-
Notifications
You must be signed in to change notification settings - Fork 0
/
utils.R
174 lines (153 loc) · 6.23 KB
/
utils.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
library(epitools)
library(stringr)
my_riskratio <- function (x, conf.level = 0.95,
correction = FALSE, verbose = FALSE)
# with continuity correction:
# https://stats.stackexchange.com/questions/21298/confidence-interval-around-the-ratio-of-two-proportions
# http://www.zen103156.zen.co.uk/rr.pdf
# https://stats.stackexchange.com/questions/3112/calculation-of-relative-risk-confidence-interval
{
x <- matrix(x, ncol = 2, byrow = TRUE)
tmx <- table.margins(x)
p.exposed <- sweep(tmx, 2, tmx["Total", ], "/")
p.outcome <- sweep(tmx, 1, tmx[, "Total"], "/")
Z <- qnorm(0.5 * (1 + conf.level))
nr <- nrow(x)
small <- matrix(NA, nr, 3)
small[1, 1] <- 1
for (i in 2:nr) {
a0 <- x[1, 2]
b0 <- x[1, 1]
a1 <- x[i, 2]
b1 <- x[i, 1]
n1 <- a1 + b1
n0 <- a0 + b0
m0 <- b0 + b1
m1 <- a0 + a1
est <- (a1/n1)/((a0)/(n0))
logRR <- log(est)
# Delta method
# choose delta = 0.1 for only a slight correction, avoid divide by 0
SElogRR <- sqrt(1/(a1+0.1) - 1/(n1) + 1/(a0+0.1) - 1/(n0))
ci <- exp(logRR + c(-1, 1) * Z * SElogRR)
# Poisson model method
# https://books.google.com/books?id=oDlBZgyx54kC&pg=PP19&lpg=PP19
# alpha <- 1 - conf.level
# rr_low <- ifelse(a1 == 0, 0, n0/n1 * a1/(a0+1) / qf(1-alpha/2, 2*(a0+1), 2*a1))
# rr_high <- n0/n1 * (a1+1)/a0 * qf(1-alpha/2, 2*(a1+1), 2*a0)
# ci <- c(rr_low, rr_high)
small[i, ] <- c(est, ci)
}
pv <- tab2by2.test(x, correction = correction)
colnames(small) <- c("estimate", "lower", "upper")
rownames(small) <- rownames(x)
cn2 <- paste("risk ratio with", paste(100 * conf.level, "%",
sep = ""), "C.I.")
names(dimnames(small)) <- c(names(dimnames(x))[1], cn2)
rr <- list(x = x, data = tmx, p.exposed = p.exposed, p.outcome = p.outcome,
measure = small, conf.level = conf.level, p.value = pv$p.value,
correction = pv$correction)
rrs <- list(data = tmx, measure = small, p.value = pv$p.value,
correction = pv$correction)
attr(rr, "method") <-
attr(rrs, "method") <-
"small sample-adjusted UMLE & normal approx (Wald) CI"
if (verbose == FALSE) {
rrs
}
else rr
}
my_fish <- function(df) {
res <- df %>%
unlist() %>%
matrix(ncol = 2, byrow = TRUE) %>%
my_riskratio(correction = TRUE)
res$measure[2, 1:3] %>%
c(p_value = res$p.value[2,2]) %>% # get p-value from fisher exact
as.matrix() %>% t() %>% data.frame()
}
heat_theme_top <- function(){
theme(panel.grid = element_blank(),
strip.text.y = element_text(angle = 0),
axis.text.x = element_blank(),
legend.text = element_text(color = "white"),
legend.title = element_text(color = 'white'),
plot.margin = unit(c(1.5,0.2,1,0.5), "lines"),
panel.spacing.x = unit(10, "points"),
panel.spacing.y = unit(0, "points"))
}
heat_theme_bottom <- function(){
theme(panel.grid = element_blank(),
legend.title = element_text(),
strip.text.y = element_text(angle = 0),
axis.text.x = element_text(angle = 90, vjust = 0.5, hjust = 1),
plot.margin = unit(c(1.5,0.2,0.5,0.5), "lines"),
panel.spacing.x = unit(10, "points"),
panel.spacing.y = unit(0, "points"))
}
plot_enrich <- function(
plot_obs_exp_left, plot_obs_exp_right, nudge = 3){
enrichment_plot_left <- plot_obs_exp_left %>%
ggplot(aes(x = full_icd)) +
geom_hline(aes(yintercept = 0), linetype = 2) +
geom_pointrange(
aes(y = lestimate, ymin = llower, ymax = lupper, group = full_icd, color = presentation),
stroke = 0.3, fatten = 2
) +
coord_flip() +
scale_color_manual(values = levels(plot_obs_exp_left$presentation),
guide = FALSE) +
labs(x = NULL, y = bquote(Log[2] ~ 'enrichment, 95% CI')) +
theme(axis.title = element_text(size = 9),
panel.grid.minor = element_blank(),
plot.margin = margin(20, 2, 5.5, 5.5, unit = 'pt')) +
scale_x_discrete(position = 'top', labels = NULL) +
scale_y_reverse()
over_severe_icds <- plot_obs_exp_right %>%
filter(over_sev > 0) %>%
pull(full_icd)
enrichment_plot_right <- plot_obs_exp_right %>%
ggplot(aes(x = full_icd, y = Observed)) +
geom_segment(aes(xend = full_icd, yend = Expected, color = presentation)) +
scale_color_manual(values = levels(plot_obs_exp_right$presentation),
guide = FALSE) +
geom_point(data = . %>% pivot_longer(c(Observed, Expected), names_to = 'OE'),
aes(y = value, shape = OE), color = 'grey20') +
labs(x = NULL, y = 'Number of ever severe patients') +
coord_flip() +
theme(
axis.title = element_text(size = 9),
panel.grid.minor = element_blank(),
legend.position = c(0.78 , 0.1),
axis.text.y = element_text(
color = as.character(plot_obs_exp_right$presentation),
hjust = 0.5),
legend.background = element_blank(),
legend.key.height = unit(3, 'mm'),
legend.key.width = unit(3, 'mm'),
plot.margin = margin(20, 5.5, 8.5, 2, unit = 'pt')
) +
scale_y_sqrt(
breaks = c(100, 1000, 3000),
expand = expansion(add = c(0, 10))) +
geom_text(
data = . %>% filter(!(full_icd %in% over_severe_icds)),
nudge_y = - nudge, aes(label = round(over_sev, 0)), size = 2.5) +
geom_text(
data = . %>% filter(full_icd %in% over_severe_icds),
nudge_y = nudge, aes(label = round(over_sev, 0)), size = 2.5)
enrichment_plot <- cowplot::plot_grid(enrichment_plot_left, enrichment_plot_right,
rel_widths = c(1, 2.8))
enrichment_plot
}
median_age <- function(age_vec){
# calculate median age for each site, assuming uniform age distribution within
# bin containing the median. Given this is a fairly smooth distribution,
# the assumption is quite reasonable.
median_idx <- unname((age_vec['80to100']+ 1)/2)
idx_bef <- head(which(age_vec > median_idx, arr.ind = TRUE), 1)
range_bef <- names(idx_bef) %>% stringr::str_split('to') %>% unlist() %>% as.numeric()
cumu_age <- unname(c(age_vec[idx_bef - 1], age_vec[idx_bef]))
median_age <- range_bef[1] + (median_idx - cumu_age[1])/diff(cumu_age)*(diff(range_bef) + 1)
median_age
}