-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathsimulation_results.Rmd
executable file
·204 lines (176 loc) · 8.83 KB
/
simulation_results.Rmd
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
---
title: "Simulation results"
output: html_document
---
```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
library(ggplot2)
library(grid)
library(gridExtra)
library(reshape)
load_sv_data <- function(sim_dir, file) {
sv <- read.delim(paste(sim_dir, file, sep=''), stringsAsFactors=F, sep='\t')
sv[,'mean_vaf'] <- apply(data.frame(vaf1=sv$vaf1,vaf2=sv$vaf2),1,mean)
sv[,'norm_mean'] <- apply(data.frame(sv$norm1,sv$norm2),1,mean)
sv[,'split_total'] <- sv$bp1_split+sv$bp2_split
sv[,'bp1_win_norm_cov'] <- sv$bp1_win_norm/10
sv[,'bp2_win_norm_cov'] <- sv$bp2_win_norm/10
sv[,'win_norm_mean'] <- apply(data.frame(sv$bp1_win_norm_cov,sv$bp2_win_norm_cov),1,mean)
return(sv)
}
load_simu_data <- function(sim_dir, support_adjust) {
sim_run <- list.files(sim_dir)
sim_run <- sim_run[grep('^tumour', sim_run)]
simu_table <- NULL
cdf <- NULL
for(i in 1:length(sim_run)) {
info <- strsplit(sim_run[i],'_')[[1]][2:3]
sample <- paste(info,collapse='_')
purity <- as.numeric(strsplit(info[1],'p')[[1]][2])/100
var_type <- info[2]
svs <- load_sv_data(sim_dir, sim_run[i])
svs <- svs[svs$classification%in%var_type,]
svs$raw_vaf <- svs$mean_vaf
svs$adjusted_support <- svs$support * (1 + (support_adjust * purity))
svs$adjusted_norm <- svs$norm_mean
if (length(grep('DUP', sample))>0) {
svs$adjusted_norm <- svs$adjusted_norm * (1 - purity/2)
}
svs$adjusted_depth <- svs$adjusted_support + svs$adjusted_norm
svs$adjusted_vaf <- svs$adjusted_support / svs$adjusted_depth
assign(sample, svs, envir = .GlobalEnv)
target_vaf <- purity/2
cdf <- rbind(cdf,data.frame(sample=sample,
purity=purity,
target_vaf=target_vaf,
type=strsplit(sample,'_')[[1]][2],
spanning=svs$spanning,
split=svs$split_total,
support=svs$support,
side=svs$preferred_side,
norm1=svs$norm1,
norm2=svs$norm2,
norm=svs$adjusted_norm,
norm_mean=svs$norm_mean,
anomalous=svs$anomalous,
vaf=svs$adjusted_vaf))
accu_p1 <- sum(svs$adjusted_vaf>=(target_vaf-.1) &
svs$adjusted_vaf<=(target_vaf+.1))/nrow(svs)
accu_p05 <- sum(svs$adjusted_vaf>=(target_vaf-.05) &
svs$adjusted_vaf<=(target_vaf+.05))/nrow(svs)
simu_table <- rbind(simu_table,
data.frame(type=info[2],purity=purity,
target_vaf=target_vaf,
mean_vaf=mean(svs$support/(svs$norm_mean+svs$support)),
mean_vaf_adj=mean(svs$adjusted_support/svs$adjusted_depth),
accuracy_p1=accu_p1,
accuracy_p05=accu_p05,
vaf_adj_std=sqrt(var(svs$adjusted_vaf)),
vaf_std=sqrt(var(svs$mean_vaf)),
supp_adj_std=sqrt(var(svs$adjusted_support)),
norm_adj_std=sqrt(var(svs$adjusted_norm)),
supp_std=sqrt(var(svs$support)),
norm_std=sqrt(var(svs$norm_mean)))
)
}
return(list(simu_table,cdf))
}
plot_vaf_comp <- function(sv, type) {
if (type %in% c('Duplications')) {
sv1 <- data.frame(VAF=sv[,'raw_vaf'],VAF_type='Raw')
sv2 <- data.frame(VAF=sv[,'adjusted_vaf'],VAF_type='Adjusted')
sv <- rbind(sv1, sv2)
} else {
sv <- data.frame(VAF=sv[,'adjusted_vaf'],VAF_type='Raw')
}
sv_comp <- ggplot(sv, aes(fill=VAF_type)) +
geom_histogram(data=sv, aes(x=VAF, y=..density..),
binwidth=0.025, colour="black", alpha=0.6) +
geom_density(aes(x=VAF), alpha=0.1) + theme_minimal() +
theme(panel.grid.major = element_blank(),
panel.grid.minor = element_blank(), panel.background = element_blank()) +
scale_x_continuous(limits=c(0,0.8),breaks=seq(0,1,4*0.05)) + labs(fill='') +
theme(text = element_text(size=22), legend.position="bottom") +
scale_fill_manual(values = c("#33a02c","#1f78b4")) + ylab('Density') +
ggtitle(type)
if (type %in% c('Duplications')) {
sv_comp <- sv_comp + geom_histogram(data=sv2, aes(x=VAF, y=..density..),
binwidth=0.025, colour="black", alpha=0.6) +
geom_density(aes(x=VAF), alpha=0.1)
}
sv_comp <- sv_comp + geom_vline(xintercept=0.5, size=2)
return(sv_comp)
}
plot_sv_adjustment <- function(x, type, title) {
p1 <- ggplot(x[x$type%in%type,], aes(x=VAF, y=Purity, color=as.factor(VAF_type))) +
geom_abline(intercept=0, slope = 2, size=2, color='darkgrey', linetype='dotted') +
geom_point(size=2.5) + geom_line() + ggtitle(title) +
scale_y_continuous(breaks = seq(0,1,0.2), limits = c(0,1)) +
scale_x_continuous(breaks = seq(0,1,0.2), limits=c(0,0.8)) +
theme_minimal() + theme( panel.background = element_blank(),
legend.title=element_blank(), text = element_text(size=18), legend.position="bottom",
axis.text=element_text(size=15),
plot.title = element_text(hjust = 0.5)) +
scale_colour_manual(values = c("#1f78b4", "#33a02c"))
return(p1)
}
simu_names <- c("p20_DEL","p20_INV","p20_TRX","p20_DUP","p40_DEL","p40_INV","p40_TRX","p40_DUP",
"p60_DEL","p60_INV","p60_TRX","p60_DUP","p80_DEL","p80_INV","p80_TRX","p80_DUP",
"p100_DEL","p100_INV","p100_TRX","p100_DUP")
sim_dir <- 'simulation_results/'
support_adjust <- 0
results <- load_simu_data(sim_dir, support_adjust)
simu_table <- results[[1]]; cdf <- results[[2]]
```
# Simulation result using bowtie2-aligned data
```{r}
cdf$sample <- factor(cdf$sample,levels=simu_names)
mean_vaf <- mean(abs(simu_table[simu_table$target_vaf!='DUP',]$mean_vaf-simu_table[simu_table$target_vaf!='DUP',]$target_vaf))
adjusted_vaf <- mean(abs(simu_table[simu_table$target_vaf!='DUP',]$mean_vaf_adj-simu_table[simu_table$target_vaf!='DUP',]$target_vaf))
print(paste('Mean VAF deviation:', round(mean_vaf, 4),
'; Adjusted VAF deviation', round(adjusted_vaf,4)))
```
## VAF accuracy
```{r, simu_accuracy, fig.width=12, fig.height=4}
p1 <- ggplot(simu_table, aes(x=purity, y=accuracy_p05, color=type)) + theme_minimal() +
scale_y_continuous(lim=c(0,1)) + geom_point(size=5) + geom_line() + ggtitle('Within 5% accuracy')
p2 <- ggplot(simu_table, aes(x=purity, y=accuracy_p1, color=type)) + theme_minimal() +
scale_y_continuous(lim=c(0,1)) + geom_point(size=5) + geom_line() + ggtitle('SVs within 10% accuracy')
grid.arrange(p1,p2,ncol=2)
```
## VAF adjustment (100% purity)
```{r, simu_adjustment, fig.width=12, fig.height=12}
dup_comp <- plot_vaf_comp(p100_DUP, 'Duplications')
del_comp <- plot_vaf_comp(p100_DEL, 'Deletions')
inv_comp <- plot_vaf_comp(p100_INV, 'Inversions')
trx_comp <- plot_vaf_comp(p100_TRX, 'Translocations')
grid.arrange(dup_comp, del_comp, inv_comp, trx_comp, ncol = 2)
```
## VAF adjustment at all purity levels
```{r, simu_adj_pur, fig.width=12, fig.height=6}
simu_vars <- c(paste(rep('p',5), seq(20,100,20), rep('_', 5), rep('DEL',5),sep=''),
paste(rep('p',5), seq(20,100,20), rep('_', 5), rep('DUP',5),sep=''),
paste(rep('p',5), seq(20,100,20), rep('_', 5), rep('TRX',5),sep=''),
paste(rep('p',5), seq(20,100,20), rep('_', 5), rep('INV',5),sep=''))
simu_stats <- NULL
for(var in simu_vars) {
info <- strsplit(var, '_')[[1]]
svs <- get(var)
purity <- as.numeric(strsplit(info[1],'p')[[1]][2])/100
target_vaf <- purity/2
simu_stats <- rbind(simu_stats,
data.frame(type=info[2],Purity=purity,
target_vaf=target_vaf,
raw=mean(svs$raw_vaf),
adjusted=round(mean(svs$adjusted_vaf),4)
)
)
}
x <- melt(simu_stats,id=c('type','Purity','target_vaf'))
colnames(x)[4:5] <- c('VAF_type', 'VAF')
dup <- plot_sv_adjustment(x, 'DUP', 'Duplications')
oth <- plot_sv_adjustment(x[x$VAF_type%in%'raw',], c('INV', 'DEL', 'TRX'), 'Other SVs')
# pdf('vaf_adjustment_scatter_v2.pdf', width=7.5, height=3.5)
grid.arrange(dup, oth, ncol=2)
# dev.off()
```