-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathGSE124647.R
319 lines (248 loc) · 14.5 KB
/
GSE124647.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
#####################################
########### GSE124647 ################
#Expression
gse124647 <- read.delim("./../Data/GSE124647/GSE124647_series_matrix.txt", comment.char="!", header=TRUE, fill=TRUE, row.names=1)
#Header annotation
gse124647.annot <- read.delim("./../Data/GSE124647/GSE124647_series_matrix copy.txt", header=FALSE, fill=TRUE, skip=10, nrows=16)
gse124647.annot <- data.frame(t(gse124647.annot[,-1]))
rownames(gse124647.annot) <- gse124647.annot$X16
temp <- apply(gse124647.annot, 2, function (x) strsplit(as.character(x), ":"))
gse124647.annotV2 <- data.frame(matrix(data=NA, nrow=nrow(gse124647.annot), ncol=15))
for (i in 1:15) {
gse124647.annotV2[,i] <- sapply(temp[[i]], '[', 2)
colnames(gse124647.annotV2)[i] <- temp[[i]][[1]][1]
}
rownames(gse124647.annotV2) <- gse124647.annot$X16
#Array information
gse124647.genes <- read.delim("./../Data/GSE124647/GPL96-57554.txt", comment.char="#", header=TRUE)
#Pre-process expression matrix
x <- match(rownames(gse124647), gse124647.genes$ID)
agg.gse124647 <- aggregate(gse124647, by=list(gse124647.genes$Gene.Symbol[x]), FUN=mean)
gse124647.mat <- agg.gse124647[-1, 2:ncol(agg.gse124647)]
rownames(gse124647.mat) <- agg.gse124647$Group.1[-1]
gvars <- apply(gse124647.mat, 1 , var)
gse124647.s <- scale(t(gse124647.mat[-which(gvars < 0.1), ]))
### Match matrices and train the model on batch-corrected data ####-------------------------
library(sva)
temp <- DoCombat(metabric.s, t(gse124647.s))
train <- match(colnames(metabric.s), colnames(temp))
test <- match(rownames(gse124647.s), colnames(temp))
MBGSE124647.ENDORSE <- gsva(temp[, train], gset.idx.list = list("ENDORSE"=ENDORSE,
"HALLMARK_ESTROGEN_RESPONSE_EARLY" = H_ESTR_EARLY,
"HALLMARK_ESTROGEN_RESPONSE_LATE" = H_ESTR_LATE),
method="ssgsea", kcdf="Gaussian")
df <- data.frame('Time'=metabric.annotV2$OS_MONTHS,
'Event'=as.numeric(metabric.annotV2$OS_EVENT),
'Age'=metabric.annotV2$AGE_AT_DIAGNOSIS,
t(MBGSE124647.ENDORSE))
cfit.0 = coxph(Surv(Time, Event) ~ ENDORSE, data=df)
cfit.5 = coxph(Surv(Time, Event) ~ HALLMARK_ESTROGEN_RESPONSE_EARLY, data=df)
cfit.6 = coxph(Surv(Time, Event) ~ ENDORSE + HALLMARK_ESTROGEN_RESPONSE_EARLY, data=df, x = T)
### ER response and survival analysis -------
gse124647.ssgsea <- gsva(temp[, test], list("ENDORSE"=ENDORSE,
"HALLMARK_ESTROGEN_RESPONSE_EARLY" = H_ESTR_EARLY,
"HALLMARK_ESTROGEN_RESPONSE_LATE" = H_ESTR_LATE),
verbose=TRUE, kcdf="Gaussian", method='ssgsea', ssgsea.norm=T)
gse124647.ssgsea2 <- rbind(gsva(temp[, test], Hset, method='ssgsea', kcdf='Gaussian'),
gsva(temp[, test], c6set, method='ssgsea', kcdf='Gaussian'))
gse124647.os <- Surv(time=as.numeric(gse124647.annotV2$`oas months`), event=as.numeric(gse124647.annotV2$`os event`))
gse124647.pfs <- Surv(time=as.numeric(gse124647.annotV2$`pfs months`), event=as.numeric(gse124647.annotV2$pfs.event))
gse124647.ssgsea2 <- rbind(gsva(temp[, test], Hset, method='ssgsea', kcdf='Gaussian'),
gsva(temp[, test], c6set, method='ssgsea', kcdf='Gaussian'))
# predict
gse124647.pred <- predict(cfit.6, data.frame(t(gse124647.ssgsea)), type="risk")
gse124647.pred.grp <- MakeRiskGrp(gse124647.pred)
############# Apply TransCONFIRM signature #########################
gse124647.TransCONFIRM <- TransCONFIRM(gse124647.s)
############# Apply SET signature #########################
gse124647.SET <- SET_Score(gse124647.s)
x <- match(names(gse124647.SET), gse124647.annot$X16)
annot.SET <- as.numeric(sapply(strsplit(as.character(gse124647.annot[x,15]), " "), "[", 4))
cor(gse124647.SET, annot.SET, use='complete.obs')
cor.test(gse124647.SET, annot.SET, use='complete.obs')
plot(gse124647.SET, annot.SET)
df.gse124647 <- data.frame("OSTime"=as.numeric(gse124647.annotV2$`oas months`),
"OSEvent"=as.numeric(gse124647.annotV2$`os event`),
"PFSTime"=as.numeric(gse124647.annotV2$`pfs months`),
"PFSEvent"=as.numeric(gse124647.annotV2$pfs.event),
"RiskGroup"=gse124647.pred.grp,
"Risk"=gse124647.pred,
"SET"=gse124647.SET,
"TransCONFIRM"=gse124647.TransCONFIRM)
gse124647.SET_Cat <- array(data=0, dim = length(gse124647.SET))
gse124647.SET_Cat[which(gse124647.SET > median(gse124647.SET))] <- 1
df.gse124647$SET_Cat <- gse124647.SET_Cat
cpOS.E2.gse124647 <- coxph(Surv(OSTime, OSEvent) ~ RiskGroup, data=df.gse124647, x=T)
cpOS.SET.gse124647 <- coxph(Surv(OSTime, OSEvent) ~ SET, data=df.gse124647, x=T)
cpOS.TC.gse124647 <- coxph(Surv(OSTime, OSEvent) ~ TransCONFIRM, data=df.gse124647, x=T)
cpPFS.E2.gse124647 <- coxph(Surv(PFSTime, PFSEvent) ~ RiskGroup, data=df.gse124647, x=T)
cpPFS.SET.gse124647 <- coxph(Surv(PFSTime, PFSEvent) ~ SET, data=df.gse124647, x=T)
cpPFS.TC.gse124647 <- coxph(Surv(PFSTime, PFSEvent) ~ TransCONFIRM, data=df.gse124647, x=T)
summary(cpOS.E2.gse124647)
cpOS.SET.gse124647
cpOS.TC.gse124647
summary(cpPFS.E2.gse124647)
cpPFS.SET.gse124647
cpPFS.TC.gse124647
plrtest(cpOS.E2.gse124647, cpOS.SET.gse124647, nested=F)
plrtest(cpOS.E2.gse124647, cpOS.TC.gse124647, nested=F)
plrtest(cpPFS.E2.gse124647, cpPFS.SET.gse124647, nested=F)
plrtest(cpPFS.E2.gse124647, cpPFS.TC.gse124647, nested=F)
##### Survival plots --------------------------------------------------
library(survminer)
fitOS.E2.gse124647 <- survfit(Surv(OSTime, OSEvent) ~ RiskGroup, data=df.gse124647)
fitOS.SET.gse124647 <- survfit(Surv(OSTime, OSEvent) ~ SET_Cat, data=df.gse124647)
fitOS.TC.gse124647 <- survfit(Surv(OSTime, OSEvent) ~ TransCONFIRM, data=df.gse124647)
fitPFS.E2.gse124647 <- survfit(Surv(PFSTime, PFSEvent) ~ RiskGroup, data=df.gse124647)
fitPFS.SET.gse124647 <- survfit(Surv(PFSTime, PFSEvent) ~ SET_Cat, data=df.gse124647)
fitPFS.TC.gse124647 <- survfit(Surv(PFSTime, PFSEvent) ~ TransCONFIRM, data=df.gse124647)
P1 <- ggsurvplot(fitOS.E2.gse124647, data=df.gse124647, palette="jco",
censor.shape="|", censor.size=2,
break.time.by=12,
legend.title="ENDORSE", legend.labs=c("Low risk", "Medium risk", "High risk"),
surv.median.line = "hv",
risk.table=T,
#title="ER+/HER2- mBC OS",
ggtheme=theme_classic2(base_size = 14))
P2 <- ggsurvplot(fitOS.SET.gse124647, data=df.gse124647, palette="jco",
censor.shape="|", censor.size=2,
break.time.by=12,
legend.title="SET", legend.labs=c("Low SET", "High SET"),
surv.median.line = "hv",
risk.table=T,
#title="ER+/HER2- mBC OS",
ggtheme=theme_classic2(base_size = 14))
P3 <- ggsurvplot(fitOS.TC.gse124647, data=df.gse124647, palette="jco",
censor.shape="|", censor.size=2,
break.time.by=12,
legend.title="TransCONFIRM", legend.labs=c("TC Low", "TC High"),
surv.median.line = "hv",
risk.table=T,
#title="ER+/HER2- mBC PFS",
ggtheme=theme_classic2(base_size = 14))
P4 <- ggsurvplot(fitPFS.E2.gse124647, data=df.gse124647, palette="jco",
censor.shape="|", censor.size=2,
break.time.by=12,
legend.title="ENDORSE", legend.labs=c("Low risk", "Medium risk", "High risk"),
surv.median.line = "hv",
risk.table=T,
#title="ER+/HER2- mBC PFS",
ggtheme=theme_classic2(base_size = 14))
P5 <- ggsurvplot(fitPFS.SET.gse124647, data=df.gse124647, palette="jco",
censor.shape="|", censor.size=2,
break.time.by=12,
legend.title="SET", legend.labs=c("Low SET", "High SET"),
surv.median.line = "hv",
risk.table=T,
#title="ER+/HER2- mBC PFS",
ggtheme=theme_classic2(base_size = 14))
P6 <- ggsurvplot(fitPFS.TC.gse124647, data=df.gse124647, palette="jco",
censor.shape="|", censor.size=2,
break.time.by=12,
legend.title="TransCONFIRM", legend.labs=c("TC Low", "TC High"),
surv.median.line = "hv",
risk.table=T,
#title="ER+/HER2- mBC PFS",
ggtheme=theme_classic2(base_size = 14))
pdf("./../Figures/Prediction_GSE124647_os_pfs_ENDORSEv2.pdf", width=6, height=6)
print(P1)
print(P2)
print(P3)
print(P4)
print(P5)
print(P6)
dev.off()
survdiff(Surv(OSTime, OSEvent) ~ RiskGroup, data=df.gse124647)
survdiff(Surv(OSTime, OSEvent) ~ SET_Cat, data=df.gse124647)
survdiff(Surv(OSTime, OSEvent) ~ TransCONFIRM, data=df.gse124647)
survdiff(Surv(PFSTime, PFSEvent) ~ RiskGroup, data=df.gse124647)
survdiff(Surv(PFSTime, PFSEvent) ~ SET_Cat, data=df.gse124647)
survdiff(Surv(PFSTime, PFSEvent) ~ TransCONFIRM, data=df.gse124647)
#### Scatter plot
gse124647.gams <- apply(t(gse124647.ssgsea2), 2, GamApp, gse124647.pred)
write.csv(t(gse124647.gams), file="./../Output/gse124647.gams.csv")
tail(sort(gse124647.gams[3,]), 20)
df <- data.frame(cbind(t(gse124647.ssgsea2), "Risk"=gse124647.pred))
PPL <- list()
PPL[[1]] <- ggplot(df, aes(x=HALLMARK_G2M_CHECKPOINT, y=Risk, color=HALLMARK_G2M_CHECKPOINT)) +
geom_point(size=3) +
scale_colour_gradient2(low = "blue", mid = "grey80", high = "red", name="GES") +
geom_smooth(method='gam', se=T, color='navy', size=3) +
labs(x="Cell cycle progression", y="ENDORSE") +
theme_classic(base_size = 16) +
theme(axis.title = element_text(size=14, face="bold"),
title = element_text(size=16, face="bold", color="grey25"))
PPL[[2]] <- ggplot(df, aes(x=E2F1_UP.V1_UP, y=Risk, color=E2F1_UP.V1_UP)) +
geom_point(size=3) +
scale_colour_gradient2(low = "blue", mid = "grey80", high = "red", name="GES") +
geom_smooth(method='gam', se=T, color='navy', size=3) +
labs(x="E2F1 upregulation", y="ENDORSE") +
theme_classic(base_size = 16) +
theme(axis.title = element_text(size=14, face="bold"),
title = element_text(size=16, face="bold", color="grey25"))
PPL[[3]] <- ggplot(df, aes(x=MTOR_UP.V1_UP, y=Risk, color=MTOR_UP.V1_UP)) +
geom_point(size=3) +
scale_colour_gradient2(low = "blue", mid = "grey80", high = "red", name="GES") +
geom_smooth(method='gam', se=T, color='navy', size=3) +
labs(x="MTOR upregulation", y="ENDORSE") +
theme_classic(base_size = 16) +
theme(axis.title = element_text(size=14, face="bold"),
title = element_text(size=16, face="bold", color="grey25"))
pdf("./../Figures/GSE124647.pathways.pdf", height = 4, width=15)
egg::ggarrange(plots=PPL, nrow=1)
dev.off()
######### Compare with the empirical and estrogen response signatures ############
gse124647.pred.c0 <- predict(cfit.0, data.frame(t(gse124647.ssgsea)), type="risk")
gse124647.pred.grp.c0 <- MakeRiskGrp(gse124647.pred.c0)
gse124647.pred.c5 <- predict(cfit.5, data.frame(t(gse124647.ssgsea)), type="risk")
gse124647.pred.grp.c5 <- MakeRiskGrp(gse124647.pred.c5)
df.gse124647.2 <- data.frame("OSTime"=as.numeric(gse124647.annotV2$`oas months`),
"OSEvent"=as.numeric(gse124647.annotV2$`os event`),
"PFSTime"=as.numeric(gse124647.annotV2$`pfs months`),
"PFSEvent"=as.numeric(gse124647.annotV2$pfs.event),
"RiskGroup"=gse124647.pred.grp,
"Risk"=gse124647.pred,
"RiskGroup0"=gse124647.pred.grp.c0,
"Risk0"=gse124647.pred.c0,
"RiskGroup5"=gse124647.pred.grp.c5,
"Risk5"=gse124647.pred.c5)
cpPFS.c0.gse124647 <- coxph(Surv(PFSTime, PFSEvent) ~ Risk0, data=df.gse124647.2, x=T) #7.735e-05
cpPFS.c5.gse124647 <- coxph(Surv(PFSTime, PFSEvent) ~ Risk5, data=df.gse124647.2, x=T) #0.01597
cpPFS.c6.gse124647 <- coxph(Surv(PFSTime, PFSEvent) ~ Risk, data=df.gse124647.2, x=T) #4.184e-06
cpOS.c0.gse124647 <- coxph(Surv(OSTime, OSEvent) ~ Risk0, data=df.gse124647.2, x=T) #0.0003096
cpOS.c5.gse124647 <- coxph(Surv(OSTime, OSEvent) ~ Risk5, data=df.gse124647.2, x=T) #0.0254
cpOS.c6.gse124647 <- coxph(Surv(OSTime, OSEvent) ~ Risk, data=df.gse124647.2, x=T) #4.782e-05
plrtest(cpOS.c6.gse124647, cpOS.c0.gse124647, nested=T) #H1 p = 0.0208 #H1A p = 0.036
plrtest(cpOS.c6.gse124647, cpOS.c5.gse124647, nested=T) #H1 p = 0.000199 #H1A p = 0.000472
plrtest(cpPFS.c6.gse124647, cpPFS.c0.gse124647, nested=T) #H1 p = 0.00484 #H1A p = 0.0109
plrtest(cpPFS.c6.gse124647, cpPFS.c5.gse124647, nested=T) #H1 p = 6e-04 #H1A p = 0.00026
fitOS.c0.gse124647 <- survfit(Surv(OSTime, OSEvent) ~ RiskGroup0, data=df.gse124647.2)
fitOS.c5.gse124647 <- survfit(Surv(OSTime, OSEvent) ~ RiskGroup5, data=df.gse124647.2)
fitPFS.c0.gse124647 <- survfit(Surv(PFSTime, PFSEvent) ~ RiskGroup0, data=df.gse124647.2)
fitPFS.c5.gse124647 <- survfit(Surv(PFSTime, PFSEvent) ~ RiskGroup5, data=df.gse124647.2)
pdf("./../Figures/GSE124647_Emp_Hstr.pdf", width=6, height=6)
ggsurvplot(fitOS.c0.gse124647, data=df.gse124647.2, palette="jco",
censor.shape="|", censor.size=2,
break.time.by=12,
legend.title="Empirical", legend.labs=c("Low risk", "Medium risk", "High risk"),
surv.median.line = "hv",
ggtheme=theme_classic2(base_size = 14))
ggsurvplot(fitOS.c5.gse124647, data=df.gse124647.2, palette="jco",
censor.shape="|", censor.size=2,
break.time.by=12,
legend.title="Estrogen response", legend.labs=c("Low risk", "Medium risk"),
surv.median.line = "hv",
ggtheme=theme_classic2(base_size = 14))
ggsurvplot(fitPFS.c0.gse124647, data=df.gse124647.2, palette="jco",
censor.shape="|", censor.size=2,
break.time.by=12,
legend.title="Empirical", legend.labs=c("Low risk", "Medium risk", "High risk"),
surv.median.line = "hv",
ggtheme=theme_classic2(base_size = 14))
ggsurvplot(fitPFS.c5.gse124647, data=df.gse124647.2, palette="jco",
censor.shape="|", censor.size=2,
break.time.by=12,
legend.title="Estrogen response", legend.labs=c("Low risk", "Medium risk"),
surv.median.line = "hv",
ggtheme=theme_classic2(base_size = 14))
dev.off()