-
Notifications
You must be signed in to change notification settings - Fork 0
/
mi.R
315 lines (255 loc) · 12.4 KB
/
mi.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
set.seed(1)
source('set_up_script.R')
nIter <- 10000
draws <- 10000
cores <<- 16
registerDoParallel(cores=cores)
enrollment_scalar_samples <<- runif(draws,0,1)
highrisk_scalar_samples <<- runif(draws,0,1)
neighbour_scalar_samples <<- runif(draws,0,1)
recall_samples <<- rbeta(draws,5,1)
variables <- list(highrisk_scalar_samples,neighbour_scalar_samples,recall_samples)
number_sampled <- 100#sample(range_informative_clusters,1)
true_contact_list <<- contact_list
true_contact_of_contact_list <<- contact_of_contact_list
true_household_list <<- household_list
true_high_risk_list <<- high_risk_list
true_neighbour_scalar <<- 0.39
true_high_risk_scalar <<- 2.17
true_enrollment_scalar <<- 0.7
trim_contact_networks <- function(recall){
contact_list <<- trim_contact_network(true_contact_list,recall)
contact_of_contact_list <<- trim_contact_network(true_contact_of_contact_list,recall)
household_list <<- trim_contact_network(true_household_list,recall)
high_risk_list <<- trim_contact_network(true_high_risk_list,recall)
}
trim_contact_network <- function(full_contact_list,recall){
trimmed_contact_list <- full_contact_list
for(i in 1:length(full_contact_list)){
size <- length(full_contact_list[[i]])
if(size>0){
ssize <- rbinom(1,size,recall)
if(ssize>0)
trimmed_contact_list[[i]] <- sample(full_contact_list[[i]],ssize)
}
}
return(trimmed_contact_list)
}
## type 1 error ############################################################
direct_VE <<- 0
netwk_list <- list()
neighbour_scalar <<- true_neighbour_scalar
highrisk_scalar <<- true_high_risk_scalar
enrollment_scalar <<- true_enrollment_scalar
for(iter in 1:nIter){
## select random person to start
first_infected <- sample(g_name,1)
inf_period <- rgamma(length(first_infected),shape=infperiod_shape,rate=infperiod_rate)
#hosp_time <- rgamma(length(first_infected),shape=hosp_shape_index,rate=hosp_rate_index)
hosp_time <- rtruncnorm(length(first_infected),a=0,mean=hosp_mean_index,sd=hosp_sd_index)
inf_time <- min(inf_period,hosp_time)
netwk <- simulate_contact_network(first_infected,inf_time,start_day=iter,from_source=0,cluster_flag=0)
netwk_list[[iter]] <- netwk
}
trial_summary <- lapply(netwk_list,summarise_trial,ve_est_temp=0)
tte <- do.call(rbind,lapply(1:length(trial_summary),function(cluster)if(!is.null(trial_summary[[cluster]]))cbind(trial_summary[[cluster]],cluster)))
trial_summary <- c()
ntwks <- unique(tte$cluster)
#profvis({
par_results <- do.call(rbind,mclapply(1:draws,function(cl){
clusters_sampled <- sample(ntwks,number_sampled,replace=F)
neighbour_scalar <<- qbeta(neighbour_scalar_samples[cl],3,3)
high_risk_scalar <<- 1/qbeta(highrisk_scalar_samples[cl],3,3)
enrollment_scalar <<- qbeta(enrollment_scalar_samples[cl],3,3)
recall <- recall_samples[cl]
trim_contact_networks(recall)
pv <- iterate_ph_model(netwk_list[clusters_sampled])
trial_summary <- list()
#for(i in 1:length(clusters_sampled)) trial_summary[[i]] <- summarise_trial(netwk_list[[clusters_sampled[i]]],ve_est_temp=pv[2])
#netwk_list <- c()
for(cluster in 1:length(clusters_sampled)) {
trial_summary[[cluster]] <- summarise_trial(netwk_list[[clusters_sampled[cluster]]],ve_est_temp=pv[2])
#if(!is.null(trial_summary[[cluster]]))
# trial_summary[[cluster]] <- cbind(trial_summary[[cluster]],cluster)
}
tte <- do.call(bind_rows,trial_summary)
#tte <- do.call(rbind,lapply(1:length(trial_summary),function(cluster)if(!is.null(trial_summary[[cluster]]))cbind(trial_summary[[cluster]],cluster)))
trial_summary <- c()
ttemat <- as.matrix(tte)
survmodel <- coxph(Surv(time,outcome)~vaccinated,weights=weight,tte)
vaccEffEst <- 1-exp(survmodel$coefficient + c(0, 1.96, -1.96)*as.vector(sqrt(survmodel$var)))
zval <- survmodel$coefficient/sqrt(survmodel$var)
pval <- pnorm(zval, lower.tail = vaccEffEst[1]>0)*2
return(c(pval,sum(ttemat[,7]==T),nrow(ttemat),vaccEffEst[1], sum(ttemat[,2]), sum(ttemat[,6]),sum(ttemat[,7]==T&ttemat[,8]==T),sum(ttemat[ttemat[,7]==T&ttemat[,8]==T,2]) )) ## output weights and exposures (time)
},mc.cores=cores))
#})
#netwk_list <- c()
variables <- list(highrisk_scalar_samples,neighbour_scalar_samples,recall_samples,par_results[,5],par_results[,6],par_results[,7],par_results[,8])
saveRDS(list(par_results,variables),'storage/mit1e.Rds')
t1e <- par_results[,1]
VEt1e <- par_results[,4]
outcomes <- list(par_results[,1],par_results[,4])
evppi <- sapply(variables,function(x)
sapply(outcomes,
function(y)mutinformation(discretize(x),discretize(y))
)
)
colnames(evppi) <- c('High-risk scalar','Neighbour scalar','Contact recall','Total weight','Total exposure','Vaccinated cases','Weighted vaccinated cases')
rownames(evppi) <- c('p value (VE=0)','VE estimate (VE=0)')
print(evppi)
x=readRDS('storage/mit1e.Rds')
t1e=x[[1]][,1]
VEt1e=x[[1]][,4]
variables<-x[[2]]
{pdf('figures/outcomebidensitiest1e.pdf',width=15); par(mfrow=c(2,5),mar=c(5,5,2,2))
for(i in 1:length(outcomes))
for(j in 1:length(variables))
plot(variables[[j]],outcomes[[i]],frame=F,xlab=colnames(evppi)[j],ylab=rownames(evppi)[i],main='',cex.lab=1.5,cex.axis=1.5,pch=15)
dev.off()}
## power ############################################################
direct_VE <<- 0.8
netwk_list <- list()
neighbour_scalar <<- true_neighbour_scalar
highrisk_scalar <<- true_high_risk_scalar
enrollment_scalar <<- true_enrollment_scalar
contact_list <<- true_contact_list
contact_of_contact_list <<- true_contact_of_contact_list
household_list <<- true_household_list
high_risk_list <<- true_high_risk_list
set.seed(1)
for(iter in 1:nIter){
## select random person to start
first_infected <- sample(g_name,1)
inf_period <- rgamma(length(first_infected),shape=infperiod_shape,rate=infperiod_rate)
#hosp_time <- rgamma(length(first_infected),shape=hosp_shape_index,rate=hosp_rate_index)
hosp_time <- rtruncnorm(length(first_infected),a=0,mean=hosp_mean_index,sd=hosp_sd_index)
inf_time <- min(inf_period,hosp_time)
netwk <- simulate_contact_network(first_infected,inf_time,start_day=iter,from_source=0,cluster_flag=0,direct_VE = direct_VE)
netwk_list[[iter]] <- netwk
}
trial_summary <- lapply(netwk_list,summarise_trial,ve_est_temp=0.8)
tte <- do.call(rbind,lapply(1:length(trial_summary),function(cluster)if(!is.null(trial_summary[[cluster]]))cbind(trial_summary[[cluster]],cluster)))
trial_summary <- c()
ntwks <- unique(tte$cluster)
#profvis({
par_results <- do.call(rbind,mclapply(1:draws,function(cl){
clusters_sampled <- sample(ntwks,number_sampled,replace=F)
neighbour_scalar <<- qbeta(neighbour_scalar_samples[cl],3,3)
high_risk_scalar <<- 1/qbeta(highrisk_scalar_samples[cl],3,3)
enrollment_scalar <<- qbeta(enrollment_scalar_samples[cl],3,3)
recall <- recall_samples[cl]
trim_contact_networks(recall)
pv <- iterate_ph_model(netwk_list[clusters_sampled])
trial_summary <- list()
#for(i in 1:length(clusters_sampled)) trial_summary[[i]] <- summarise_trial(netwk_list[[clusters_sampled[i]]],ve_est_temp=pv[2])
#netwk_list <- c()
for(cluster in 1:length(clusters_sampled)) {
trial_summary[[cluster]] <- summarise_trial(netwk_list[[clusters_sampled[cluster]]],ve_est_temp=pv[2])
#if(!is.null(trial_summary[[cluster]]))
# trial_summary[[cluster]] <- cbind(trial_summary[[cluster]],cluster)
}
tte <- do.call(bind_rows,trial_summary)
#tte <- do.call(rbind,lapply(1:length(trial_summary),function(cluster)if(!is.null(trial_summary[[cluster]]))cbind(trial_summary[[cluster]],cluster)))
trial_summary <- c()
ttemat <- as.matrix(tte)
survmodel <- coxph(Surv(time,outcome)~vaccinated,weights=weight,tte)
vaccEffEst <- 1-exp(survmodel$coefficient + c(0, 1.96, -1.96)*as.vector(sqrt(survmodel$var)))
zval <- survmodel$coefficient/sqrt(survmodel$var)
pval <- pnorm(zval, lower.tail = vaccEffEst[1]>0)*2
return(c(pval,sum(ttemat[,7]==T),nrow(ttemat),vaccEffEst[1], sum(ttemat[,2]), sum(ttemat[,6]) ,sum(ttemat[,7]==T&ttemat[,8]==T) ,sum(ttemat[ttemat[,7]==T&ttemat[,8]==T,2]) )) ## output weights and exposures (time)
},mc.cores=cores))
#})
#netwk_list <- c()
variables <- list(highrisk_scalar_samples,neighbour_scalar_samples,recall_samples,par_results[,5],par_results[,6],par_results[,7],par_results[,8])
saveRDS(list(par_results,variables),'storage/mipower.Rds')
power <- par_results[,1]
VE <- par_results[,4]
outcomes2 <- list(power,VE)
evppi2 <- sapply(variables,function(x)
sapply(outcomes2,
function(y)mutinformation(infotheo::discretize(x),infotheo::discretize(y))/infotheo::entropy(infotheo::discretize(y))
)
)
colnames(evppi2) <- c('High-risk scalar','Neighbour scalar','Contact recall','Total weight','Total exposure','Vaccinated cases','Weighted vaccinated cases')
rownames(evppi2) <- c('p value (VE=0.8)','VE estimate (VE=0.8)')
outcomes3 <- list(t1e,VEt1e,power,VE)
print(evppi2)
evppi3 <- rbind(evppi,evppi2)*100
print(evppi3)
sapply(1:4,function(i)infotheo::entropy(infotheo::discretize(outcomes3[[i]])))
#1/(c(var(old_par_results[,1]),var(old_par_results[,4]))/c(var(par_results[,1]),var(par_results[,4])))
{pdf('figures/scalarmi.pdf',height=6,width=2+length(outcomes3));
# {x11(height=6,width=2+length(outcomes3));
par(mar=c(17,12,3.5,10))
outcome_labels <- rownames(evppi3)
labs <- rev(colnames(evppi3))
get.pal=colorRampPalette(brewer.pal(9,"Reds"))
redCol=rev(get.pal(12))
bkT <- seq(max(evppi3)+1e-10, 0,length=13)
cex.lab <- 1.5
maxval <- round(bkT[1],digits=1)
col.labels<- c(0,maxval/2,maxval)
cellcolors <- vector()
for(ii in 1:length(unlist(evppi3)))
cellcolors[ii] <- redCol[tail(which(unlist(evppi3[ii])<bkT),n=1)]
color2D.matplot(evppi3,cellcolors=cellcolors,main="",xlab="",ylab="",cex.lab=2,axes=F,border='white')
fullaxis(side=2,las=2,at=0:(length(outcomes3)-1)+1/2,labels=rev(outcome_labels),line=NA,pos=NA,outer=FALSE,font=NA,lwd=0,cex.axis=1.2)
fullaxis(side=1,las=2,at=(length(labs)-1):0+0.5,labels=labs,line=NA,pos=NA,outer=FALSE,font=NA,lwd=0,cex.axis=1.2)
mtext(3,text='Mutual information',line=1,cex=1.3)
color.legend(length(labs)+0.5,0,length(labs)+0.8,length(outcomes3),col.labels,rev(redCol),gradient="y",cex=1.2,align="rb")
for(i in seq(0,length(outcomes3),by=1)) abline(h=i)
for(i in seq(0,length(labs),by=1)) abline(v=i)
dev.off()
}
{pdf('figures/outcomedensities.pdf'); par(mfrow=c(2,2),mar=c(3,5,5,2))
for(i in 1:length(outcomes3)) plot(density(outcomes3[[i]]),frame=F,lwd=2,xlab='',ylab='Density',main=rownames(evppi3)[i],cex.lab=1.5,cex.axis=1.5)
dev.off()
}
dis_fun <- function (x, numBins, r = range(x), b = NULL){
if(is.null(b)) b = seq(from = r[1], to = r[2], length.out = numBins + 1)
y = table(cut(x, breaks = b, include.lowest = TRUE))
return(list(y,b))
}
x=readRDS('storage/mipower.Rds')
power=x[[1]][,1]
VE=x[[1]][,4]
outcomes3 <- list(power,VE)
variables=x[[2]]
highrisk_scalar_samples=variables[[1]]
neigbour_scalar_samples=variables[[2]]
recall_samples=variables[[3]]
{pdf('figures/outcomebidensitiespower.pdf',width=15); par(mfrow=c(2,5),mar=c(5,5,2,2))
for(i in 1:length(outcomes3))
for(j in 1:length(variables))
plot(variables[[j]],outcomes3[[i]],frame=F,xlab=colnames(evppi3)[j],ylab=rownames(evppi3)[i],main='',cex.lab=1.5,cex.axis=1.5,pch=15)
dev.off()}
#old_power <- old_par_results[,1]
#numBins <- length(unlist(unique(infotheo::discretize(old_power))))
#bins <- dis_fun(old_power,numBins=numBins)[[2]]
#entropy(dis_fun(old_power,numBins=numBins,b=bins)[[1]])
#entropy(dis_fun(power,numBins=numBins,b=bins)[[1]])
sample_values <- list(1/qbeta(highrisk_scalar_samples,3,3),
qbeta(neighbour_scalar_samples,3,3),
recall_samples)
var_names <- c('High-risk scalar','Neighbour scalar','Contact recall')
measures <- list(t1e,power)
m_names <- c('Type 1 error','Power')
truth <- c(true_high_risk_scalar,true_neighbour_scalar,1)
x11(width=15,height=5);#
pdf('figures/powert1evsparams.pdf',width=15,height=5);
par(mar=c(5,5,2,2),mfrow=c(2,3))
for(j in 1:2){
measure <- measures[[j]]
for(i in 1:3){
param <- sample_values[[i]]
breaks <- c(quantile(param,seq(0,1,length=6))[1:5]-1e-10,max(param))
labels <- cut(param,breaks,labels=F)
powers <- sapply(1:5,function(x){subpow <- measure[labels==x];
sum(subpow<0.05)/length(subpow)})
plot(breaks[1:5]+diff(breaks)/2,powers,frame=F,col='navyblue',pch=16,cex=2,cex.axis=1.5,cex.lab=1.5,
xlab=var_names[i],ylab=m_names[j],xlim=range(param),log=ifelse(i==1,'x',''))
abline(v=breaks,col='grey',lwd=2,lty=2)
abline(v=truth[i],col='hotpink',lwd=2)
}
}
dev.off()