-
Notifications
You must be signed in to change notification settings - Fork 14
/
Copy pathrun_ensemble_clustering.R
164 lines (123 loc) · 6.83 KB
/
run_ensemble_clustering.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
#########################################################################################
# R script to run ensemble clustering
#
# Lukas Weber, September 2016
#########################################################################################
library(clue)
# load clustering results (runtime: 15 mins)
source("../evaluate_results/evaluate_ClusterX.R")
source("../evaluate_results/evaluate_DensVM.R")
source("../evaluate_results/evaluate_FLOCK.R")
source("../evaluate_results/evaluate_flowClust.R")
source("../evaluate_results/evaluate_flowMeans.R")
source("../evaluate_results/evaluate_flowPeaks.R")
source("../evaluate_results/evaluate_FlowSOM.R")
source("../evaluate_results/evaluate_immunoClust.R")
source("../evaluate_results/evaluate_kmeans.R")
source("../evaluate_results/evaluate_PhenoGraph.R")
source("../evaluate_results/evaluate_Rclusterpp.R")
source("../evaluate_results/evaluate_SamSPECTRAL.R")
source("../evaluate_results/evaluate_SPADE.R")
source("../evaluate_results/evaluate_SWIFT.R")
source("../evaluate_results/evaluate_Xshift.R")
###########################
### ENSEMBLE CLUSTERING ###
###########################
# calculate ensemble clustering using all methods for each data set, excluding the
# following:
# - methods that use subsampled data, since consensus clustering requires same data for
# each method (see parameters spreadsheet)
# - methods that remove outliers, since consensus clustering requires same data for each
# method (SamSPECTRAL for all data sets; flowClust for Nilsson_rare and Mosmann_rare;
# SWIFT for data sets with multiple populations; X-shift for all data sets except
# Samusik_01 and Nilsson_rare)
# - methods that give a large number of small clusters (FlowSOM_pre; SWIFT for data sets
# with multiple populations), since this greatly slows down runtime
# create partitions in format required by CLUE
partition_Levine_32dim <- list(as.cl_partition(clus_FLOCK[[1]]),
as.cl_partition(clus_flowMeans[[1]]),
as.cl_partition(clus_flowPeaks[[1]]),
as.cl_partition(clus_FlowSOM[[1]]),
as.cl_partition(clus_kmeans[[1]]),
as.cl_partition(clus_PhenoGraph[[1]]),
as.cl_partition(clus_Rclusterpp[[1]]))
partition_Levine_13dim <- list(as.cl_partition(clus_ClusterX[[2]]),
as.cl_partition(clus_FLOCK[[2]]),
as.cl_partition(clus_flowMeans[[2]]),
as.cl_partition(clus_flowPeaks[[2]]),
as.cl_partition(clus_FlowSOM[[2]]),
as.cl_partition(clus_immunoClust[[2]]),
as.cl_partition(clus_kmeans[[2]]),
as.cl_partition(clus_PhenoGraph[[2]]),
as.cl_partition(clus_Rclusterpp[[2]]),
as.cl_partition(clus_SPADE[[2]]))
partition_Samusik_01 <- list(as.cl_partition(clus_ClusterX[[3]]),
as.cl_partition(clus_DensVM[[3]]),
as.cl_partition(clus_FLOCK[[3]]),
as.cl_partition(clus_flowMeans[[3]]),
as.cl_partition(clus_flowPeaks[[3]]),
as.cl_partition(clus_FlowSOM[[3]]),
as.cl_partition(clus_immunoClust[[3]]),
as.cl_partition(clus_kmeans[[3]]),
as.cl_partition(clus_PhenoGraph[[3]]),
as.cl_partition(clus_Rclusterpp[[3]]),
as.cl_partition(clus_SPADE[[3]]),
as.cl_partition(clus_Xshift[[3]]))
partition_Samusik_all <- list(as.cl_partition(clus_FLOCK[[4]]),
as.cl_partition(clus_flowPeaks[[4]]),
as.cl_partition(clus_FlowSOM[[4]]),
as.cl_partition(clus_kmeans[[4]]),
as.cl_partition(clus_PhenoGraph[[4]]),
as.cl_partition(clus_SPADE[[4]]))
partition_Nilsson_rare <- list(as.cl_partition(clus_ClusterX[[5]]),
as.cl_partition(clus_DensVM[[5]]),
as.cl_partition(clus_FLOCK[[5]]),
as.cl_partition(clus_flowMeans[[5]]),
as.cl_partition(clus_flowPeaks[[5]]),
as.cl_partition(clus_FlowSOM[[5]]),
as.cl_partition(clus_immunoClust[[5]]),
as.cl_partition(clus_kmeans[[5]]),
as.cl_partition(clus_PhenoGraph[[5]]),
as.cl_partition(clus_Rclusterpp[[5]]),
as.cl_partition(clus_SPADE[[5]]),
as.cl_partition(clus_SWIFT[[5]]),
as.cl_partition(clus_Xshift[[5]]))
partition_Mosmann_rare <- list(as.cl_partition(clus_FLOCK[[6]]),
as.cl_partition(clus_flowMeans[[6]]),
as.cl_partition(clus_flowPeaks[[6]]),
as.cl_partition(clus_FlowSOM[[6]]),
as.cl_partition(clus_immunoClust[[6]]),
as.cl_partition(clus_kmeans[[6]]),
as.cl_partition(clus_PhenoGraph[[6]]),
as.cl_partition(clus_SPADE[[6]]),
as.cl_partition(clus_SWIFT[[6]]))
# create cluster ensembles
partitions <- list(partition_Levine_32dim,
partition_Levine_13dim,
partition_Samusik_01,
partition_Samusik_all,
partition_Nilsson_rare,
partition_Mosmann_rare)
ensembles <- lapply(partitions, function(p) cl_ensemble(list = p))
# calculate consensus clustering (runtime: 35 mins)
consensus <- vector("list", length(ensembles))
for (i in 1:length(consensus)) {
set.seed(123)
consensus[[i]] <- cl_consensus(ensembles[[i]])
}
# get class IDs
clus_consensus <- lapply(consensus, cl_class_ids)
####################
### SAVE RESULTS ###
####################
# save cluster labels
datasets <- c("Levine_32dim", "Levine_13dim", "Samusik_01", "Samusik_all", "Nilsson_rare", "Mosmann_rare")
files_out <- paste0("../../results/ensemble/ensemble_labels_", datasets, ".txt")
for (i in 1:length(files_out)) {
res <- data.frame(label = as.numeric(clus_consensus[[i]]))
write.table(res, file = files_out[i], row.names = FALSE, quote = FALSE, sep = "\t")
}
# save session information
sink(file = "../../results/ensemble/session_info_ensemble.txt")
sessionInfo()
sink()