-ignition checking

arg.check <- NULL # for function debbuging
checked.arg.names <- NULL # for function debbuging
ee <- expression(arg.check <- c(arg.check, tempo$problem) , checked.arg.names <- c(checked.arg.names, tempo$param.name))

# initializations
tempo <- fun_param_check(data = erase.objects, class = "logical", length = 1) ; eval(ee)
tempo <- fun_param_check(data = erase.graphs, class = "logical", length = 1) ; eval(ee)
tempo <- fun_param_check(data = script, class = "character", length = 1) ; eval(ee)
tempo <- fun_param_check(data = args, class = "character", length = length(tempo.arg.names)) ; eval(ee)
tempo <- fun_param_check(data = tempo.arg.names, class = "character", length = length(tempo.arg.names)) ; eval(ee)

# imported objects
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
147
148
149
150
tempo <- fun_param_check(data = path.lib, class = "character") ; eval(ee)
if(tempo$problem == FALSE & ! all(path.lib == "none")){
    if( ! all(dir.exists(path.lib))){
        cat(paste0("\n\n============\n\nERROR: DIRECTORY PATH INDICATED IN THE path.lib PARAMETER DOES NOT EXISTS:\n", paste(path.lib, collapse = "\n"), "\n\n============\n\n"))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
151
        arg.check <- TRUE
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
152
153
154
155
156
    }
}
tempo <- fun_param_check(data = path.in, class = "character", length = 1) ; eval(ee)
if(tempo$problem == FALSE & ! dir.exists(path.in)){
    cat(paste0("\n\n============\n\nERROR: DIRECTORY PATH INDICATED IN THE path.in PARAMETER DOES NOT EXISTS: ", path.in, "\n\n============\n\n"))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
157
    arg.check <- TRUE
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
158
}
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
159
160
161
if(tempo$problem == FALSE & ! ("loop1_res_data.RData" %in% list.files(paste0(path.in, "loop1/")))){
    cat(paste0("\n\n============\n\nERROR: loop1_res_data.RData SHOULD BE PRESENT IN ", paste0(path.in, "loop1/"), ":\nCHECK THE path.in PARAMETER\n\n============\n\n"))
    arg.check <- TRUE
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
162
163
164
165
}
tempo <- fun_param_check(data = path.out, class = "character", length = 1) ; eval(ee)
if(tempo$problem == FALSE & ! dir.exists(path.out)){
    cat(paste0("\n\n============\n\nERROR: DIRECTORY PATH INDICATED IN THE path.out PARAMETER DOES NOT EXISTS: ", path.out, "\n\n============\n\n"))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
166
    arg.check <- TRUE
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
167
}
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
168
169
# path.function1 fully tested above
tempo <- fun_param_check(data = path.function1, class = "character", length = 1) ; eval(ee)
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
170
tempo <- fun_param_check(data = project.name, class = "character", length = 1) ; eval(ee)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
171
tempo <- fun_param_check(data = label.size, typeof = "integer", length = 1, double.as.integer.allowed = TRUE, neg.values = FALSE) ; eval(ee)
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
172
173
174
175
176
177
178
179
180
181
182
tempo <- fun_param_check(data = optional.text, class = "character", length = 1) ; eval(ee)
tempo <- fun_param_check(data = slurm.loop.nb, typeof = "integer", length = 1, double.as.integer.allowed = TRUE, neg.values = FALSE) ; eval(ee)
tempo <- fun_param_check(data = analysis.kind, options = c("longit", "valid_boot", "full_cross_validation"), length = 1) ; eval(ee)
if(any(arg.check) == TRUE){
    stop()
}

################ Ignition

ini.time <- as.numeric(Sys.time()) # time of process begin, converted into seconds
analysis.nb <- trunc(ini.time) # to provide a specific number ot each analysis
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
183
log.file <- paste0("r_", project.name, "_", analysis.nb,"_report.txt")
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
184
185
186
187
188
# name.dir <- paste0(project.name, "_", analysis.nb)
# path.out<-paste0(path.out, name.dir)
# suppressWarnings(dir.create(path.out))
backup.name <- NULL # names of the object to save
fun_export_data(data = paste0("\n\n################################ ", log.file, " ################"), output = log.file, no.overwrite = FALSE, path = path.out, sep = 4)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
189
fun_export_data(path = path.out, data = "################################ INITIAL DATA", output = log.file)
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
190
191
fun_export_data(path = path.out, data = paste0("SCRIPT USED: ", script), output = log.file)
fun_export_data(path = path.out, data = paste0("KIND OF ANALYSIS PERFORMED: ", analysis.kind), output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
192
fun_export_data(path = path.out, data = paste0("NUMBER OF LOOPS USED FOR COMPILATION: ", slurm.loop.nb), output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
193
fun_export_data(path = path.out, data = paste0("FOR INFO: THE RESPONSE USED IS THE COLUMN: response_ASDAS_R_NR"), output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
194

Gael  MILLOT's avatar
interm  
Gael MILLOT committed
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209

################ Graphical parameter ignition

zone.ini <- matrix(1, ncol=1) # to reset the layout if required
if(erase.graphs == TRUE){
    graphics.off()
}else{
    tempo.cat <- paste0("BEWARE: GRAPHICS HAVE NOT BEEN ERASED. GRAPHICAL PARAMETERS MAY HAVE NOT BEEN REINITIALIZED")
    fun_export_data(path = path.out, data = tempo.cat, output = log.file)
}
if(optional.text == "no.txt"){
    optional.text <- ""
}else{
    fun_export_data(path = path.out, data = paste0("OPTIONAL TEXT:\n", optional.text), output = log.file)
}
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
210
par.ini <- fun_open_window(pdf.disp = activate.pdf, path.fun = path.out, pdf.name.file = paste0("final_graphs"), width.fun = 7, height.fun = 7, paper = "special", no.pdf.overwrite = TRUE, return.output = TRUE)$ini.par
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
211
212
213
214
215
if(activate.pdf == TRUE){
    pdf.nb <- dev.cur()
}else{
    tempo <- dev.off()
}
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
216
217
218

################ Data import

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
219
220
for(i0 in 1:slurm.loop.nb){
    load(paste0(path.in, "loop", i0, "/loop", i0, "_res_data.RData"))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
221
222
223
    if(i0 == 1){
        load(paste0(path.in, "loop", i0, "/complete.data.table.RData")) # import df.nano
    }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
}

# > class(loop1_ttab)
# [1] "data.frame"
# > class(mod.gene.names)
# [1] "character"
# > class(confmat1.genes.logreg$result)
# [1] "matrix"
# class(data.pred1.genes.logreg)
# [1] "data.frame"
# class(data.roc1.genes.rpart)
# [1] "data.frame"

# > class(final.gene.list)
# [1] "data.frame"

################ Data compilation

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
242
fun_export_data(path = path.out, data = "################################ COMPILATION", output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
243
244
245
246
247
# ttab
if(analysis.kind == "full_cross_validation"){
    final.ttab <- data.frame(NULL, stringsAsFactors = FALSE)
    for(i0 in 1:slurm.loop.nb){
        if(nrow(get(paste0("loop", i0, "_ttab"))) > 0){
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
248
            final.ttab <- rbind(final.ttab, data.frame(get(paste0("loop", i0, "_ttab")), LOOP_NB = paste0("loop", i0), stringsAsFactors = FALSE))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
249
250
251
252
253
254
255
256
257
258
        }else{
            final.ttab <- rbind(final.ttab, data.frame(logFC = NA, AveExpr = NA, t = NA, P.Value = NA, adj.P.Val = NA, B = NA, LOOP_NB = paste0("loop", i0), stringsAsFactors = FALSE))
        }
    }
}else{
    final.ttab <- get(paste0("loop1_ttab"))
    fun_export_data(path = path.out, data = paste0("final.ttab OBJECT CONTAINS ONLY LOOP1 RESULTS BECAUSE NO LOOP PERFORMED FOR THIS USING ", analysis.kind, " ANALYSIS"), output = log.file)
}
backup.name <- c(backup.name, "final.ttab")

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
# select.gene.curve
if(analysis.kind == "full_cross_validation"){
    final.select.gene.curve <- data.frame(NULL, stringsAsFactors = FALSE)
    for(i0 in 1:slurm.loop.nb){
        if(nrow(get(paste0("loop", i0, "_select.gene.curve"))) > 0){
            final.select.gene.curve <- rbind(final.select.gene.curve, data.frame(get(paste0("loop", i0, "_select.gene.curve")), LOOP_NB = paste0("loop", i0), stringsAsFactors = FALSE))
        }else{
            final.select.gene.curve <- rbind(final.select.gene.curve, data.frame(x = NA, y = NA, PANEL = NA, group = NA, shape = NA, colour = NA, size = NA, fill = NA, alpha = NA, stroke = NA, LOOP_NB = paste0("loop", i0), stringsAsFactors = FALSE))
        }
    }
}else{
    final.select.gene.curve <- get(paste0("loop1_select.gene.curve"))
    fun_export_data(path = path.out, data = paste0("final.select.gene.curve OBJECT CONTAINS ONLY LOOP1 RESULTS BECAUSE NO LOOP PERFORMED FOR THIS USING ", analysis.kind, " ANALYSIS"), output = log.file)
}
backup.name <- c(backup.name, "final.select.gene.curve")

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
275
276
277
278
279
280
281
282
283
284
285
# mod.gene.names
if(analysis.kind == "full_cross_validation"){
    final.mod.gene.names <- data.frame(NULL, stringsAsFactors = FALSE)
    for(i0 in 1:slurm.loop.nb){
        if(length(get(paste0("loop", i0, "_mod.gene.names"))) > 0){
            final.mod.gene.names <- rbind(final.mod.gene.names, data.frame(GENE = get(paste0("loop", i0, "_mod.gene.names")), LOOP_NB = paste0("loop", i0), stringsAsFactors = FALSE))
        }else{
            final.mod.gene.names <- rbind(final.mod.gene.names, data.frame(GENE = NA, LOOP_NB = paste0("loop", i0), stringsAsFactors = FALSE))
        }
    }
}else{
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
286
    final.mod.gene.names <- data.frame(GENE = get(paste0("loop1_mod.gene.names")), LOOP_NB = "loop1", stringsAsFactors = FALSE)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
287
288
289
290
    fun_export_data(path = path.out, data = paste0("final.mod.gene.names OBJECT CONTAINS ONLY LOOP1 RESULTS BECAUSE NO LOOP PERFORMED FOR THIS USING ", analysis.kind, " ANALYSIS"), output = log.file)
}
backup.name <- c(backup.name, "final.mod.gene.names")

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
# gene.importance
if(analysis.kind == "full_cross_validation"){
    final.gene.importance <- data.frame(NULL, stringsAsFactors = FALSE)
    for(i0 in 1:slurm.loop.nb){
        if(nrow(get(paste0("loop", i0, "_gene.importance"))) > 0){
            final.gene.importance <- rbind(final.gene.importance, data.frame(get(paste0("loop", i0, "_gene.importance")), LOOP_NB = paste0("loop", i0), stringsAsFactors = FALSE))
        }else{
            final.gene.importance <- rbind(final.gene.importance, data.frame(features = NA, importance = NA, LOOP_NB = paste0("loop", i0), stringsAsFactors = FALSE))
        }
    }
}else{
    final.gene.importance <- get(paste0("loop1_gene.importance"))
    fun_export_data(path = path.out, data = paste0("final.gene.importance OBJECT CONTAINS ONLY LOOP1 RESULTS BECAUSE NO LOOP PERFORMED FOR THIS USING ", analysis.kind, " ANALYSIS"), output = log.file)
}
backup.name <- c(backup.name, "final.gene.importance")

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
307
308
309
310
311
312
313
314
# loop1_confmat1.genes.rf, data.pred1.genes.logreg, data.roc1.genes.logreg
data.kind <- c("genes", "crp", "genes.crp") # beware, the order is important
ana.kind <- c("rf", "logreg", "rpart")
for(i0 in 1:length(data.kind)){
    for(i1 in 1:length(ana.kind)){
        assign(paste0("final.confmat", i0, ".", data.kind[i0], ".", ana.kind[i1]), data.frame(NULL, stringsAsFactors = FALSE))
        assign(paste0("final.data.pred", i0, ".", data.kind[i0], ".", ana.kind[i1]), data.frame(NULL, stringsAsFactors = FALSE))
        assign(paste0("final.data.roc", i0, ".", data.kind[i0], ".", ana.kind[i1]), data.frame(NULL, stringsAsFactors = FALSE))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
315
        assign(paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1]), data.frame(NULL, stringsAsFactors = FALSE))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
        for(i2 in 1:slurm.loop.nb){
            # matrix confusion
            if(nrow(get(paste0("loop", i2, "_confmat", i0, ".", data.kind[i0], ".", ana.kind[i1]))$result) > 0){
                # transform the table into a 1 line data frame
                tempo1 <- get(paste0("loop", i2, "_confmat", i0, ".", data.kind[i0], ".", ana.kind[i1]))
                tempo2 <- c(c(tempo1$result[1:2, 1:2]), tempo1$task.desc$size)
                tempo3 <- c("TRUE_NR_PRED_NR", "TRUE_R_PRED_NR", "TRUE_NR_PRED_R", "TRUE_R_PRED_R", "SIZE")
                tempo4 <- eval(parse(text = paste0("data.frame(", paste(tempo3, tempo2, sep = "=", collapse = ","), ", LOOP_NB = 'loop", i2, "', stringsAsFactors = FALSE)")))
                assign(paste0("final.confmat", i0, ".", data.kind[i0], ".", ana.kind[i1]), rbind(get(paste0("final.confmat", i0, ".", data.kind[i0], ".", ana.kind[i1])), tempo4))
            }else{
                assign(paste0("final.confmat", i0, ".", data.kind[i0], ".", ana.kind[i1]), rbind(get(paste0("final.confmat", i0, ".", data.kind[i0], ".", ana.kind[i1])), data.frame(TRUE_NR_PRED_NR = NA, TRUE_R_PRED_NR = NA, TRUE_NR_PRED_R = NA, TRUE_R_PRED_R = NA, SIZE = NA, LOOP_NB = paste0("loop", i0), stringsAsFactors = FALSE)))
            }
            # predictions
            if(nrow(get(paste0("loop", i2, "_data.pred", i0, ".", data.kind[i0], ".", ana.kind[i1]))) > 0){
                assign(paste0("final.data.pred", i0, ".", data.kind[i0], ".", ana.kind[i1]), rbind(get(paste0("final.data.pred", i0, ".", data.kind[i0], ".", ana.kind[i1])), data.frame(get(paste0("loop", i2, "_data.pred", i0, ".", data.kind[i0], ".", ana.kind[i1])), LOOP_NB = paste0("loop", i2), stringsAsFactors = FALSE)))
            }else{
                assign(paste0("final.data.pred", i0, ".", data.kind[i0], ".", ana.kind[i1]), rbind(get(paste0("final.data.pred", i0, ".", data.kind[i0], ".", ana.kind[i1])), data.frame(truth = NA, prob.NR = NA, prob.R = NA, response = NA, LOOP_NB = paste0("loop", i0), stringsAsFactors = FALSE)))
            }
            # roc curves
            if(nrow(get(paste0("loop", i2, "_data.roc", i0, ".", data.kind[i0], ".", ana.kind[i1]))) > 0){
                assign(paste0("final.data.roc", i0, ".", data.kind[i0], ".", ana.kind[i1]), rbind(get(paste0("final.data.roc", i0, ".", data.kind[i0], ".", ana.kind[i1])), data.frame(get(paste0("loop", i2, "_data.roc", i0, ".", data.kind[i0], ".", ana.kind[i1])), LOOP_NB = paste0("loop", i2), stringsAsFactors = FALSE)))
            }else{
                assign(paste0("final.data.roc", i0, ".", data.kind[i0], ".", ana.kind[i1]), rbind(get(paste0("final.data.roc", i0, ".", data.kind[i0], ".", ana.kind[i1])), data.frame(x = NA, y = NA, PANEL = NA, group = NA, coloup = NA, sizep = NA, linetypep = NA, alphap = NA, LOOP_NB = paste0("loop", i0), stringsAsFactors = FALSE)))
            }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
340
341
342
343
344
345
            # AUC
            if(length(get(paste0("loop", i2, "_auc", i0, ".", data.kind[i0], ".", ana.kind[i1]))) > 0){
                assign(paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1]), rbind(get(paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1])), data.frame(AUC = get(paste0("loop", i2, "_auc", i0, ".", data.kind[i0], ".", ana.kind[i1])), LOOP_NB = paste0("loop", i2), stringsAsFactors = FALSE)))
            }else{
                assign(paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1]), rbind(get(paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1])), data.frame(AUC = NA, LOOP_NB = paste0("loop", i2), stringsAsFactors = FALSE)))
            }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
346
347
348
349
        }
        backup.name <- c(backup.name, paste0("final.confmat", i0, ".", data.kind[i0], ".", ana.kind[i1]))
        backup.name <- c(backup.name, paste0("final.data.pred", i0, ".", data.kind[i0], ".", ana.kind[i1]))
        backup.name <- c(backup.name, paste0("final.data.roc", i0, ".", data.kind[i0], ".", ana.kind[i1]))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
350
        backup.name <- c(backup.name, paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1]))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
351
352
353
354
355
356
    }
}

# final.gene.list
final.gene.list <- data.frame(NULL, stringsAsFactors = FALSE)
for(i0 in 1:slurm.loop.nb){
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
357
    if(nrow(get(paste0("loop", i0, "_final.gene.list"))) > 0){
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
358
359
360
361
362
        final.gene.list <- rbind(final.gene.list, data.frame(get(paste0("loop", i0, "_final.gene.list")), LOOP_NB = paste0("loop", i0), stringsAsFactors = FALSE))
    }else{
        final.gene.list <- rbind(final.gene.list, data.frame(sfeats = NA, Freq = NA, LOOP_NB = paste0("loop", i0), stringsAsFactors = FALSE))
    }
}
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
363
final.gene.list <- final.gene.list[order(final.gene.list$Freq, decreasing = TRUE), ]
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
364
backup.name <- c(backup.name, "final.gene.list")
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
365
fun_export_data(path = path.out, data = paste0("THE COMPILED DATA WILL BE SAVED IN: ", paste0(path.out, "compiled_data.RData")), output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
366

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
367
368
369
370
371





Gael  MILLOT's avatar
interm    
Gael MILLOT committed
372
################ 3 Differential analysis with limma
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
373

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
374
fun_export_data(path = path.out, data = "################################ LIMMA ANALYSIS", output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
375
376
377
378
379

if(nrow(final.ttab) == 0 | nrow(na.omit(final.ttab)) == 0){
    fun_export_data(path = path.out, data = paste0("NO GENE LIST RESULTS FROM THE LIMMA ANALYSIS (P VALUES ABOVE 0.05 AFTER CORRECTION FOR INSTANCE)"), output = log.file)
    fun_export_data(path = path.out, data = final.ttab, output = log.file)
}else if(analysis.kind == "valid_boot"){
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
380
    fun_export_data(path = path.out, data = paste0("BEWARE: NO COMPILATION FOR THE LIMMA ANALYSIS (", analysis.kind, " KIND OF ANALYSIS)"), output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
381
}else{
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
382
383
384
    tempo0 <- final.ttab[ ! is.na(final.ttab$adj.P.Val), ]
    tempo1 <- c(table(rownames(tempo0)))
    tempo2 <- aggregate(tempo0$adj.P.Val, list(rownames(tempo0)), median, na.rm = TRUE)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
385
    names(tempo2) <- c("GENE", "ADJ_P_VALUE_MEDIAN")
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
386
    tempo2 <- data.frame(
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
387
        tempo2,
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
388
389
        ADJ_P_VALUE_CI95_INF = aggregate(tempo0$adj.P.Val, list(rownames(tempo0)), quantile, probs = 0.025, na.rm = TRUE)$x,
        ADJ_P_VALUE_CI95_SUP = aggregate(tempo0$adj.P.Val, list(rownames(tempo0)), quantile, probs = 0.975, na.rm = TRUE)$x
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
390
    )
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
391
    final.ttab.freq <- data.frame(tempo2, NB = unname(tempo1), FREQ = unname(tempo1)/slurm.loop.nb)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
392
    rownames(final.ttab.freq) <- NULL
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
393
394
395
396
397
398
399
400
    final.ttab.freq <- plyr::arrange(final.ttab.freq, plyr::desc(FREQ), ADJ_P_VALUE_MEDIAN) # reorder according to what is written
    if(nrow(final.ttab.freq) > 300){
        fun_export_data(path = path.out, data = "FREQUENCIES (RELATED TO NB OF TIME GENES HAVE BEEN SELECTED PER LOOP) AND P VALUES MEDIANS & CI (ONLY THE 300 FIRST FREQ, see the final.ttab.freq FILE FOR COMPLETE DATA): ", output = log.file)
        fun_export_data(path = path.out, data = final.ttab.freq[1:300, ], output = log.file)
    }else{
        fun_export_data(path = path.out, data = "FREQUENCIES (RELATED TO NB OF TIME GENES HAVE BEEN SELECTED PER LOOP) AND P VALUES MEDIANS & CI: ", output = log.file)
        fun_export_data(path = path.out, data = final.ttab.freq, output = log.file)
    }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
401
    # horiz barplot
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
402
403
404
405
406
    bar.width <- 0.5
    log.scale <- FALSE
    amplif <- label.size
    tempo.gg.name <- "gg.indiv.plot."
    tempo.gg.count <- 0
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
407
    if(nrow(final.ttab.freq) > 20){
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
408
        tempo1 <- table(final.ttab.freq$FREQ, useNA = "no")
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
409
410
411
        tempo2 <- tempo1[order(names(tempo1), decreasing = TRUE)]
        tempo3 <- as.numeric(names(cumsum(tempo2)[cumsum(tempo2) <= 20]))
        tempo.plot <- final.ttab.freq[round(final.ttab.freq$FREQ, 7) %in% round(tempo3, 7), ]
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
412
413
414
415
416
417
        if(length(tempo3) == 0){
            tempo.plot <- final.ttab.freq
            tempo.txt <- paste0("BEWARE: UNIQUE FREQUENCY ", names(tempo2), " FOR ALL THE ", unname(tempo2), " GENES -> CRAZY PLOT")
        }else{
            tempo.txt <- paste0("BEWARE: ONLY THE ", nrow(tempo.plot), " MOST FREQUENT GENES PLOTTED, AMONG ", nrow(final.ttab.freq))
        }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
418
419
420
421
    }else{
        tempo.plot <- final.ttab.freq
        tempo.txt <- ""
    }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
422
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggplot(data = tempo.plot, mapping = ggplot2::aes(x = reorder(GENE, FREQ), y = FREQ))) # reorder from higher to lower
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
423
424
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_bar(stat = "identity", position = "dodge", color = "black", fill = grey(0.8), width = bar.width))
    # assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_errorbar(ggplot2::aes(ymin = PROP.CI95.inf, ymax = PROP.CI95.sup), position = ggplot2::position_dodge(width = bar.width), color = "black", width = 0))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
425
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggtitle(paste0("LIMMA GENE LIST\n(PROP OF TIMES THE GENE IS SIGNIFICANT FOR ", slurm.loop.nb, " LOOPS)\n", tempo.txt)))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
426
427
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::xlab(""))
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ylab("PROPORTION"))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
    # assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), m.gg <- ggplot2::theme_classic(base_size = amplif))
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), m.gg <- ggplot2::theme(
        # plot.margin = margin(up.space.mds, right.space.mds, down.space.mds, left.space.mds, "inches"), 
        plot.title = ggplot2::element_text(hjust=1, vjust=1, size = amplif * 0.5), 
        axis.text.x = ggplot2::element_text(angle = 90, hjust = 1),
        axis.text = ggplot2::element_text(size = amplif * 0.75), 
        axis.title = ggplot2::element_text(size = amplif), 
        legend.text = ggplot2::element_text(size = amplif), 
        legend.title = ggplot2::element_text(size = amplif), 
        strip.text = ggplot2::element_text(size = amplif), 
        legend.key = ggplot2::element_blank(), 
        panel.background = ggplot2::element_rect(fill = "white"), 
        panel.border = ggplot2::element_rect(colour = "black", fill = NA), 
        panel.grid.major.x = ggplot2::element_line(colour = "grey75"), 
        panel.grid.major.y = ggplot2::element_line(colour = "grey75"), 
        panel.grid.minor.x = ggplot2::element_blank(), 
        panel.grid.minor.y = ggplot2::element_blank(), 
        strip.background = ggplot2::element_rect(fill = "white", colour = "black")
    ))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
447
    coord <- ggplot2::ggplot_build(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + "))))$data # to have the summary statistics of the plot. is interesting: x = coord[[2]]$x, y = coord[[2]]$ymax_final
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
448
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1),  ggplot2::annotate(geom = "text", x = coord[[1]]$x, y = coord[[1]]$ymax * 1.05, label = round(coord[[1]]$y, 3), size = amplif/3, color = "black", vjust = "center", hjust = "left")) # beware: no need of order() for labels because coord[[1]]$x set the order
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
449
450
451
452
    if(log.scale == TRUE){
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::annotation_logticks(sides = "l"))
    }else{
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::scale_y_continuous(expand = c(0,0), breaks = scales::extended_breaks(10)))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
453
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::coord_flip(ylim = c(0, max(final.ttab.freq$FREQ, na.rm = TRUE) * 1.1))) # fix limits
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
454
455
456
457
458
459
460
    }
    if(activate.pdf == TRUE){
        tempo <- dev.set(pdf.nb) # assign to avoid the message
    }else{
        windows(5, 5)
    }
    print(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + "))))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
461
462
    fun_export_data(path = path.out, data = paste0("SEE THE final.ttab AND THE final.ttab.freq OBJECTS IN THE compiled_data.RData FILE PRESENT IN: ", path.out), output = log.file)
    backup.name <- c(backup.name, "final.ttab.freq")
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
463
}
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
464

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
465

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
466
467
468



Gael  MILLOT's avatar
interm    
Gael MILLOT committed
469
470
471
472
################ 4 Machine Learning based analysis

######## 4.1 Discovery: Learning the Random Forest

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
473
fun_export_data(path = path.out, data = "################################ RANDOM FOREST LEARNING USING THE DISCOVERY SET", output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
474

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
475
#select.gene.curve (nb of features)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
476

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
477
fun_export_data(path = path.out, data = "################ OPTIMAL NUMBER OF GENES DURING THE MODEL ELABORATION", output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
478

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
479
480
481
if(nrow(final.select.gene.curve) == 0 | nrow(na.omit(final.select.gene.curve)) == 0){
    fun_export_data(path = path.out, data = paste0("NO OPTIMAL RESULTS FROM THE RANDOM FOREST TRAINING"), output = log.file)
    fun_export_data(path = path.out, data = final.select.gene.curve, output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
482
}else if(analysis.kind == "valid_boot"){
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
483
    fun_export_data(path = path.out, data = paste0("BEWARE: NO COMPILATION FOR THE THE RANDOM FOREST TRAINING (", analysis.kind, " KIND OF ANALYSIS)"), output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
484
485
    # fun_export_data(path = path.out, data = final.select.gene.curve, output = log.file)
    optimal.gene.nb <- min(final.select.gene.curve$x[which.min(final.select.gene.curve$y)], na.rm = TRUE)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
486
}else{
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
487
488
489
490
491
492
493
494
495
    tempo1 <- aggregate(final.select.gene.curve$y, list(final.select.gene.curve$x), median, na.rm = TRUE)
    names(tempo1) <- c("X", "MEDIAN")
    final.select.gene.curve.median <- cbind(tempo1,
                                            CI95.inf = aggregate(final.select.gene.curve$y, list(final.select.gene.curve$x), quantile, probs = 0.025, na.rm = TRUE)[, 2],
                                            CI95.sup = aggregate(final.select.gene.curve$y, list(final.select.gene.curve$x), quantile, probs = 0.975, na.rm = TRUE)[, 2],
                                            stringsAsFactors = FALSE
    )
    fun_export_data(path = path.out, data = "OPTIMAL NUMBER OF GENES: ", output = log.file)
    fun_export_data(path = path.out, data = final.select.gene.curve.median, output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
496
    
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
497
    # line + CI area
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
498
499
500
501
502
    bar.width <- 0.5
    log.scale <- FALSE
    amplif <- label.size
    tempo.gg.name <- "gg.indiv.plot."
    tempo.gg.count <- 0
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggplot(data = final.select.gene.curve.median, mapping = ggplot2::aes(x = X, y = MEDIAN)))
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_point(color = "red", shape = 16, size = 2))
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_line(color = "red"))
    # assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_errorbar(ggplot2::aes(ymin = CI95.inf, ymax = CI95.sup), position = ggplot2::position_dodge(), color = "black", width = 0))
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_ribbon(ggplot2::aes(ymin = CI95.inf, ymax = CI95.sup), color = NA, fill = "red", alpha = 0.1))
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggtitle(paste0("OPTIMAL NUMBER OF GENES (n = ", slurm.loop.nb,")")))
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::xlab("NUMBER OF GENES"))
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ylab("MEDIAN & 95CI"))
    # assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), m.gg <- ggplot2::theme_classic(base_size = amplif))
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), m.gg <- ggplot2::theme(
        # plot.margin = margin(up.space.mds, right.space.mds, down.space.mds, left.space.mds, "inches"), 
        plot.title = ggplot2::element_text(hjust=1, vjust=1, size = amplif * 0.5), 
        axis.text.x = ggplot2::element_text(angle = 90, hjust = 1),
        axis.text = ggplot2::element_text(size = amplif * 0.75), 
        axis.title = ggplot2::element_text(size = amplif), 
        legend.text = ggplot2::element_text(size = amplif), 
        legend.title = ggplot2::element_text(size = amplif), 
        strip.text = ggplot2::element_text(size = amplif), 
        legend.key = ggplot2::element_blank(), 
        panel.background = ggplot2::element_rect(fill = "white"), 
        panel.border = ggplot2::element_rect(colour = "black", fill = NA), 
        panel.grid.major.x = ggplot2::element_line(colour = "white"), 
        panel.grid.major.y = ggplot2::element_line(colour = "grey75"), 
        panel.grid.minor.x = ggplot2::element_blank(), 
        panel.grid.minor.y = ggplot2::element_blank(), 
        strip.background = ggplot2::element_rect(fill = "white", colour = "black")
    ))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
530
531
532
    if(log.scale == TRUE){
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::annotation_logticks(sides = "l"))
    }else{
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
533
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::scale_x_continuous(expand = c(0,0)))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
534
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::scale_y_continuous(expand = c(0,0), breaks = scales::extended_breaks(10)))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
535
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::coord_cartesian(xlim= c(min(final.select.gene.curve.median$X, na.rm = TRUE), max(final.select.gene.curve.median$X, na.rm = TRUE)), ylim = c(0, max(final.select.gene.curve.median$CI95.sup, na.rm = TRUE) * 1.1))) # fix limits
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
536
537
538
539
540
541
542
    }
    if(activate.pdf == TRUE){
        tempo <- dev.set(pdf.nb) # assign to avoid the message
    }else{
        windows(5, 5)
    }
    print(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + "))))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
543
544
545
546
547
548
549
550
551
552
553
    fun_export_data(path = path.out, data = paste0("SEE THE final.select.gene.curve AND THE final.select.gene.curve.median OBJECTS IN THE compiled_data.RData FILE PRESENT IN: ", path.out), output = log.file)
    backup.name <- c(backup.name, "final.select.gene.curve.median")

    tempo.data <- final.select.gene.curve.median[ ! is.na(final.select.gene.curve.median$MEDIAN), ]
    if(nrow(tempo.data) == 0){
        tempo.cat <- paste0("NO OPTIMAL NUMBER OF GENE SELECTED BECAUSE OF NA IN final.select.gene.curve.median$MEDIAN")
        fun_export_data(path = path.out, data = tempo.cat, output = log.file)
        stop(tempo.cat)
    }else{
        optimal.gene.nb <- min(tempo.data$X[which.min(tempo.data$MEDIAN)], na.rm = TRUE)
    }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
554
}
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
555
556


Gael  MILLOT's avatar
interm    
Gael MILLOT committed
557
#mod.gene.names (final gene list)
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
558

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
559
fun_export_data(path = path.out, data = "################ LIST OF GENES SELECTED DURING THE MODEL ELABORATION", output = log.file)
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
560

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
561
562
563
if(nrow(final.mod.gene.names) == 0 | nrow(na.omit(final.mod.gene.names)) == 0){
    fun_export_data(path = path.out, data = paste0("NO GENE LIST RESULTS FROM THE RANDOM FOREST TRAINING"), output = log.file)
    fun_export_data(path = path.out, data = final.mod.gene.names, output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
564
}else if(analysis.kind == "valid_boot"){
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
565
    fun_export_data(path = path.out, data = paste0("BEWARE: NO COMPILATION FOR THE THE RANDOM FOREST TRAINING (", analysis.kind, " KIND OF ANALYSIS)"), output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
566
567
    # fun_export_data(path = path.out, data = final.mod.gene.names, output = log.file)
    selected.gene.names <- as.character(final.mod.gene.names$GENE)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
568
}else{
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
    tempo1 <- c(table(final.mod.gene.names$GENE, useNA = "no"))
    final.mod.gene.names.freq <- data.frame(GENE = names(tempo1), NB = unname(tempo1), FREQ = unname(tempo1)/slurm.loop.nb)
    final.mod.gene.names.freq <- final.mod.gene.names.freq[order(final.mod.gene.names.freq$FREQ, decreasing = TRUE), ]
    rownames(final.mod.gene.names.freq) <- NULL
    if(nrow(final.mod.gene.names.freq) > optimal.gene.nb){
        tempo1 <- table(final.mod.gene.names.freq$FREQ, useNA = "no")
        tempo2 <- tempo1[order(names(tempo1), decreasing = TRUE)]
        tempo3 <- as.numeric(names(cumsum(tempo2)[cumsum(tempo2) <= optimal.gene.nb]))
        if(length(tempo3) == 0){
            fun_export_data(path = path.out, data = paste0("BEWARE: THE OPTIMAL GENE NUMBER ", optimal.gene.nb, " COULD NOT BE APPLIED TO final.mod.gene.names.freq BECAUSE THE FIRST CLASS OF FREQ IS OVER ", optimal.gene.nb), output = log.file, sep = 1)
            fun_export_data(path = path.out, data = tempo2[1], output = log.file)
            fun_export_data(path = path.out, data = "THIS CLASS OF FREQ IS USED FOR THE GRAPH AND FINAL final.mod.gene.names.freq", output = log.file)
            final.mod.gene.names.freq <- final.mod.gene.names.freq[round(final.mod.gene.names.freq$FREQ, 7) %in% round(as.numeric(names(tempo2[1])), 7), ]
        }else{
            final.mod.gene.names.freq <- final.mod.gene.names.freq[round(final.mod.gene.names.freq$FREQ, 7) %in% round(tempo3, 7), ]
            if(nrow(final.mod.gene.names.freq) != optimal.gene.nb){
                fun_export_data(path = path.out, data = paste0("BEWARE: THE OPTIMAL GENE NUMBER ", optimal.gene.nb, " IS FINALLY LOWER (", nrow(final.mod.gene.names.freq), ") IN final.mod.gene.names.freq BECAUSE THE NUMBER FALLS INTO THE MIDDLE OF GENES WITH IDENTICAL FREQUENCIES"), output = log.file, sep = 1)
            }
        }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
588
    }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
589
590
591
592
593
    optimal.gene.nb <- nrow(final.mod.gene.names.freq) # nb reduced as explained above
    selected.gene.names <- as.character(final.mod.gene.names.freq$GENE)
    fun_export_data(path = path.out, data = paste0("THE FINAL SELECTED GENE LIST (n = ", optimal.gene.nb, ") DERIVES FROM THE FREQUENCIES OF getFilteredFeatures(mod) ON MULTIPLE LOOPS"), output = log.file, sep = 1)
    fun_export_data(path = path.out, data = paste0("SELECTED GENE LIST (n = ", optimal.gene.nb, ") AND FREQUENCIES: "), output = log.file)
    fun_export_data(path = path.out, data = final.mod.gene.names.freq, output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
594
    
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
595
    # horiz barplot
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
596
597
598
599
600
    bar.width <- 0.5
    log.scale <- FALSE
    amplif <- label.size
    tempo.gg.name <- "gg.indiv.plot."
    tempo.gg.count <- 0
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
601
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggplot(data =  final.mod.gene.names.freq, mapping = ggplot2::aes(x = reorder(GENE, FREQ), y = FREQ))) # reorder from higher to lower
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
602
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_bar(stat = "identity", position = "dodge", color = "black", fill = grey(0.8), width = bar.width))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
603
604
    # assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_errorbar(ggplot2::aes(ymin = PROP.CI95.inf, ymax = PROP.CI95.sup), position = ggplot2::position_dodge(width = bar.width), color = "black", width = 0))
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggtitle(paste0("RANDOM FOREST MODELISATION GENE LIST\n(PROP OF TIMES THE GENES HAVE BEEN SELECTED FOR ", slurm.loop.nb, " LOOPS)")))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
605
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::xlab(""))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
606
607
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ylab("PROPORTION"))
    # assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), m.gg <- ggplot2::theme_classic(base_size = amplif))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
608
609
610
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), m.gg <- ggplot2::theme(
        # plot.margin = margin(up.space.mds, right.space.mds, down.space.mds, left.space.mds, "inches"), 
        plot.title = ggplot2::element_text(hjust=1, vjust=1, size = amplif * 0.5), 
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
611
        axis.text.x = ggplot2::element_text(angle = 90, hjust = 1),
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
612
613
614
615
616
617
        axis.text = ggplot2::element_text(size = amplif * 0.75), 
        axis.title = ggplot2::element_text(size = amplif), 
        legend.text = ggplot2::element_text(size = amplif), 
        legend.title = ggplot2::element_text(size = amplif), 
        strip.text = ggplot2::element_text(size = amplif), 
        legend.key = ggplot2::element_blank(), 
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
618
        panel.background = ggplot2::element_rect(fill = "white"), 
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
619
620
621
622
623
624
625
626
        panel.border = ggplot2::element_rect(colour = "black", fill = NA), 
        panel.grid.major.x = ggplot2::element_line(colour = "grey75"), 
        panel.grid.major.y = ggplot2::element_line(colour = "grey75"), 
        panel.grid.minor.x = ggplot2::element_blank(), 
        panel.grid.minor.y = ggplot2::element_blank(), 
        strip.background = ggplot2::element_rect(fill = "white", colour = "black")
    ))
    coord <- ggplot2::ggplot_build(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + "))))$data # to have the summary statistics of the plot. is interesting: x = coord[[2]]$x, y = coord[[2]]$ymax_final
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
627
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1),  ggplot2::annotate(geom = "text", x = coord[[1]]$x, y = coord[[1]]$ymax * 1.05, label = round(coord[[1]]$y, 3), size = amplif/3, color = "black", vjust = "center", hjust = "left")) # beware: no need of order() for labels because coord[[1]]$x set the order
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
628
629
630
631
    if(log.scale == TRUE){
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::annotation_logticks(sides = "l"))
    }else{
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::scale_y_continuous(expand = c(0,0), breaks = scales::extended_breaks(10)))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
632
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::coord_flip(ylim = c(0, max(final.mod.gene.names.freq$FREQ, na.rm = TRUE) * 1.1))) # fix limits
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
633
634
635
636
637
638
639
    }
    if(activate.pdf == TRUE){
        tempo <- dev.set(pdf.nb) # assign to avoid the message
    }else{
        windows(5, 5)
    }
    print(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + "))))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
640
    fun_export_data(path = path.out, data = paste0("SEE THE final.mod.gene.names AND THE final.mod.gene.names.freq OBJECTS IN THE compiled_data.RData FILE PRESENT IN: ", path.out), output = log.file)
Gael  MILLOT's avatar
Gael MILLOT committed
641
    backup.name <- c(backup.name, "final.mod.gene.names.freq", "optimal.gene.nb", "selected.gene.names")
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
642
}
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
643
644


Gael  MILLOT's avatar
interm    
Gael MILLOT committed
645
#gene.importance
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
646

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
647
fun_export_data(path = path.out, data = "################ GENE IMPORTANCE DURING THE MODEL ELABORATION", output = log.file)
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
648

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
649
650
651
if(nrow(final.gene.importance) == 0 | nrow(na.omit(final.gene.importance)) == 0){
    fun_export_data(path = path.out, data = paste0("NO GENE IMPORTANCE RESULTS FROM THE RANDOM FOREST TRAINING"), output = log.file)
    fun_export_data(path = path.out, data = final.gene.importance, output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
652
}else if(analysis.kind == "valid_boot"){
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
653
    fun_export_data(path = path.out, data = paste0("BEWARE: NO COMPILATION FOR THE THE RANDOM FOREST TRAINING (", analysis.kind, " KIND OF ANALYSIS)"), output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
654
    # fun_export_data(path = path.out, data = final.gene.importance, output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
655
}else{
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
656
657
658
659
660
    tempo1 <- aggregate(final.gene.importance$importance, list(final.gene.importance$features), median, na.rm = TRUE)
    names(tempo1) <- c("GENE", "MEDIAN")
    tempo2 <- aggregate(final.gene.importance$features, list(final.gene.importance$features), function(x){length(x[ ! is.na(x)])}) # nb
    if(identical(tempo1$GENE, tempo2$Group.1)){
        tempo1 <- data.frame(tempo1, NB = tempo2$x)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
661
    }else{
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
662
663
664
665
666
        cat(paste0("\n\n============\n\nERROR: IN THE PLOT OF THE final.gene.importance OBJECT, tempo1 AND tempo2 SHOULD HAVE THE SAME GENE COLUMN: \n\n============\n\n"))
        tempo1
        tempo2
        cat("\n\n============\n\n")
        stop()
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
667
    }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
668
669
670
671
672
673
674
675
676
    final.gene.importance.median <- cbind(tempo1,
                                          CI95.inf = aggregate(final.gene.importance$importance, list(final.gene.importance$features), quantile, probs = 0.025, na.rm = TRUE)[, 2],
                                          CI95.sup = aggregate(final.gene.importance$importance, list(final.gene.importance$features), quantile, probs = 0.975, na.rm = TRUE)[, 2],
                                          stringsAsFactors = FALSE
    )
    final.gene.importance.median <- final.gene.importance.median[order(final.gene.importance.median$MEDIAN, decreasing = TRUE), ]
    if(sum(final.gene.importance.median$GENE %in% selected.gene.names, na.rm = TRUE) == 0){
        tempo.cat <- paste0("NO IMPORTANCE OF GENE SELECTED BECAUSE NO MATCH BETWEEN GENES IN final.gene.importance.median AND THE SELECTED GENE LIST")
        fun_export_data(path = path.out, data = tempo.cat, output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
677
    }else{
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
        final.gene.importance.median <- final.gene.importance.median[final.gene.importance.median$GENE %in% selected.gene.names, ]
        fun_export_data(path = path.out, data = paste0("IMPORTANCE OF THE SELECTED GENES (n = ", nrow(final.gene.importance.median), "): "), output = log.file)
        fun_export_data(path = path.out, data = final.gene.importance.median, output = log.file)
        
        # barplot + CI
        bar.width <- 0.5
        log.scale <- FALSE
        amplif <- label.size
        tempo.gg.name <- "gg.indiv.plot."
        tempo.gg.count <- 0
        if(nrow(final.gene.importance.median) > 20){
            tempo1 <- table(final.gene.importance.median$MEDIAN, useNA = "no")
            tempo2 <- tempo1[order(names(tempo1), decreasing = TRUE)]
            tempo3 <- as.numeric(names(cumsum(tempo2)[cumsum(tempo2) <= 20]))
            tempo.plot <- final.gene.importance.median[round(final.gene.importance.median$MEDIAN, 7) %in% round(tempo3, 7), ]
            assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggplot(data = tempo.plot, mapping = ggplot2::aes(x = reorder(GENE, -MEDIAN), y = MEDIAN))) # reorder from higher to lower
            tempo.txt <- paste0("BEWARE: ONLY THE ", nrow(tempo.plot), " MOST IMPORTANT GENES PLOTTED, AMONG ", nrow(final.gene.importance.median))
        }else{
            tempo.plot <- final.gene.importance.median
            assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggplot(data =  tempo.plot, mapping = ggplot2::aes(x = reorder(GENE, -MEDIAN), y = MEDIAN))) # reorder from higher to lower
            tempo.txt <- ""
        }
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_bar(stat = "identity", position = "dodge", color = "black", fill = grey(0.8), width = bar.width))
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_errorbar(ggplot2::aes(ymin = CI95.inf, ymax = CI95.sup), position = ggplot2::position_dodge(width = bar.width), color = "black", width = 0))
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggtitle(paste0("GENE IMPORTANCE (NUMBERS ARE NUMBER OF TIMES THE GENE IS SEEN FOR ", slurm.loop.nb, " LOOPS)")))
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::xlab(""))
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ylab("MEDIAN & 95CI"))
        #     assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), m.gg <- ggplot2::theme_classic(base_size = amplif))
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), m.gg <- ggplot2::theme(
            # plot.margin = margin(up.space.mds, right.space.mds, down.space.mds, left.space.mds, "inches"), 
            plot.title = ggplot2::element_text(hjust=1, vjust=1, size = amplif * 0.5), 
            axis.text.x = ggplot2::element_text(angle = 90, hjust = 1),
            axis.text = ggplot2::element_text(size = amplif * 0.75), 
            axis.title = ggplot2::element_text(size = amplif), 
            legend.text = ggplot2::element_text(size = amplif), 
            legend.title = ggplot2::element_text(size = amplif), 
            strip.text = ggplot2::element_text(size = amplif), 
            legend.key = ggplot2::element_blank(), 
            panel.background = ggplot2::element_rect(fill = "grey95"), 
            panel.border = ggplot2::element_rect(colour = "black", fill = NA), 
            panel.grid.major.x = ggplot2::element_line(colour = "grey75"), 
            panel.grid.major.y = ggplot2::element_line(colour = "grey75"), 
            panel.grid.minor.x = ggplot2::element_blank(), 
            panel.grid.minor.y = ggplot2::element_blank(), 
            strip.background = ggplot2::element_rect(fill = "white", colour = "black")
        ))
        coord <- ggplot2::ggplot_build(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + "))))$data # to have the summary statistics of the plot. is interesting: x = coord[[2]]$x, y = coord[[2]]$ymax_final
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1),  ggplot2::annotate(geom = "text", x = coord[[2]]$x, y = coord[[2]]$ymax * 1.05, label = tempo.plot$NB, size = amplif/2, color = "black", vjust = "bottom", hjust = "center", angle = 00)) # beware: no need of order() for labels because coord[[1]]$x set the order
        if(log.scale == TRUE){
            assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::annotation_logticks(sides = "l"))
        }else{
            assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::scale_y_continuous(expand = c(0,0), breaks = scales::extended_breaks(10)))
            assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::coord_cartesian(ylim = c(0, max(tempo.plot$CI95.sup, na.rm = TRUE) * 1.1))) # fix limits
        }
        if(activate.pdf == TRUE){
            tempo <- dev.set(pdf.nb) # assign to avoid the message
        }else{
            windows(5, 5)
        }
        print(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + "))))
        fun_export_data(path = path.out, data = paste0("SEE THE final.gene.importance AND THE final.gene.importance.median OBJECTS IN THE compiled_data.RData FILE PRESENT IN: ", path.out), output = log.file)
        backup.name <- c(backup.name, "final.gene.importance.median")
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
740
741
    }
}
Gael  MILLOT's avatar
interm  
Gael MILLOT committed
742

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
743
# plotting the heatmap adnd corplot
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
744
745
fun_export_data(path = path.out, data = "################ HEATMAP AND CORRELATION PLOT ", output = log.file)
if(analysis.kind == "valid_boot"){
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
746
    if(length(selected.gene.names) == 0){
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
747
748
749
750
751
752
753
754
        fun_export_data(path = path.out, data = paste0("NO GENE LIST FROM THE RANDOM FOREST TRAINING"), output = log.file)
        fun_export_data(path = path.out, data = final.mod.gene.names, output = log.file)
    }else{
        fun_export_data(path = path.out, data = paste0("BEWARE: NO COMPILATION FOR THE THE RANDOM FOREST TRAINING (", analysis.kind, " KIND OF ANALYSIS)"), output = log.file)
        # fun_export_data(path = path.out, data = final.mod.gene.names, output = log.file)
        tempo.data <- dat[order(dat$Y), ]
        annot.rows <- dat[, "Y", drop = FALSE] # allow to keep the data.frame structure
        colnames(annot.rows) <- "ASDAS R/NR"
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
755
756
757
758
759
760
761
762
763
764
765
        if(length(na.omit(match(selected.gene.names, names(tempo.data)))) != length(selected.gene.names)){
            tempo.cat <- cat(paste0("\n\n============\n\nERROR: THE COLUMN NAMES OF tempo.data OBJECT SHOULD MATCH ALL THE NAMES IN selected.gene.names.\n
                    PROBLEM IN selected.gene.names:\n", selected.gene.names[ ! selected.gene.names %in% names(tempo.data)], "
                    PROBLEM IN names(tempo.data):\n", names(tempo.data)[ ! names(tempo.data) %in% selected.gene.names], "  
                       \n\n============\n\n"))
            fun_export_data(path = path.out, data = tempo.cat, output = log.file)
            stop(tempo.cat)
        }else{
            fun_export_data(path = path.out, data = "IN THE nano.rf.selected.genes OBJECT, COLUMN ARE ORDERED AS IN THE FREQ OF THE GENE LIST selected.gene.names", output = log.file)
            nano.rf.selected.genes <- tempo.data[, match(selected.gene.names, names(tempo.data))]
        }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
766
    }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
767
}else{
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
768
    if(length(selected.gene.names) == 0){
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
769
770
771
772
773
774
        fun_export_data(path = path.out, data = paste0("NO GENE LIST FROM THE RANDOM FOREST TRAINING"), output = log.file)
        fun_export_data(path = path.out, data = final.mod.gene.names.freq, output = log.file)
    }else{
        tempo.data <- dat[order(dat$Y), ]
        annot.rows <- dat[, "Y", drop = FALSE] # allow to keep the data.frame structure
        colnames(annot.rows) <- "ASDAS R/NR"
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
775
776
777
778
779
780
781
782
783
784
785
        if(length(na.omit(match(selected.gene.names, names(tempo.data)))) != length(selected.gene.names)){
            tempo.cat <- cat(paste0("\n\n============\n\nERROR: THE COLUMN NAMES OF tempo.data OBJECT SHOULD MATCH ALL THE NAMES IN selected.gene.names.\n
                                    PROBLEM IN selected.gene.names:\n", selected.gene.names[ ! selected.gene.names %in% names(tempo.data)], "
                                    PROBLEM IN names(tempo.data):\n", names(tempo.data)[ ! names(tempo.data) %in% selected.gene.names], "  
                                    \n\n============\n\n"))
            fun_export_data(path = path.out, data = tempo.cat, output = log.file)
            stop(tempo.cat)
        }else{
            fun_export_data(path = path.out, data = "IN THE nano.rf.selected.genes OBJECT, COLUMN ARE ORDERED AS IN THE FREQ OF THE GENE LIST selected.gene.names", output = log.file)
            nano.rf.selected.genes <- tempo.data[, match(selected.gene.names, names(tempo.data))]
        }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
786
787
788
789
790
791
792
793
794
795
        ann_colors = list("ASDAS R/NR" = c(R = "steelblue", NR = "tomato"))
        # pheatmap(t(scale(nano.rf.selected.genes)), silent = TRUE, annotation_col = annot.rows, cluster_cols = FALSE, show_colnames = FALSE, border_color = NA,  color = colorRampPalette(c("red", "black", "green"))(499), annotation_colors = ann_colors, fontsize_row = label.size, fontsize_col = label.size)
        tempo <- dev.set(pdf.nb) # assign to avoid the message
        heatmap.plot <- pheatmap(t(scale(nano.rf.selected.genes)), silent = TRUE, annotation_col = annot.rows, cluster_cols = FALSE, show_colnames = FALSE, border_color = NA,  color = colorRampPalette(c("red", "black", "green"))(499), annotation_colors = ann_colors, fontsize_row = label.size, fontsize_col = label.size)
        print(ggplot2::ggplot()+ggplot2::theme_classic())
        print(heatmap.plot)
        tempo <- dev.set(pdf.nb) # assign to avoid the message
        corrplot::corrplot(cor(nano.rf.selected.genes), tl.col = "black", tl.cex = label.size / 10)
        fun_export_data(path = path.out, data = paste0("SEE THE nano.rf.selected.genes OBJECTS IN THE compiled_data.RData FILE PRESENT IN: ", path.out), output = log.file)
    }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
796
}
Gael  MILLOT's avatar
Gael MILLOT committed
797
backup.name <- c(backup.name, "nano.rf.selected.genes", "dat")
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
798
799


Gael  MILLOT's avatar
interm  
Gael MILLOT committed
800
801
######## 4.2 Validate the model
# 
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
802
fun_export_data(path = path.out, data = "################################ VALIDATION", output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
803
804
805
806
807
fun_export_data(path = path.out, data = "################ ROC CURVES", output = log.file)
data.kind <- c("genes", "crp", "genes.crp") # beware, the order is important
ana.kind <- c("rf", "logreg", "rpart")
for(i0 in 1:length(data.kind)){
    for(i1 in 1:length(ana.kind)){
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
808
809
        tempo.name <- paste0("final.data.roc", i0, ".", data.kind[i0], ".", ana.kind[i1])
        tempo.name.median <- paste0("final.data.roc", i0, ".", data.kind[i0], ".", ana.kind[i1], ".median")
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
810
        fun_export_data(path = path.out, data = paste0("######## ", data.kind[i0], " AND ", ana.kind[i1], " ANALYSIS"), output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
811
        if(nrow(get(tempo.name)) == 0 | nrow(na.omit(get(tempo.name))) == 0){
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
812
            fun_export_data(path = path.out, data = paste0("NO ROC CURVE FOR THE ", data.kind[i0], " AND ", ana.kind[i1], " ANALYSIS"), output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
813
            fun_export_data(path = path.out, data = get(tempo.name), output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
814
        }else{
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
815
816
            # roc curve compil
            tempo1 <- aggregate(get(tempo.name)$x, list(get(tempo.name)$CUTOFF), median, na.rm = TRUE)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
817
            names(tempo1) <- c("CUTOFF", "X_MEDIAN")
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
818
            tempo2 <- aggregate(get(tempo.name)$y, list(get(tempo.name)$CUTOFF), median, na.rm = TRUE)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
819
            names(tempo2) <- c("CUTOFF", "Y_MEDIAN")
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
820
821
822
823
824
825
826
827
            assign(tempo.name.median, cbind(tempo1, tempo2["Y_MEDIAN"]))
            assign(tempo.name.median, get(tempo.name.median)[order(get(tempo.name.median)$X_MEDIAN, get(tempo.name.median)$Y_MEDIAN, decreasing = FALSE), ]) # to get proper lines on the plot
            # auc compil
            assign(paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1], ".median"), data.frame(
                MEDIAN = median(get(paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1]))$AUC, na.rm = TRUE),
                CI95.inf = quantile(get(paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1]))$AUC, probs = 0.025, na.rm = TRUE),
                CI95.sup = quantile(get(paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1]))$AUC, probs = 0.975, na.rm = TRUE)
            ))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
828
            
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
829
            # roc curves plot
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
830
831
832
833
834
            bar.width <- 0.5
            log.scale <- FALSE
            amplif <- label.size
            tempo.gg.name <- "gg.indiv.plot."
            tempo.gg.count <- 0
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
835
            assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggplot(data = get(tempo.name.median), mapping = ggplot2::aes(x = X_MEDIAN, y = Y_MEDIAN)))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
836
837
838
            for(i2 in 1:slurm.loop.nb){
                # roc curves
                if(nrow(get(paste0("loop", i2, "_data.roc", i0, ".", data.kind[i0], ".", ana.kind[i1]))) > 0){
Gael  MILLOT's avatar
Gael MILLOT committed
839
                    backup.name <- c(backup.name, paste0("loop", i2, "_data.roc", i0, ".", data.kind[i0], ".", ana.kind[i1]))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
840
                    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_line(data = get(paste0("loop", i2, "_data.roc", i0, ".", data.kind[i0], ".", ana.kind[i1]))[order(get(paste0("loop", i2, "_data.roc", i0, ".", data.kind[i0], ".", ana.kind[i1]))$x, get(paste0("loop", i2, "_data.roc", i0, ".", data.kind[i0], ".", ana.kind[i1]))$y, decreasing = FALSE), ], mapping = ggplot2::aes(x = x, y = y), color = "grey", alpha = 0.5))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
841
842
                }
            }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
843
844
            assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_segment(data = data.frame(x = 0, y = 0, xend = 1, yend = 1), mapping = ggplot2::aes(x = x, y = y, xend = xend, yend = yend), color = "black", linetype = "dashed"))
            assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_line(color = "red", size =0.75))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
845
846
847
848
849
850
851
            assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggtitle(paste0("ROC CURVE ", data.kind[i0], " AND ", ana.kind[i1], " (n = ", slurm.loop.nb,")\nAUC = ",
                        round(get(paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1], ".median"))$MEDIAN, 3), " (",
                        round(get(paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1], ".median"))$CI95.inf, 3), " - ",
                        round(get(paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1], ".median"))$CI95.sup, 3), ")"
                    )))
            assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::xlab("FALSE POSITIVE RATE (1 - SPECIFICITY)"))
            assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ylab("TRUE POSITIVE RATE (SENSITIVITY)"))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
852
853
            assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), m.gg <- ggplot2::theme_classic(base_size = amplif))
            assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::scale_y_continuous(expand = c(0,0), breaks = scales::extended_breaks(10)))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
854
            assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::scale_x_continuous(expand = c(0,0), breaks = scales::extended_breaks(10)))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
855
856
857
858
859
860
861
862
            assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::coord_cartesian(xlim = c(0, 1), ylim = c(0, 1))) # fix limits
            if(activate.pdf == TRUE){
                tempo <- dev.set(pdf.nb) # assign to avoid the message
            }else{
                windows(5, 5)
            }
            print(eval(parse(text = paste(paste0(tempo.gg.name, 1:tempo.gg.count), collapse = " + "))))
        }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
863
        # fun_export_data(path = path.out, data = paste0("SEE THE ", tempo.name.median, " AND THE ", paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1], ".median"), " OBJECTS IN THE compiled_data.RData FILE PRESENT IN: ", path.out), output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
864
865
866
867
868
869
870
871
        backup.name <- c(backup.name, tempo.name.median, paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1], ".median"))
        
        # confusion matrix
        fun_export_data(path = path.out, data = paste0("CONFUSION MATRIX (CI95):"), output = log.file)
        tempo.name <- paste0("final.confmat", i0, ".", data.kind[i0], ".", ana.kind[i1])
        tempo1 <- apply(get(tempo.name)[1:4], 2, median, na.rm = TRUE)
        tempo2 <- apply(get(tempo.name)[1:4], 2, quantile, probs = 0.025, na.rm = TRUE)
        tempo3 <- apply(get(tempo.name)[1:4], 2, quantile, probs = 0.975, na.rm = TRUE)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
872
873
        tempo4 <- paste(tempo1, round(tempo2, 3), sep = " (")
        tempo4 <- paste(tempo4, round(tempo3, 3), sep = " - ")
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
874
875
        tempo4 <- paste0(tempo4, ")")
        assign(paste0("final.confmat", i0, ".", data.kind[i0], ".", ana.kind[i1], ".median"), as.table(matrix(tempo4, ncol = 2, dimnames = list(TRUTH = c("NR", "R"), PREDICTION = c("NR", "R")))))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
876
        fun_export_data(path = path.out, data = get(paste0("final.confmat", i0, ".", data.kind[i0], ".", ana.kind[i1], ".median")), output = log.file, rownames.kept = TRUE)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
877
878
        fun_export_data(path = path.out, data = paste0("SEE THE ", tempo.name.median, " AND THE ", paste0("final.auc", i0, ".", data.kind[i0], ".", ana.kind[i1], ".median"), " OBJECT IN THE compiled_data.RData FILE PRESENT IN: ", path.out), output = log.file)
        backup.name <- c(backup.name, tempo.name.median, paste0("final.confmat", i0, ".", data.kind[i0], ".", ana.kind[i1], ".median"))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
879
880
881
    }
}

Gael  MILLOT's avatar
interm  
Gael MILLOT committed
882
883
884
885


######## 4.6 Plot top feature boxplots

Gael  MILLOT's avatar
interm    
Gael MILLOT committed
886
fun_export_data(path = path.out, data = "################ TOP PREDICTOR GENES", output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
887
boxdat <- data.frame(ASDAS = df.nano$response_ASDAS_R_NR, 
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
888
889
890
891
                     Cohort.name = df.nano$cohort_id,
                     RF.COHORTE = df.nano$training_validation, 
                     nano.rf.selected.genes
)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
892
boxdat_melt <- reshape2::melt(boxdat, id.vars = c("ASDAS", "Cohort.name", "RF.COHORTE"), variable.name = "Gene")
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
893
boxdat_melt$Gene <- factor(boxdat_melt$Gene, levels = unique(boxdat_melt$Gene))
Gael  MILLOT's avatar
Gael MILLOT committed
894
backup.name <- c(backup.name, "boxdat_melt")
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
895
896
897
898
899

# boxplot
bar.width <- 0.5
log.scale <- FALSE
amplif <- label.size
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
900
901
facet.nrow = 4
facet.ncol = 4
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
902
pages.print.nb <- ceiling(length(levels(boxdat_melt$Gene)) / (facet.nrow * facet.ncol))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
903
904
905
increm <- facet.nrow * facet.ncol
for(i0 in 1:pages.print.nb){
    select <- (increm * i0 - increm + 1):(increm * i0)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
906
907
    if(max(select) > length(levels(boxdat_melt$Gene))){
        select <- (increm * i0 - increm + 1):length(levels(boxdat_melt$Gene))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
908
909
910
911
912
913
914
    }
    tempo.gg.name <- "gg.indiv.plot."
    tempo.gg.count <- 0
    if(analysis.kind == "valid_boot"){
        if(i0 == 1){
            fun_export_data(path = path.out, data = paste0("WITH ", analysis.kind, " KIND OF ANALYSIS, BOXPLOTS ARE PROVIDED USING THE TRAINING AND THE VALIDATING COHORTES, AS DEFINED IN THE training_validation COLUMN OF THE df.nano DATA FRAME", path.out), output = log.file)
        }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
915
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggplot(data = boxdat_melt[boxdat_melt$Gene %in% levels(boxdat_melt$Gene)[select], ], mapping = ggplot2::aes(x=ASDAS, y=value, colour=ASDAS, shape = RF.COHORTE, linetype = RF.COHORTE)))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
916
917
    }else if(analysis.kind == "full_cross_validation"){
        if(i0 == 1){
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
918
            fun_export_data(path = path.out, data = paste0("WITH ", analysis.kind, " KIND OF ANALYSIS, BOXPLOTS CANNOT BE DIFFERENCIATED BETWEEN THE TRAINING AND THE VALIDATING COHORTES"), output = log.file)
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
919
        }
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
920
        assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::ggplot(data = boxdat_melt[boxdat_melt$Gene %in% levels(boxdat_melt$Gene)[select], ], mapping = ggplot2::aes(x=ASDAS, y=value, colour=ASDAS)))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
921
922
923
924
925
    }else{
        stop(paste0("\n\n============\n\nERROR 1 IN R CODE\n\n============\n\n"))
    }
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggbeeswarm::geom_beeswarm(dodge.width = 0.75, alpha = 0.7))
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::geom_boxplot(outlier.shape = NA, fill = NA))
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
926
    assign(paste0(tempo.gg.name, tempo.gg.count <- tempo.gg.count + 1), ggplot2::facet_wrap(~ Gene, nrow = facet.nrow, ncol = facet.ncol, scales = "fixed", shrink = FALSE, drop = TRUE)) # scales = "fixed", shrink = FALSE to have same scale range
Gael  MILLOT's avatar
interm    
Gael MILLOT committed
927
928
929
930
931
932
933