-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathMR_functions_v1.0.R
303 lines (280 loc) · 9.56 KB
/
MR_functions_v1.0.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
MRHCA <- function(cor_M, MR_E, step_size0 = 50, p_sig0 = 0.05, Hit_score_cutoff = 100) {
cor_c <- cor_M
diag(cor_c) <- 0
cor_c <- abs(cor_c)
cor_r_rank1 <- apply(cor_c, 1, rank)
cor_r_rank2 <- t(apply(cor_c, 2, rank))
cor_r_rank1 <- nrow(cor_r_rank1) - cor_r_rank1 + 1
cor_r_rank2 <- nrow(cor_r_rank2) - cor_r_rank2 + 1
MR_M <- cor_r_rank1 * cor_r_rank2
ddd <- MR_E
null_growth_rate <- c()
step_size <- step_size0 #parameter
print("Generating Empirical Null Growth Rate")
for (i in 1:nrow(ddd)) {
x0 <- sqrt(sort(ddd[i,]))
null_growth_rate <- rbind(null_growth_rate, calculate_growth_rate(x0, step = step_size))
}
print("Hub Identification")
hub_info_all <- c()
sig_N <- ncol(null_growth_rate) #parameter
p_sig <- p_sig0 #parameter
hub_sig_cut <- Hit_score_cutoff #parameter
for (i in 1:nrow(MR_M)) {
x0 <- sqrt(sort(MR_M[i,]))
tg_growth_rate <- calculate_growth_rate(x0, step = 50)
is_sig <- growth_rate_significance(tg_growth_rate, null_growth_rate, sig_N) < p_sig
S_all <- growth_rate_scoring(is_sig)
cluster_bound <- 0
if_hub <- 0
if ((max(S_all) > hub_sig_cut) & (sum(tg_growth_rate[1:50] > 0.7) == 0)) {
cluster_bound <- max(which(S_all == max(S_all)))
if_hub <- 1
}
hub_info <- c(if_hub, cluster_bound)
hub_info_all <- rbind(hub_info_all, hub_info)
}
rownames(hub_info_all) <- rownames(MR_M)
return(list(hub_info_all, MR_M))
}
Empirical_null_distribution_MR <- function(data, Rounds = 1000) {
N <- nrow(data)
aaa <- runif(N * Rounds, 0, 1)
bbb1 <- rep(0, length(aaa))
bbb2 <- rep(0, length(aaa))
for (i in 1:length(aaa)) {
bbb1[i] <- rbinom(1, N - 1, aaa[i])
bbb2[i] <- rbinom(1, N - 1, aaa[i])
}
ccc <- (bbb1 + 1) * (bbb2 + 1)
ddd <- matrix(ccc, Rounds, N)
return(ddd)
}
growth_rate_scoring <- function(is_sig, up_score = 1, down_score = -2, delete_top = 10) {
S <- 0
S_all <- c()
for (i in 1:length(is_sig)) {
if (i < delete_top) {
S <- S + 0
S_all <- c(S_all, S)
}
else {
if (is_sig[i] == 1) {
S <- S + up_score
S_all <- c(S_all, S)
}
if (is_sig[i] == 0) {
S <- S + down_score
S_all <- c(S_all, S)
}
}
}
return(S_all)
}
growth_rate_significance <- function(tg_growth_rate, null_growth_rate, sig_N) {
growth_rate_sig <- rep(0, sig_N)
for (i in 1:sig_N) {
growth_rate_sig[i] <- sum(tg_growth_rate[i] >= null_growth_rate[, i])
}
growth_rate_sig <- growth_rate_sig / nrow(null_growth_rate)
return(growth_rate_sig)
}
calculate_growth_rate <- function(x0, step = 20) {
gr_all <- rep(0, length(x0))
for (i in 1:length(x0)) {
sid <- max(1, i - step)
eid <- min(length(x0), i + step)
tg_id <- c(sid:eid)
l <- eid - sid + 1
gr_all[i] <- (x0[eid] - x0[sid]) / l
}
return(gr_all)
}
calculate_MR <- function(tg_id, Rank_index) {
N <- nrow(Rank_index)
K <- ncol(Rank_index)
MR_result <- rep(0, N)
for (i in 1:N) {
R1 <- K + 1
R2 <- K + 1
if (sum(tg_id == Rank_index[i, ]) > 0) {
R1 <- which(Rank_index[i,] == tg_id)
}
if (sum(i == Rank_index[tg_id, ]) > 0) {
R2 <- which(Rank_index[tg_id,] == i)
}
MR_result[i] <- R1 * R2
#if(i%%1000==1)
#{
#print(i)
#}
}
MR_result[tg_id] <- K ^ 2
return(MR_result)
}
MRHCA_small_hub_identificaiton <- function(MRHCA_output, data, MR_E) {
hubs_small_result <- MRHCA_output
data_c <- data
MR_M <- hubs_small_result[[2]]
MR_M0 <- MR_M
for (i in 1:nrow(MR_M0)) {
MR_M0[i,] <- sort(MR_M0[i,])
}
############
MR_E0 <- MR_E
for (i in 1:nrow(MR_E0)) {
MR_E0[i,] <- sort(MR_E0[i,])
}
#############
ccc <- hubs_small_result[[1]]
tg_cut_id <- 100
tg_hubs <- c()
if (sum(ccc[, 1] == 1) > 0) {
tg_cut_id <- max(min(ccc[which(ccc[, 1] == 1), 2]), 100)
tg_hubs <- names(which(hubs_small_result[[1]][, 1] == 1))
}
print("Cut Smallest Module Size!!!\n")
print(tg_cut_id)
MR_M_c <- sqrt(MR_M0[, 1:tg_cut_id])
MR_E_c <- sqrt(MR_E0[, 1:tg_cut_id])
p_cut <- 0.01
stat_MR <- MR_M_c * 0
for (i in 1:ncol(stat_MR)) {
stat_MR[, i] <- MR_M_c[, i] < quantile(MR_E_c[, i], p_cut)
}
cumsum_MR <- stat_MR * 0
for (i in 1:nrow(stat_MR)) {
cumsum_MR[i,] <- cumsum(stat_MR[i,])
}
for (i in 1:nrow(cumsum_MR)) {
if (length(which(cumsum_MR[i, ] != 0)) == 0) {
fff <- rep(1, length(cumsum_MR[i,]))
}
else {
tg_ids <- min(which(cumsum_MR[i,] != 0))
fff <- c(rep(1, tg_ids - 1), 1:(length(cumsum_MR[i,]) - tg_ids + 1))
}
cumsum_MR[i,] <- cumsum_MR[i,] / fff
}
#intersect(names(which(apply(cumsum_MR>0.8,1,sum)>5)),tg_hubs)
tg_id_ccc <- setdiff(names(which(apply(cumsum_MR > 0.8, 1, sum) > 2)), tg_hubs)
small_hub_info <- c()
small_hubs <- c()
small_hub_info_selected <- c()
if (length(tg_id_ccc) > 0) {
for (i in 1:length(tg_id_ccc)) {
x0 <- sqrt(sort(MR_M[tg_id_ccc[i],]))
tg_growth_rate <- calculate_growth_rate(x0, step = 10)
ccc <- cumsum(tg_growth_rate < 0.5) - cumsum(tg_growth_rate > 1) * 100 - cumsum(tg_growth_rate > 0.85) * 5
ddd <- ccc / c(1:length(ccc))
if ((sum(ccc[1:10] < 0) == 0) & (max(ccc) > 0.8)) {
if (sum(ddd < 0) > 0) {
small_hub_info <- rbind(small_hub_info, c(1, min(which(ddd < 0))))
}
else {
small_hub_info <- rbind(small_hub_info, c(1, ncol(MR_M)))
}
small_hubs <- c(small_hubs, tg_id_ccc[i])
}
}
rownames(small_hub_info) <- small_hubs
small_hub_info_selected <- small_hub_info[which(small_hub_info[, 2] <= 300),]
if (length(which(small_hub_info[, 2] <= 300)) == 1) {
tg_id <- rownames(small_hub_info)[which(small_hub_info[, 2] <= 300)]
small_hub_info_selected <- t(as.matrix(small_hub_info_selected))
rownames(small_hub_info_selected) <- tg_id
}
print("small hub selected: ")
print(length(which(small_hub_info[, 2] <= 300)))
print("small hub removed: ")
print(length(which(small_hub_info[, 2] > 300)))
#hist(small_hub_info[,2],breaks=100)
}
return(small_hub_info_selected)
}
########################
#Fast version functions
########################
Empirical_null_distribution_MR_F <- function(data, K = 1000, Rounds = 1000) {
M<-K
N <- nrow(data)
aaa <- runif(N * Rounds, 0, 1)
bbb1 <- rep(0, length(aaa))
bbb2 <- rep(0, length(aaa))
for (i in 1:length(aaa)) {
bbb1[i] <- rbinom(1, N - 1, aaa[i])
bbb2[i] <- rbinom(1, N - 1, aaa[i])
}
bbb1[which(bbb1 > (M))] <- M
bbb2[which(bbb2 > (M))] <- M
ccc <- (bbb1 + 1) * (bbb2 + 1)
ddd <- matrix(ccc, Rounds, N)
return(ddd)
}
Compute_rank_index <- function(data, K) {
Rank_index <- c()
for (i in 1:nrow(data)) {
cor_cc <- cor(t(data), data[i,])
cor_cc[i] <- 0
Rank_index <- rbind(Rank_index, order(-abs(cor_cc))[1:K])
}
rownames(Rank_index) <- rownames(data)
return(Rank_index)
}
MRHCA_fast_version <- function(Rank_index, MR_E, K = 500, step_size0 = 50, p_sig0 = 0.05, Hit_score_cutoff = 100, TN_p = 1) {
sig_N0 <- floor(K * 0.75)
hub_sig_cut0 <- Hit_score_cutoff
MR_id <- 1:nrow(Rank_index)
if (TN_p < 1) {
MR_id <- names(sort(table(Rank_index), decreasing = T))
MR_id <- as.numeric(MR_id)
}
MR_EM <- c()
TN <- floor(nrow(Rank_index) * TN_p)
rnames <- c()
print("Compute MR")
for (i in 1:TN) {
tg_id <- MR_id[i]
a1 <- calculate_MR(tg_id, Rank_index)
rnames <- c(rnames, tg_id)
MR_EM <- rbind(MR_EM, a1)
}
rownames(MR_EM) <- rnames
colnames(MR_EM) <- c(1:length(a1))
#print("Compute Empirical Null Distribution")
####calculate null distribution of MR
ddd2 <- MR_E
####calculate null distribution of MR growth rate
null_growth_rate <- c()
step_size <- step_size0 #parameter
for (i in 1:nrow(ddd2)) {
x0 <- sqrt(sort(ddd2[i,]))
null_growth_rate <- rbind(null_growth_rate, calculate_growth_rate(x0, step = step_size))
}
print("Hub Identification")
hub_info_all <- c()
sig_N <- sig_N0 #parameter
p_sig <- p_sig0 #parameter
hub_sig_cut <- hub_sig_cut0 #parameter
for (i in 1:nrow(MR_EM)) {
x0 <- sqrt(sort(MR_EM[i,]))
#x0<-sqrt(sort(calculate_MR(i,Rank_index)))
tg_growth_rate <- calculate_growth_rate(x0, step = 50)
is_sig <- growth_rate_significance(tg_growth_rate, null_growth_rate, sig_N) < p_sig
S_all <- growth_rate_scoring(is_sig)
cluster_bound <- 0
if_hub <- 0
if ((max(S_all) > hub_sig_cut) & (sum(tg_growth_rate[1:50] > 0.7) == 0)) {
cluster_bound <- max(which(S_all == max(S_all)))
if_hub <- 1
}
hub_info <- c(if_hub, cluster_bound, 0)
if (cluster_bound == sig_N) {
hub_info[3] <- max(which(tg_growth_rate != 0))
}
hub_info_all <- rbind(hub_info_all, hub_info)
}
hub_info_all[which(hub_info_all[, 2] == sig_N), 2] <- 0
rownames(hub_info_all) <- rownames(MR_EM)
return(list(hub_info_all, MR_EM))
}