-
Notifications
You must be signed in to change notification settings - Fork 0
/
12-GRN_final.R
343 lines (265 loc) · 12.8 KB
/
12-GRN_final.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
### Project Setup ==================================================================================
out<-"outputs/12-GRN_final"
dir.create(out)
source("scripts/utils/new_utils.R")
library(Seurat)
library(Signac)
####Functions####
GetMotifIDs<-function(object,motif.names,assay=NULL,return_dt=FALSE){
if(is.null(assay))assay<-DefaultAssay(object)
idx<-match(motif.names,object@assays[[assay]]@motifs@motif.names)
if(return_dt){
return(
data.table(motif.name=motif.names,
motif.id=names(object@assays[[assay]]@motifs@motif.names[idx]))
)
}else{
return(names(object@assays[[assay]]@motifs@motif.names[idx]))
}
}
CheckMotif<-function(object,peaks,motif.name,assay = NULL,return.peaks=FALSE){
require("Signac")
if(is.null(assay))assay<-DefaultAssay(object)
motif<-GetMotifID(object,motif.name,assay=assay)
motif.all <- GetMotifData(
object = object, assay = assay, slot = "data"
)
motifs_peaks_tf <- motif.all[peaks,motif , drop = FALSE]
if(return.peaks){
motifs_peaks_tf<-rownames(motifs_peaks_tf)[as.vector(motifs_peaks_tf==1)]
return(motifs_peaks_tf)
}else{
motifs_peaks_tf_vec<-as.vector(motifs_peaks_tf==1)
names(motifs_peaks_tf_vec)<-rownames(motifs_peaks_tf)
return(motifs_peaks_tf_vec)
}
}
### Analysis =======================================================================================
#clean regulons list based on atac
regulons_list<-readRDS("outputs/09-SCENIC/cbps_14k/regulons_list.rds")
atacs<-readRDS("outputs/07-DMCs_atac_integr/cbps_atacs.rds")
atacs[["lin_peaks"]]<-readRDS("outputs/07-DMCs_atac_integr/cbps_lin_spe_peaks_assay.rds")
atacs@assays$lin_peaks@motifs<-readRDS("outputs/07-DMCs_atac_integr/atacs_cbps_lin_peaks_motif_object.rds")
DefaultAssay(atacs)<-"lin_peaks"
#for EGR1
peaks_hsc_genes<-fread("outputs/07-DMCs_atac_integr/peaks_hsc_genes_anno.csv.gz")
peaks_close_EGR1_target<-peaks_hsc_genes[gene_name%in%regulons_list$EGR1]$query_region
egr1_peaks<-CheckMotif(atacs,
peaks =peaks_close_EGR1_target ,
motif.name = "EGR1",
return.peaks = TRUE)
length(egr1_peaks)/length(peaks_close_EGR1_target) #30%
egr1_genes<-intersect(peaks_hsc_genes[query_region%in%egr1_peaks]$gene_name,
regulons_list$EGR1)
length(egr1_genes)/length(regulons_list$EGR1) #96% ! (25/26)
#egr1_extended
peaks_close_EGR1_target<-peaks_hsc_genes[gene_name%in%regulons_list$EGR1e]$query_region
egr1_peaks<-CheckMotif(atacs,
peaks =peaks_close_EGR1_target ,
motif.name = "EGR1",
return.peaks = TRUE)
length(egr1_peaks)/length(peaks_close_EGR1_target) #17%
egr1_genes<-intersect(peaks_hsc_genes[query_region%in%egr1_peaks]$gene_name,
regulons_list$EGR1e)
length(egr1_genes)/length(regulons_list$EGR1e) #93% (393/424)
#for all
tfs_scenic<-unique(str_remove(names(regulons_list),"e$"))
regulons_tf_atac<-unlist(atacs@assays$lin_peaks@motifs@motif.names[atacs@assays$lin_peaks@motifs@motif.names%in%tfs_scenic])
length(regulons_tf_atac)/length(tfs_scenic)#107/157
regulons_atac_list<-regulons_list[str_remove(names(regulons_list),"e$")%in%regulons_tf_atac]
length(regulons_atac_list) #174
length(regulons_list) #250
regulons_atac_listf<-lapply(names(regulons_atac_list), function(regulon_name){
targets<-regulons_atac_list[[regulon_name]]
motif_name<-str_remove(regulon_name,"e$")
peaks_close_targets<-peaks_hsc_genes[gene_name%in%targets]$query_region
tf_peaks<-CheckMotif(atacs,
peaks =peaks_close_targets ,
motif.name = motif_name,
return.peaks = TRUE)
filtered_targets<-intersect(peaks_hsc_genes[query_region%in%tf_peaks]$gene_name,
targets)
return(filtered_targets)
})
names(regulons_atac_listf)<-names(regulons_atac_list)
cat(unlist(lapply(1:length(regulons_atac_listf),
function(i)paste(names(regulons_atac_listf)[i],"=",round(length(regulons_atac_listf[[i]])/length(regulons_atac_list[[i]])*100),"%"))),sep = "\n")
mean(unlist(lapply(1:length(regulons_atac_listf),
function(i)length(regulons_atac_listf[[i]])/length(regulons_atac_list[[i]])))) #59%
#make a df of interactions tf > targets
regulons<-Reduce(rbind,lapply(names(regulons_atac_listf), function(t)data.table(tf=rep(t,length(regulons_atac_listf[[t]])),target=regulons_atac_listf[[t]])))
regulons[,extended:=str_detect(tf,"e$")]
regulons[,tf:=str_remove(tf,"e$")]
regulons[(extended)] #25397 tf > target interaction
regulons[(!extended)] #4808 tf > target interaction with high confidence
fwrite(regulons,fp(out,"tf_target_interactions.csv"))
regulons<-fread(fp(out,"tf_target_interactions.csv"))
#%TF-target conserved
regulonsf<-fread(fp(out,"tf_target_interactions.csv"))[!(extended)]
regulons_old<-fread("outputs/10-SCENIC/regulons.csv")
res_conserved<-sapply(unique(regulons_old$tf), function(t)length(intersect(regulons_old[tf==t]$gene,regulonsf[tf==t]$target))/nrow(regulons_old[tf==t]))
res_conserved[c("EGR1","KLF2","KLF4")]
#start build network only with tf> interact with high conf
regf<-regulons[(!extended)]
length(unique(regf$tf)) #72 tfs
regf[,n.target:=.N,by="tf"]
summary(unique(regf,by="tf")$n.target)
# Min. 1st Qu. Median Mean 3rd Qu. Max.
# 4.00 16.50 32.50 66.78 75.50 462.00
#show network using ggnet
#renv::install("briatte/ggnet")
library(ggnet)
library(network)
library(sna)
?network
regf<-regf[!is.na(target)]
net<-as.network(regf[,.(tf,target)],loops = T,directed = T)
net
# Network attributes:
# vertices = 1802
# directed = TRUE
# hyper = FALSE
# loops = TRUE
# multiple = FALSE
# bipartite = FALSE
# total edges= 4808
# missing edges= 0
# non-missing edges= 4808
#
# Vertex attribute names:
# vertex.names
#
# Edge attribute names not shown
saveRDS(net,fp(out,"network_tf_target_hi_conf.rds"))
#only with tf of interest
egr1_modul<-c("KLF2","EGR1","KLF4")
reg_egr1<-regf[tf%in%c(egr1_modul)] #add only targets of the tfs altered
fwrite(reg_egr1,fp(out,"egr1_KLF2_KLF4_network_tf_target_interactions.csv"))
net_genes<-union(reg_egr1$tf,reg_egr1$target)
reg_egr1r1<-unique(rbind(reg_egr1,regf[target%in%egr1_modul&tf%in%net_genes])) #add also regulators of this tfs in this newtwork
fwrite(reg_egr1r1,fp(out,"egr1_network_plus_tf_regulators_tf_target_interactions.csv"))
# reg_egr1r2<-regf[tf%in%c(egr1_modul)|target%in%egr1_modul] #add upstream regulators of egr1_modul
# fwrite(reg_egr1r2,fp(out,"egr1_network_plus_tf_regulators.csv"))
#reg_egr1<-unique(rbind(reg_egr1,regf[tf%in%net_genes&target%in%net_genes])) #add all interactions of this genes presents
tfs<-unique(reg_egr1r1$tf)
net_egr1<-as.network(reg_egr1r1[,.(tf,target)],loops = T,directed = T)
net_egr1
# Network attributes:
# vertices = 123
# directed = TRUE
# hyper = FALSE
# loops = TRUE
# multiple = FALSE
# bipartite = FALSE
# total edges= 161
# missing edges= 0
# non-missing edges= 161
#
# Vertex attribute names:
# vertex.names
#
# No edge attributes
#add a vertex attributes wich indicates if the gene is a tf or not
net_egr1 %v% "type" = ifelse(network.vertex.names(net_egr1) %in% regf$tf, "tf", "gene")
#add methyl info
res_m<-fread("outputs/02-gene_score_calculation_and_validation/res_genes.csv.gz")
res_m[gene_score_add>500,meth:="darkred"]
res_m[gene_score_add<=500,meth:="black"]
net_egr1 %v% "meth" = sapply(res_m[network.vertex.names(net_egr1),on="gene"]$meth,function(x)ifelse(is.na(x),"cornsilk3",x))
#add expr info
res_e<-fread("outputs/06-LGA_vs_Ctrl_RNA/res_pseudobulkDESeq2_by_lineage.csv.gz")[lineage=="HSC"]
res_e[padj>0.05,deg:="cornsilk3"]
res_e[padj<=0.05&log2FoldChange>0,deg:="coral2"]
res_e[padj<=0.05&log2FoldChange>0.5,deg:="coral3"]
res_e[padj<=0.05&log2FoldChange<(0),deg:="cadetblue3"]
res_e[padj<=0.05&log2FoldChange<(-0.5),deg:="cadetblue4"]
res_e[padj<=0.05&log2FoldChange<(-0.25)]
net_egr1 %v% "deg" = res_e[network.vertex.names(net_egr1),on="gene"]$deg
#add atac info
#on vertice
#need add target info
res_a<-fread("outputs/08-chromatin_change_LGA_vs_Ctrl/differential_peaks_accessibility_lga_vs_ctrl_hsc_logFC0.csv.gz")
res_a<-res_a[!str_detect(peak,"chr[XY]")]
peaks_hsc_genes[,peak:=query_region]
res_at<-merge(res_a,peaks_hsc_genes,by="peak")
res_at[,target:=gene_name]
res_at[p_val_adj<0.001&avg_log2FC>0.25,da:="red"]
res_at[p_val_adj<0.001&avg_log2FC<(-0.25),da:="blue"]
res_at[is.na(da),da:="grey75"]
net_egr1 %v% "da" = res_at[network.vertex.names(net_egr1),on="target"]$da
#on edge : df tf-target link with if peak with motif found, FC / pval of the change
#need merge network df with res_atac df
#add TF info on res_atac => for each peak, merge with tf(of the network)-peak dt
tfs<-unique(reg_egr1r1[,.(tf,target)]$tf) #"KLF4" "EGR1" "KLF2" "ATF4" "ATF3" "JUN" "FOS" "JUNB"
peaks<-unique(res_at[target%in%reg_egr1r1$target]$peak)
length(peaks)#690
motif.all <- GetMotifData(
object = atacs, assay = "lin_peaks", slot = "data"
)
motifs_peaks_tfs <- motif.all[peaks,GetMotifIDs(atacs,tfs) , drop = FALSE]
tf_peak_dt<-melt(data.table(data.frame(as.matrix(motifs_peaks_tfs==1)),keep.rownames = "peak"),id.vars = "peak",variable.name ="motif.id" ,value.name = "is_present")
tf_peak_dt<-merge(tf_peak_dt,GetMotifIDs(atacs,tfs,return_dt=TRUE))
tf_peak_dt<-tf_peak_dt[is_present==TRUE]
res_at_tf<-merge(res_at,tf_peak_dt,by="peak")
res_at_tf[,tf:=motif.name]
#merge with network df
reg_egr1r1_peaks<-merge(reg_egr1r1,
res_at_tf[,.(tf,target,peak,p_val,p_val_adj,avg_log2FC,pct.1,pct.2,type,da)],
by = c("tf","target"),
all.x = T)
reg_egr1r1_peaks[,n.tf.target.peaks:=.N,by=.(tf,target)]
reg_egr1r1_peaks[,biggest_change:=p_val==min(p_val),.(tf,target)]
reg_egr1r1_peaks[(biggest_change)|is.na(biggest_change)]
unique(reg_egr1r1_peaks[(biggest_change)|is.na(biggest_change)],by=c("tf","target"))#ok
reg_egr1r1_peaks[,da.peak:=p_val_adj<0.001&abs(avg_log2FC)>0.25]
reg_egr1r1_peaks[,n.da.tf.target.peaks:=sum(da.peak),.(tf,target)]
reg_egr1r1_peaks[da.peak==T] #9
reg_egr1r1_peak1<-reg_egr1r1_peaks[(biggest_change)|is.na(biggest_change)]
#add DMCs infos on edge
#need merge peaks DMCs df with reg_egr1 df
peaks_cpgs<-fread("outputs/07-DMCs_atac_integr/cpgs_in_lin_OCRs.csv.gz")
peaks_meth<-merge(peaks_cpgs,fread("outputs/01-lga_vs_ctrl_limma_DMCs_analysis/res_limma.tsv.gz"))
peaks_meth[,peak:=peaks]
peaks_meth_hsc<-peaks_meth[peak%in%peaks_hsc_genes$peak]
reg_egr1r1_peak1_meth<-merge(reg_egr1r1_peak1,
peaks_meth_hsc[,.(peak,cpg_id,logFC,P.Value,adj.P.Val)],
by="peak",
all.x=T)
reg_egr1r1_peak1_meth[,n.cpg.peak:=.N,by=.(peak,tf)]
reg_egr1r1_peak1_meth[,biggest_meth_change:=P.Value==min(P.Value),.(peak,tf)]
reg_egr1r1_peak1_meth[(biggest_meth_change)|is.na(biggest_meth_change)] #ok
reg_egr1r1_peak1_meth[,dmcs:=P.Value<0.001&abs(logFC)>25]
reg_egr1r1_peak1_meth[,n.dmcs.peak:=sum(dmcs),.(peak,tf)]
reg_egr1r1_peak1_meth[dmcs==T] #24
reg_egr1r1_peak1_meth1<-reg_egr1r1_peak1_meth[(biggest_meth_change)|is.na(biggest_meth_change)]
#ADD edge atttibute (tf> target) : color depend of if atac based tf> target interact is altered by chromatin change
#if tf-gene peak dn : blue if tf-gene peak up : red
net_egr1_a<-network(reg_egr1r1_peak1_meth1[,-c("extended","biggest_change","biggest_meth_change","peak")],loops = T,directed = T)
list.edge.attributes(net_egr1_a)
as.matrix(net_egr1_a,attrname='da')
net_egr1_a %e% "da"
net_egr1_a %e% "da"=sapply(net_egr1_a %e% "da",function(x)ifelse(x=="grey75","darkgrey",x))
net_egr1_a %e% "dmc_line"=net_egr1_a %e% "n.dmcs.peak"+1
net_egr1_a %e% "dmc_line"=sapply(net_egr1_a %e% "dmc_line",function(x)ifelse(is.na(x),1,x))
#set.edge.attribute(net_egr1, "color", ifelse(net_egr1 %e% "dap" > 1, "black", "grey75"))
#add vertices attributes
net_egr1_a %v% "type" = ifelse(network.vertex.names(net_egr1_a) %in% regf$tf, "tf", "gene")
net_egr1_a %v% "deg" = res_e[network.vertex.names(net_egr1_a),on="gene"]$deg
net_egr1_a %v% "deg" = sapply(net_egr1_a %v% "deg",function(x)ifelse(is.na(x),"cornsilk3",x))
net_egr1_a %v% "meth" = sapply(res_m[network.vertex.names(net_egr1_a),on="gene"]$meth,function(x)ifelse(is.na(x),"black",x))
#genes_of_interest<-union(res_e[padj<=0.05&abs(log2FoldChange)>0.5]$gene,union(res_m[gene_score_add>500]$gene,unique(reg_egr1$tf)))
#GRN sans selection,label all genes
ggnet2(net_egr1_a,
color = "deg",
label = T,label.color = "meth",label.size = 2,
size = "type" ,size.palette = c("tf"=3,"gene"=1),
shape = "type",
edge.alpha = 0.8,
edge.size=0.5,
edge.color = "da",
arrow.size = 5,
edge.lty = "dmc_line",
arrow.gap =0.02) +
theme(panel.background = element_rect(fill = "white"))
ggsave(fp(out,"final_network_EGR1_KLF2_KLF4_tf_targets_2.pdf"))
reg_egr1r1_peak1_meth1[n.dmcs.peak>1]