-
Notifications
You must be signed in to change notification settings - Fork 1
/
wTO.Complete2.R
411 lines (336 loc) · 17 KB
/
wTO.Complete2.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
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
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
#' @title wTO.Complete
#' @author Deisy Morselli Gysi <deisy at bioinf.uni-leipzig.de>
#' @param k Number of threads to be used for computing the weight Topological Overlap. Default is set to 1.
#' @param n Number of resamplings, used to compute the empirical distribuitions of the links. Default is set to 100.
#' @param Data data.frame containing the count / expression data for the correlation.
#' @param Overlap Set of nodes of interest, where the Overlapping weights will be computed.
#' @param method Type of the correlation that should be used. "s" / "spearman" will compute the rank spearman correlation, "p" / "pearson" will compute the linear correlation. If no value is given, the default is to use "p".
#' @param method_resampling method of the resampling. Bootstrap, BlockBootstrap or Reshuffle. Bootstrap null hypothesis is that the wTO is random, and Reshuffle tests if the wTO is equal to zero.
#' @param pvalmethod method to compute the multiple test correction for the pvalue. for more information check the function \code{\link[stats]{p.adjust}}.
#' @param savecor T/F if need to save the correlation.
#' @param expected.diff Difference expected between the real wTO and resampled wTO By default, it is set to 0.2.
#' @param lag time dependency, lag, if you are using the BlockedBootstrap.
#' @param ID ID of the samples for the blocked bootstrap (for repeated measures).
#' @param normalize T/F Should the data be normalized?
#' @param plot T/F Should the diagnosis plot be plotted?
#'
#' @description Compute the wTO and also the bootstraps. Proposed at: arXiv:1711.04702
#' @return a list with results.
#' \itemize{
#' \item wTO is a data.frame containig the Nodes, the wTO computed using the signed correlations, the pvalue and the adj.pvalue.
#' \item abs.wTO is a data.frame containig the Nodes, the wTO computed using the absolute correlations, the pvalue and the adj.pvalue.
#' \item Correlation is a data.frame containing the correlation between all the nodes.
#' \item Empirical.Quantile quantile values for the empirical distribution.
#' \item Quantile quantile values for the sample distribution.
#' }
#' @importFrom parallel makeCluster clusterExport clusterApplyLB stopCluster
#' @importFrom data.table rbindlist dcast
#' @importFrom som normalize
#' @importFrom stats cor p.adjust reshape
#' @importFrom graphics plot axis par abline legend
#'
#'
#' @examples
#' \dontrun{
#' # Using spearman rank correlation and bonferroni correction for the pvalues.
#' wTO.Complete( k =8, n = 1000, Data = Microarray_Expression1,
#' Overlap = ExampleGRF$x, method = "s", pvalmethod = "bonferroni")
#' # Changing the resampling method to Reshuffle.
#' wTO.Complete( k =1, n = 1000, Data = Microarray_Expression1,
#' Overlap = ExampleGRF$x, method_resampling = "Reshuffle")
#' # Changing the resampling method to BlockBootstrap, with a lag of 2.
#' row.names(metagenomics_abundance) = metagenomics_abundance$OTU
#' metagenomics_abundance = metagenomics_abundance[,-1]
#' wTO.Complete( k =1, n = 1000, Data = metagenomics_abundance, method = "s",
#' Overlap = row.names(metagenomics_abundance), method_resampling = "BlockBootstrap", lag = 2)
#' wTO.Complete( k =2, n = 1000, Data = Microarray_Expression1, method = "s",
#' Overlap = ExampleGRF$x, method_resampling = "BlockBootstrap", ID = rep(1:9,each = 2))
#' X = wTO.Complete( k =1, n = 1000, Data = Microarray_Expression1,
#' Overlap = ExampleGRF$x, method = "p", plot = FALSE)
#' }
#' @export
wTO.Complete = function(k = 1 ,n = 100, Data , Overlap = row.names(Data),
method = "p", method_resampling = "Bootstrap",
pvalmethod = "BH", savecor = F,
expected.diff = 0.20, lag = NULL, ID = NULL,
normalize = F, plot = T){
N = k
Overlap = unique(as.character(Overlap))
`%ni%` <- Negate(`%in%`)
##### Messages
if(is.numeric(k) == F){
stop("k must be numeric.")
}
if(k <= 0){
stop("k must be greater than 0.")
}
if(is.numeric(n) == F){
stop("n must be numeric.")
}
if(n <= 0){
stop("n must be greater than 0.")
}
if(is.data.frame(Data) == F){
stop("Data must be a data.frame.")
}
if(method %ni% c("s", "spearman", "p", "pearson")){
stop('Method must be: "s", "spearman", "p" or "pearson".')
}
if(method_resampling %ni% c("Bootstrap", "Reshuffle", "BlockBootstrap")){
stop('Method must be: "Bootstrap", "BlockBootstrap" or "Reshuffle".')
}
if(method_resampling %in% "BlockBootstrap"){
if (is.null(lag)&is.null(ID)){
stop('If you want to use the "BlockBootstrap" please give a lag or the indivuals ID.')
}
if(!is.null(lag)&!is.null(ID)){
stop('If you want to use the "BlockBootstrap" please give a lag OR the indivuals ID.')
}
}
if(pvalmethod %ni% c ('holm', 'hochberg', 'hommel', 'bonferroni', 'BH', 'BY', 'fdr', 'none')){
stop("pvalmethod must be: 'holm', 'hochberg', 'hommel', 'bonferroni', 'BH', 'BY', 'fdr' or 'none'")
}
if(normalize %ni% c (T, F)){
stop("normalize must be: TRUE or FALSE.")
}
if(normalize == T){
Data.n = as.data.frame(som::normalize(Data))
row.names(Data.n)= row.names(Data)
Data = Data.n
}
DIM_Overlap = nrow(subset(Data, row.names(Data) %in% Overlap))
if(DIM_Overlap == 0){
stop('There is no overlapping nodes. Please check your input "Overlap"')
}
if(!is.null(DIM_Overlap)){
message(paste('There are',DIM_Overlap, "overlapping nodes,",dim(Data)[1],
"total nodes and" , dim(Data)[2],"individuals." ))
}
message("This function might take a long time to run. Don't turn off the computer.")
PAR = par()
## For the original data
# real_Genes = Data
Saving = CorrelationOverlap(Data = Data, Overlap = Overlap, method = method)
WTO_abs = wTO(A = Saving, sign = "abs")
WTO_sign = wTO(A = Saving, sign = "sign")
Cor_real = wTO.in.line(WTO_sign)
Cor_real_abs = wTO.in.line(WTO_abs)
names(Cor_real) = names(Cor_real_abs) <-c("Node.1", "Node.2", "wTO_0")
idcol = c("Node.1", "Node.2")
rm("WTO_abs")
rm("WTO_sign")
data.table::setkeyv(Cor_real, c("Node.1", "Node.2"))
data.table::setkeyv(Cor_real_abs, c("Node.1", "Node.2"))
Orig = cbind(Rep = 0, Cor_real[Cor_real_abs])
names(Orig)= c("Rep","Node.1", "Node.2", "wTO_sign", "wTO_abs")
reps_rest = n
### If only one node
if ( k == 1){
a = 0
while ( reps_rest > 0){
# message(a)
K = 1:min(N, reps_rest)
# K = 1:n
OUTPUT = lapply(K, wTO.aux.each, Data= Data,
Overlap = Overlap, method = method, ID, lag = lag, method_resampling= method_resampling)
ALL = data.table::rbindlist(OUTPUT, idcol = idcol)
names(ALL) = names(Orig) = c("Rep", "Node.1", "Node.2", "wTO_sign" ,"wTO_abs")
ALL_DT_sig = data.table::dcast(ALL, Node.1 + Node.2 ~ Rep, value.var = "wTO_sign")
ALL_DT_abs = data.table::dcast(ALL, Node.1 + Node.2 ~ Rep, value.var = "wTO_abs")
if ( a == 0){
Ps1 = rowSums(ALL_DT_sig[,-c(1:2)] < Orig$wTO_sign - expected.diff)
Ps2 = rowSums(ALL_DT_sig[,-c(1:2)] > Orig$wTO_sign + expected.diff)
Ps = Ps1 + Ps2
Pa1 = rowSums(ALL_DT_abs[,-c(1:2)] < Orig$wTO_abs - expected.diff)
Pa2 = rowSums(ALL_DT_abs[,-c(1:2)] > Orig$wTO_abs + expected.diff)
Pa = Pa1 + Pa2
TAB_SIGN = as.data.frame(table(unlist(round(ALL_DT_sig[,-c(1:2)], 2))))
TAB_ABS = as.data.frame(table(unlist(round(ALL_DT_abs[,-c(1:2)], 2))))
}
if ( a > 0){
Ps1 = rowSums(ALL_DT_sig[,-c(1:2)] < Orig$wTO_sign - expected.diff)
Ps2 = rowSums(ALL_DT_sig[,-c(1:2)] > Orig$wTO_sign + expected.diff)
Ps = Ps + Ps1 + Ps2
Pa1 = rowSums(ALL_DT_abs[,-c(1:2)] < Orig$wTO_abs - expected.diff)
Pa2 = rowSums(ALL_DT_abs[,-c(1:2)] > Orig$wTO_abs + expected.diff)
Pa = Pa + Pa1 + Pa2
TAB_SIGN_aux = as.data.frame(table(unlist(round(ALL_DT_sig[,-c(1:2)], 2))))
TAB_ABS_aux = as.data.frame(table(unlist(round(ALL_DT_abs[,-c(1:2)], 2))))
TAB_SIGN = plyr::join(TAB_SIGN, TAB_SIGN_aux, by = "Var1")
TAB_SIGN = data.frame(Var1 = TAB_SIGN$Var1,
Sum = rowSums(TAB_SIGN[,-1]))
TAB_ABS = plyr::join(TAB_ABS, TAB_ABS_aux, by = "Var1")
TAB_ABS = data.frame(Var1 = TAB_ABS$Var1,
Sum = rowSums(TAB_ABS[,-1]))
}
rm("ALL_DT_sig", "ALL_DT_abs", "ALL", "OUTPUT")
reps_rest = (reps_rest - N)
a = a +1
}
}
else if ( k > 1){
WTO = new.env()
assign("Data", Data, envir = WTO)
assign("Overlap", Overlap, envir = WTO)
assign("method", method, envir = WTO)
assign("CorrelationOverlap", CorrelationOverlap, envir = WTO)
assign("wTO", wTO, envir = WTO)
assign("wTO.in.line", wTO, envir = WTO)
assign("wTO.aux.each", wTO.aux.each, envir = WTO)
assign("method_resampling", method_resampling, envir = WTO)
assign("sample_ind", sample_ind, envir = WTO)
assign("lag", lag, envir = WTO)
assign("ID", ID, envir = WTO)
cl = parallel::makeCluster(k)
parallel::clusterExport(cl, "Data", envir = WTO)
parallel::clusterExport(cl, "wTO.in.line", envir = WTO)
parallel::clusterExport(cl, "lag", envir = WTO)
parallel::clusterExport(cl, "Overlap", envir = WTO)
parallel::clusterExport(cl, "method", envir = WTO)
parallel::clusterExport(cl, "CorrelationOverlap", envir = WTO )
parallel::clusterExport(cl, "wTO", envir = WTO)
parallel::clusterExport(cl, 'wTO.aux.each', envir = WTO)
parallel::clusterExport(cl, 'method_resampling', envir = WTO)
parallel::clusterExport(cl, 'sample_ind', envir = WTO)
# message("cluster")
# K = 1:n
a = 0
while ( reps_rest > 0){
# message(a)
K = 1:min(N, reps_rest)
OUTPUT = parallel::clusterApply(cl, K, wTO.aux.each , Data= Data,
Overlap = Overlap, ID, lag = lag, method = method, method_resampling= method_resampling)
ALL = data.table::rbindlist(OUTPUT, idcol = idcol)
names(ALL) = names(Orig) = c("Rep", "Node.1", "Node.2", "wTO_sign" ,"wTO_abs")
ALL_DT_sig = data.table::dcast(ALL, Node.1 + Node.2 ~ Rep, value.var = "wTO_sign")
ALL_DT_abs = data.table::dcast(ALL, Node.1 + Node.2 ~ Rep, value.var = "wTO_abs")
if ( a == 0){
Ps = rowSums(ALL_DT_sig[,-c(1:2)] < Orig$wTO_sign - expected.diff) +
rowSums(ALL_DT_sig[,-c(1:2)] > Orig$wTO_sign + expected.diff)
Pa = rowSums(ALL_DT_abs[,-c(1:2)] < Orig$wTO_abs - expected.diff) +
rowSums(ALL_DT_abs[,-c(1:2)] > Orig$wTO_abs + expected.diff)
TAB_SIGN = as.data.frame(table(unlist(round(ALL_DT_sig[,-c(1:2)], 2))))
TAB_ABS = as.data.frame(table(unlist(round(ALL_DT_abs[,-c(1:2)], 2))))
}
if ( a > 0){
Ps = Ps + rowSums(ALL_DT_sig[,-c(1:2)] < Orig$wTO_sign - expected.diff) +
rowSums(ALL_DT_sig[,-c(1:2)] > Orig$wTO_sign + expected.diff)
Pa = Pa + rowSums(ALL_DT_abs[,-c(1:2)] < Orig$wTO_abs - expected.diff) +
rowSums(ALL_DT_abs[,-c(1:2)] > Orig$wTO_abs + expected.diff)
# message(Pa)
# message(Ps)
TAB_SIGN_aux = as.data.frame(table(unlist(round(ALL_DT_sig[,-c(1:2)], 2))))
TAB_ABS_aux = as.data.frame(table(unlist(round(ALL_DT_abs[,-c(1:2)], 2))))
TAB_SIGN = plyr::join(TAB_SIGN, TAB_SIGN_aux, by = "Var1")
TAB_SIGN = data.frame(Var1 = TAB_SIGN$Var1,
Sum = rowSums(TAB_SIGN[,-1]))
TAB_ABS = plyr::join(TAB_ABS, TAB_ABS_aux, by = "Var1")
TAB_ABS = data.frame(Var1 = TAB_ABS$Var1,
Sum = rowSums(TAB_ABS[,-1]))
}
rm("ALL_DT_sig", "ALL_DT_abs", "ALL", "OUTPUT")
reps_rest = (reps_rest - N)
a = a +1
}
parallel::stopCluster(cl)
}
message("Simulations are done.")
message("Computing p-values")
Orig$pval_sig = Ps / n
Orig$pval_abs = Pa / n
if(method_resampling == "Reshuffle"){
Orig$pval_sig = 1- Orig$pval_sig
Orig$pval_abs = 1- Orig$pval_abs
}
Orig$Padj_sig = (stats::p.adjust(Orig$pval_sig, method = pvalmethod))
Orig$Padj_abs = (stats::p.adjust(Orig$pval_abs, method = pvalmethod))
## Running the correlation
if( savecor == T){
Total_Correlation = as.data.frame(stats::cor(t(Data), method = method))
Total_Correlation = wTO.in.line(Total_Correlation)
names(Total_Correlation) = c("Node.1", "Node.2", "Cor")
}
if( savecor == F){
Total_Correlation = NULL
}
TAB_SIGN_aux = as.data.frame(table(round(Orig$wTO_sign,2)))
TAB_ABS_aux = as.data.frame(table(round(Orig$wTO_abs,2)))
TAB_SIGN = plyr::join(TAB_SIGN, TAB_SIGN_aux, by = "Var1")
TAB_ABS = plyr::join(TAB_ABS, TAB_ABS_aux, by = "Var1")
message("Computing cutoffs")
if(plot == TRUE){
graphics::par(mar=c(5.1, 4.1, 4.1, 8.1), xpd=TRUE, mfrow = c(3,2))
}
Cutoffs = Cut.off(TAB_SIGN, "wTO - Resampling", plot = plot)
Cutoffs_abs = Cut.off(TAB_ABS, "|wTO| - Resampling", plot = plot)
Orig = Orig[, -"Rep"]
Orig$wTO_abs = as.numeric(Orig$wTO_abs)
Orig$wTO_sign = as.numeric(Orig$wTO_sign)
Orig$pval_abs = as.numeric(Orig$pval_abs)
Orig$pval_sig = as.numeric(Orig$pval_sig)
Orig$Padj_abs = as.numeric(Orig$Padj_abs)
Orig$Padj_sig = as.numeric(Orig$Padj_sig)
Quantiles = rbind(
Cutoffs$Empirical.Quantile,
Cutoffs$Quantile ,
Cutoffs_abs$Empirical.Quantile,
Cutoffs_abs$Quantile)
row.names(Quantiles) = c( 'Empirical.Quantile',
'Quantile',
'Empirical.Quantile.abs',
'Quantile.abs')
tQ = as.data.frame(t(Quantiles))
output = list(wTO = Orig,
Correlation = Total_Correlation,
Quantiles = Quantiles
)
col = ifelse(Orig$pval_sig < 0.05 & Orig$pval_abs < 0.05, "red",
ifelse(Orig$pval_sig < 0.05, "orange",
ifelse (Orig$pval_abs < 0.05, "yellow", "black")))
if(plot == T){
# par(mar=c(5.1, 4.1, 4.1, 8.1), xpd=TRUE, mfrow = c(3,1))
graphics::plot(Orig$wTO_sign, Orig$wTO_abs, axes = F,
xlab = "|wTO|", ylab = "wTO",
main = "|wTO| vs wTO", pch = ".", xlim = c(-1,1), ylim = c(0,1),
col.main = "steelblue2", col.lab = "steelblue2", col = col)
graphics::axis(1, las = 1, cex.axis = 0.6, col = "steelblue",
col.ticks = "steelblue3", col.axis = "steelblue")
graphics::axis(2, las = 1, cex.axis = 0.6, col = "steelblue",col.ticks = "steelblue3", col.axis = "steelblue")
graphics::legend(c(0.9,0), c ("p-value < 0.05", 'wTO sign & |wTO|',
'wTO sign','|wTO|'),
col = c("transparent","red", "orange", "yellow"), pch = 16, bty = "n",
inset=c(-0.8,0), cex = 0.5 )
graphics::par(xpd=FALSE)
graphics::abline( h = 0, lty = 2, col = "gray50")
graphics::abline(v = 0, lty = 2, col = "gray50")
graphics::plot(Orig$wTO_sign, Orig$pval_sig, axes = F,
xlab = "wTO", ylab = "p-value", ylim = c(0,1), xlim = c(-1,1), col.main = "steelblue2", col.lab = "steelblue2",
main = "wTO vs p-value",
pch = 16)
graphics::axis(1, las = 1, cex.axis = 0.6, col = "steelblue",
col.ticks = "steelblue3", col.axis = "steelblue")
graphics::axis(2, las = 1, cex.axis = 0.6, col = "steelblue",col.ticks = "steelblue3", col.axis = "steelblue")
graphics::par(xpd=FALSE)
graphics::abline( v = tQ$Empirical.Quantile, lty = 2, col = c("red", "orange", "yellow", "yellow", "orange", "red"))
graphics::par(xpd=T)
graphics::legend(c(0.9,0), c ("Empirical Quantiles", '0.1%','2.5%','10%','90%','97.5%','99.9%'),
col = c("white", "red", "orange", "yellow", "yellow", "orange", "red"), lwd = 2, bty = "n",
inset=c(-0.8,0), cex = 0.5 )
graphics::par(xpd=FALSE)
graphics::plot(Orig$wTO_abs, Orig$pval_abs, axes = F,
xlab = "|wTO|", ylab = "p-value", ylim = c(0,1), xlim = c(0,1),
main = "|wTO| vs p-value",
pch = 16, col.main = "steelblue2", col.lab = "steelblue2")
graphics::axis(1, las = 1, cex.axis = 0.6, col = "steelblue",
col.ticks = "steelblue3", col.axis = "steelblue")
graphics::axis(2, las = 1, cex.axis = 0.6, col = "steelblue",col.ticks = "steelblue3", col.axis = "steelblue")
graphics::abline( v = tQ$Empirical.Quantile.abs, lty = 2, col = c("red", "orange", "yellow", "yellow", "orange", "red"))
graphics::par(xpd=T)
graphics::legend(c(0.9,0), c ("Empirical Quantiles", '0.1%','2.5%','10%','90%','97.5%','99.9%'),
col = c("white", "red", "orange", "yellow", "yellow", "orange", "red"), lwd = 2, bty = "n",
inset=c(-0.8,0), cex = 0.5 )
}
class(output)<- append('wTO', class(output))
message("Done!")
return(output)
}