-
Notifications
You must be signed in to change notification settings - Fork 0
/
sim_modules.R
434 lines (345 loc) · 14.5 KB
/
sim_modules.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
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
library(genefilter)
library(MASS)
library(pROC)
library(docopt)
library(bayesbio)
setwd("/Users/amckenz/Dropbox/zhang/diffcorr")
###############################
# steps to perform in the program
test = FALSE
# hardcode the parameters during a test
if(test){
n_samples = 200
fraction_pairs = 1
seed = floor(runif(1, 0, 100)) #42
ddcor_test = TRUE
DiffCoEx_test = TRUE
dicer_test = FALSE
}
#get in parameters from the command line when in "production" use
if(!test){
args = commandArgs(trailingOnly = TRUE)
print(args)
n_samples = as.numeric(args[1])
if(args[2] == "random"){
seed = runif(1, 0, 100000)
print("seed")
print(seed)
} else {
seed = as.numeric(args[2])
}
fraction_pairs = as.numeric(args[3])
ddcor_test = as.logical(args[4])
DiffCoEx_test = as.logical(args[5])
dicer_test = as.logical(args[6])
}
corr_classes = "strong"
#reproducibility
set.seed(seed)
#################################
# functions
#sets the non-diagonals of a matrix to the number x.
non_diag <- function(matrix, x){
for(i in 1:nrow(matrix)){
for(j in 1:ncol(matrix)){
if(j >= i){
next
} else {
matrix[i, j] = sample(c(0, x), 1, prob = c(1-fraction_pairs, fraction_pairs))
}
}
}
#make the upper triangle the same as the lower triangle
matrix = makeMatSym(matrix, replaceUpper = FALSE)
return(matrix)
}
#changes the correlation structure of a particular submatrix of a larger matrix
add_sigma_to_matrix <- function(total_matrix, row_start, nrows, x){
extract_mat = total_matrix[row_start:(row_start + nrows),
row_start:(row_start + nrows)]
mat = non_diag(extract_mat, x)
total_matrix[row_start:(row_start + nrows),
row_start:(row_start + nrows)] = mat
return(total_matrix)
}
#find gene pair names in the relevant submatrix
find_gene_pair_names <- function(total_dc_genes, n_genes){
res = apply(combn(letters_unique[(total_dc_genes + 1):(total_dc_genes + n_genes + 1)], m = 2), 2, paste, collapse = " ")
return(res)
}
#updates the corrrelation matrix and adds the gene pairs that were modified to a class for subsequent extraction
update_corr_structure <- function(sigma, n_genes, variance, corr_factor,
dcor = NULL, dcor_specific = NULL){
total_dc_genes = sigma[["total_dc_genes"]]
sigma[["sigma_tot"]] = add_sigma_to_matrix(sigma[["sigma_tot"]],
row_start = total_dc_genes + 1,
nrows = n_genes, x = variance * corr_factor)
if(!is.null(dcor)){
if(dcor == "goc"){
sigma[["goc_gene_pairs"]] = c(sigma[["goc_gene_pairs"]],
find_gene_pair_names(total_dc_genes, n_genes))
}
if(dcor == "loc"){
sigma[["loc_gene_pairs"]] = c(sigma[["loc_gene_pairs"]],
find_gene_pair_names(total_dc_genes, n_genes))
}
}
if(!is.null(dcor_specific)){
if(dcor_specific == "+/0"){
sigma[["pos_none_gene_pairs"]] = c(sigma[["pos_none_gene_pairs"]],
find_gene_pair_names(total_dc_genes, n_genes))
}
if(dcor_specific == "+/-"){
sigma[["pos_neg_gene_pairs"]] = c(sigma[["pos_neg_gene_pairs"]],
find_gene_pair_names(total_dc_genes, n_genes))
}
if(dcor_specific == "0/+"){
sigma[["none_pos_gene_pairs"]] = c(sigma[["none_pos_gene_pairs"]],
find_gene_pair_names(total_dc_genes, n_genes))
}
if(dcor_specific == "0/-"){
sigma[["none_neg_gene_pairs"]] = c(sigma[["none_neg_gene_pairs"]],
find_gene_pair_names(total_dc_genes, n_genes))
}
if(dcor_specific == "-/+"){
sigma[["neg_pos_gene_pairs"]] = c(sigma[["neg_pos_gene_pairs"]],
find_gene_pair_names(total_dc_genes, n_genes))
}
if(dcor_specific == "-/0"){
sigma[["neg_none_gene_pairs"]] = c(sigma[["neg_none_gene_pairs"]],
find_gene_pair_names(total_dc_genes, n_genes))
}
}
sigma[["total_dc_genes"]] = total_dc_genes + n_genes
return(sigma)
}
#for gene name generation, i.e. "aa", "ab", "ac", etc
combine <- function (vecs, number){
combn(vecs, number, paste, collapse = "")
}
letters_unique = combine(c(letters[1:26], letters[1:26]), 2)
letters_unique = sort(unique(letters_unique))
#################################
# parameters for simulation data
n_non_expr_genes = 200
n_housekeeping_genes = 100
n_activated_genes = 300
n_tot_genes = n_non_expr_genes + n_housekeeping_genes + n_activated_genes
low_mean = 2
high_mean = 50
low_var = 50
high_var = 100
positive_correlation = 0.9
negative_correlation = 0 #if this is set too low, then the matrix will be non-positive definite
######################################
#create overall correlation matrices and classes to store them
sigma_tot_a <- structure(list(
sigma_tot = matrix(0, nrow = n_tot_genes, ncol = n_tot_genes),
total_dc_genes = 0
), class = "sigma")
sigma_tot_b <- structure(list(
sigma_tot = matrix(0, nrow = n_tot_genes, ncol = n_tot_genes),
total_dc_genes = 0,
loc_gene_pairs = vector(),
goc_gene_pairs = vector(),
pos_none_gene_pairs = vector(),
pos_neg_gene_pairs = vector(),
none_pos_gene_pairs = vector(),
none_neg_gene_pairs = vector(),
neg_pos_gene_pairs = vector(),
neg_none_gene_pairs = vector()
), class = "sigma")
if(corr_classes == "strong"){
#A+, B- = loss of correlation
n_apbm = 29
sigma_tot_a = update_corr_structure(sigma = sigma_tot_a, n_genes = n_apbm,
variance = high_var, corr_factor = positive_correlation, dcor = NULL)
sigma_tot_b = update_corr_structure(sigma = sigma_tot_b, n_genes = n_apbm,
variance = high_var, corr_factor = negative_correlation, dcor = "loc",
dcor_specific = "+/-")
#A-, B+ = gain of correlation
n_ambp = 29
sigma_tot_a = update_corr_structure(sigma = sigma_tot_a, n_genes = n_ambp,
variance = high_var, corr_factor = negative_correlation, dcor = NULL)
sigma_tot_b = update_corr_structure(sigma = sigma_tot_b, n_genes = n_ambp,
variance = high_var, corr_factor = positive_correlation, dcor = "goc",
dcor_specific = "-/+")
}
sigma_tot_a_mat = sigma_tot_a[["sigma_tot"]]
sigma_tot_b_mat = sigma_tot_b[["sigma_tot"]]
########################################
#set the means and variances of the genes
#set the variance
var_diag = c(rep(high_var, n_activated_genes),
rep(low_var, n_housekeeping_genes),
rep(high_var, n_non_expr_genes))
#set the mean for the high expression genes
mu = rnbinom(n = (n_activated_genes + n_housekeeping_genes), mu = high_mean, size = 0.5)
#set the mean for the low expression genes
mu = c(mu, rnbinom(n = n_non_expr_genes, mu = low_mean, size = 0.5))
#if mu is less than one, set it to one
mu[mu < 1] = 1
diag(sigma_tot_a_mat) = var_diag
diag(sigma_tot_b_mat) = var_diag
#generate the simulation matrix
sim_data_a = as.matrix(mvrnorm(n_samples, mu = mu, Sigma = sigma_tot_a_mat, tol = 0.4))
sim_data_b = as.matrix(mvrnorm(n_samples, mu = mu, Sigma = sigma_tot_b_mat, tol = 0.4))
true_cases = c(sigma_tot_b$loc_gene_pairs, sigma_tot_b$goc_gene_pairs)
####################################
#run the ddcor differential correlation pipeline on the simulated data
if(ddcor_test){
library(DGCA)
print("starting the ddcor pipeline")
sim_data_a = t(sim_data_a)
sim_data_b = t(sim_data_b)
sim_data_merge = cbind(sim_data_a, sim_data_b)
rownames(sim_data_merge) = letters_unique[1:nrow(sim_data_merge)]
conditions = c(rep("cond_a", ncol(sim_data_a)),
rep("cond_b", ncol(sim_data_b)))
design_mat = model.matrix(~ conditions + 0)
labels = c("cond_a", "cond_b")
colnames(design_mat) = labels
npairs = (nrow(sim_data_merge)^2)/2 - nrow(sim_data_merge)
ddcor_start = Sys.time()
ddcor_res = ddcorAll(inputMat = sim_data_merge, design = design_mat, compare = labels, corr_cutoff = 0.95, adjust = "perm", corrType = "pearson", nPerm = 10, heatmapPlot = FALSE, nPairs = "all")
print(paste0("time taken to run the ddcor pipeline with ", corr_classes, n_samples, " samples..."))
ddcor_time = difftime(ddcor_start, Sys.time(), units = "secs")
print(difftime(ddcor_start, Sys.time()))
ddcor_res$combos = paste(ddcor_res$Gene1, ddcor_res$Gene2, sep = " ")
ddcor_res$true = ddcor_res$combos %in% true_cases
library(MEGENA)
library(doMC)
n.perm = 100
hub.pval = 0.05
module.pval = 1
min.size = 15
max.size = 1000
n.cores = 4
save.output = FALSE
doPar = TRUE
ddcor_res_sig_05 = ddcor_res[ddcor_res$pValDiff_adj < 0.05, ]
ddcor_res_megena = ddcor_res_sig_05[ , colnames(ddcor_res_sig_05) %in% c("Gene1", "Gene2", "zScoreDiff")]
ddcor_res_megena$zScoreDiff = abs(ddcor_res_megena$zScoreDiff)
pfn_res = MEGENA::calculate.PFN(ddcor_res_megena, doPar = doPar, num.cores = n.cores)
#normalize to follow the convention that weights are in [0, 1]
pfn_res$weight = (pfn_res$weight/max(pfn_res$weight)) * 0.999999999
g = igraph::graph.data.frame(pfn_res, directed = FALSE)
MEGENA.output = MEGENA::do.MEGENA(g, mod.pval = module.pval, hub.pval = hub.pval, remove.unsig = TRUE,
min.size = min.size, max.size = max.size, doPar = doPar, num.cores = n.cores, n.perm = n.perm,
save.output = save.output)
MEGENA_modules = MEGENA.output$module.outpu$modules
dgcamegena_modules =
data.frame(Genes = unlist(MEGENA_modules), Modules = rep(names(MEGENA_modules), sapply(MEGENA_modules, length)))
if(is.null(dgcamegena_modules)){
grey_module = data.frame(Genes = rownames(sim_data_merge), Modules = rep("grey", nrow(sim_data_merge)))
dgcamegena_modules = grey_module
} else {
genes_in_grey_module = rownames(sim_data_merge)[!rownames(sim_data_merge) %in% dgcamegena_modules$Genes]
grey_module = data.frame(Genes = genes_in_grey_module, Modules = rep("grey", length(genes_in_grey_module)))
dgcamegena_modules = rbind(dgcamegena_modules, grey_module)
}
}
###################################
#run the DiffCoEx differential correlation pipeline on the simulated data
if(DiffCoEx_test){
print("starting the DiffCoEx pipeline")
library(flashClust)
beta1 = 6 #user defined parameter for soft thresholding
AdjMatC1 = sign(cor(t(sim_data_a), method="spearman"))*(cor(t(sim_data_a), method="spearman"))^2
AdjMatC2 = sign(cor(t(sim_data_b), method="spearman"))*(cor(t(sim_data_b), method="spearman"))^2
diag(AdjMatC1) = 0
diag(AdjMatC2) = 0
collectGarbage()
dissTOMC1C2 = TOMdist((abs(AdjMatC1-AdjMatC2)/2)^(beta1/2))
collectGarbage()
geneTreeC1C2 = flashClust(as.dist(dissTOMC1C2), method = "average");
dynamicModsHybridC1C2 = cutreeDynamic(dendro = geneTreeC1C2, distM = dissTOMC1C2,method="hybrid",cutHeight=0.996,deepSplit = T, pamRespectsDendro = FALSE, minClusterSize = 15);
dynamicColorsHybridC1C2 = labels2colors(dynamicModsHybridC1C2)
#module merging fails if all of the colors are "grey" (i.e., there is only one module), so avoid this possibility
if(any(!dynamicColorsHybridC1C2 == "grey")){
#the next step merges clusters which are close (see WGCNA package documentation)
mergedColorC1C2 = mergeCloseModules(rbind(t(sim_data_a),t(sim_data_b)),dynamicColorsHybridC1C2,cutHeight=.2)$color
} else {
mergedColorC1C2 = dynamicColorsHybridC1C2
}
#We write each module to an individual file containing affymetrix probeset IDs
diffcoex_modules =
data.frame(Genes = letters_unique[1:nrow(sim_data_a)], Modules = mergedColorC1C2)
}
###########################################
if(dicer_test){
print("starting the dicer pipeline")
library(bayesbio)
ID = 1:ncol(sim_data_merge)
sim_df = rbind(ID, sim_data_merge)
write.delim(sim_df,
paste0("dicer_sim/sim_gnxp.tsv"), row.names = TRUE, col.names = FALSE)
sim_classes = data.frame(a = c(0, ncol(sim_data_a)), b = c(ncol(sim_data_a) + 1, (ncol(sim_data_a) + ncol(sim_data_b))), c = c("cond_a", "cond_b"))
write.delim(sim_classes, "dicer_sim/sim_classes.tsv", col.names = FALSE)
#dicer.jar needs to be in the same directory
system("java -jar dicer.jar dicer_sim/sim_gnxp.tsv dicer_sim/sim_classes.tsv 0 dicer_sim/sim_res.txt")
dicer_modules = tryCatch(read.table("dicer_sim/sim_res.txt", header = FALSE), error=function(e) NULL)
if(!is.null(dicer_modules)){
colnames(dicer_modules) = c("Genes", "Modules")
dicer_line = c(seed, fraction_pairs, n_samples, nrow(dicer_modules))
} else {
print("no dicer modules detected")
dicer_line = c(seed, fraction_pairs, n_samples, 0)
genes_in_grey_module = rownames(sim_data_merge)
grey_module = data.frame(Genes = genes_in_grey_module, Modules = rep("grey", length(genes_in_grey_module)))
dicer_modules = rbind(grey_module)
}
write.table(as.matrix(t(dicer_line)), file = "dicer_sim/dicer_sim_module_res.txt",
sep = "\t", append = TRUE, col.names = FALSE, row.names = FALSE)
}
#############################
# compare the modules detected
module_goc = unique(unlist(strsplit(sigma_tot_b$goc_gene_pairs, " ")))
module_loc = unique(unlist(strsplit(sigma_tot_b$loc_gene_pairs, " ")))
#find the modules that best overlap for each, based on jaccard index
calculate_module_jaccard <- function(modules, true_module){
unique_module_names = unique(modules$Modules)
jaccard_stats_true_module = vector()
for(i in 1:length(unique_module_names)){
jaccard_stats_true_module[i] = jaccardSets(true_module,
modules[modules$Modules == unique_module_names[i], "Genes"])
}
modules_max_jaccard = max(jaccard_stats_true_module)
return(modules_max_jaccard)
}
calculate_module_sensitivity <- function(modules, true_module){
unique_module_names = unique(modules$Modules)
sensitivity_stats_true_module = vector()
for(i in 1:length(unique_module_names)){
if(length(modules[modules$Modules == unique_module_names[i], "Genes"]) > 50){
sensitivity_stats_true_module[i] = 0
} else {
sensitivity_stats_true_module[i] = length(intersect(true_module,
modules[modules$Modules == unique_module_names[i], "Genes"]))/length(true_module)
}
}
modules_max_sensitivity = max(sensitivity_stats_true_module)
return(modules_max_sensitivity)
}
find_true_module_jaccard_sensitivity <- function(modules, true_module, data_set, module_type){
max_jaccard = calculate_module_jaccard(modules, true_module)
line_out = c(seed, fraction_pairs, n_samples, data_set, module_type, max_jaccard)
write.table(as.matrix(t(line_out)), file = "module_DC_jaccard.txt",
sep = "\t", append = TRUE, col.names = FALSE, row.names = FALSE, quote = FALSE)
max_sensitivity = calculate_module_sensitivity(modules, true_module)
line_out = c(seed, fraction_pairs, n_samples, data_set, module_type, max_sensitivity)
write.table(as.matrix(t(line_out)), file = "module_DC_sensitivity.txt",
sep = "\t", append = TRUE, col.names = FALSE, row.names = FALSE, quote = FALSE)
}
if(!is.null(dgcamegena_modules)){
find_true_module_jaccard_sensitivity(dgcamegena_modules, module_goc, "DGCA_MEGENA", "GOC")
find_true_module_jaccard_sensitivity(dgcamegena_modules, module_loc, "DGCA_MEGENA", "LOC")
}
if(!is.null(diffcoex_modules)){
find_true_module_jaccard_sensitivity(diffcoex_modules, module_goc, "diffcoex", "GOC")
find_true_module_jaccard_sensitivity(diffcoex_modules, module_loc, "diffcoex", "LOC")
}
if(!is.null(dicer_modules)){
find_true_module_jaccard_sensitivity(dicer_modules, module_goc, "DICER", "GOC")
find_true_module_jaccard_sensitivity(dicer_modules, module_loc, "DICER", "LOC")
}