Tech Log Plot
2026年3月20日 研究日志
绘制可用的会议展示用图
↓根据之前生成的生存分析结果绘制KM生存曲线
library(readr)
library(tidyr)
library(here)
library(ggplot2)
library(dplyr)
library(patchwork)
library(gridExtra)
library(tibble)
library(scales)
options(repr.plot.width=8.8, repr.plot.height=6, repr.plot.res=600)
# Basic information
PatientsInSGL <- 98897
PatientsInGLP <- 98902
# Input data(Additional annotations have been remove before)
data <- read.csv(here("Outcome_7_Result_b_KM_graph.csv"))
colnames(data) <- c("Time",
"Surv1", "Lower1", "Upper1",
"Surv2", "Lower2", "Upper2")
# transfer to long format
plot_data <- data %>%
pivot_longer(
cols = -Time,
names_to = c(".value", "Cohort"),
names_pattern = "([A-Za-z]+)(\\d)"
) %>%
mutate(
Cohort = paste0("Cohort", Cohort)
)
plot_data <- plot_data %>%
mutate(Cohort = recode(Cohort,
"Cohort1" = "SGLT2i",
"Cohort2" = "GLP1RA"))
# Check data frame
if (nrow(plot_data) == 0 || all(is.na(plot_data$Lower))) {
stop("Data conversion error: No confidence interval data was detected after conversion. Please check the original CSV structure.")
}
# Plot
# Color
color_SGLT2i <- "#8E00FA"
fill_SGLT2i <- adjustcolor(color_SGLT2i, alpha.f = 0.2)
color_GLP1RA <- "#117833"
fill_GLP1RA <- adjustcolor(color_GLP1RA, alpha.f = 0.2)
# label data
label_data <- data.frame(
Cohort = c("SGLT2i", "GLP1RA"),
Time = c(1000, 1000), # fixed in x=1000
Surv = c(
approx(
x = plot_data$Time[plot_data$Cohort == "SGLT2i"],
y = plot_data$Surv[plot_data$Cohort == "SGLT2i"],
xout = 1000
)$y,
approx(
x = plot_data$Time[plot_data$Cohort == "GLP1RA"],
y = plot_data$Surv[plot_data$Cohort == "GLP1RA"],
xout = 1000
)$y
)
) %>%
mutate(
Surv_offset = ifelse(Cohort == "GLP1RA", Surv - 0.05, Surv + 0.05)
)
# Create number table
number_table <- data %>%
filter(Time %in% c(250, 500, 750, 1000)) %>%
select(Surv1, Surv2) %>%
`row.names<-`(as.character(c(250, 500, 750, 1000))) %>%
rbind(data.frame(Surv1 = 1, Surv2 = 1, row.names = "0")) %>%
slice(c(n(),1:n()-1)) %>%
mutate(Surv1 = Surv1 * PatientsInSGL, Surv2 = Surv2 * PatientsInGLP) %>%
mutate(across(everything(), round))
Ready_for_plot <- t(number_table) %>%
as.data.frame(check.names = FALSE) %>%
rownames_to_column("Drug") %>%
mutate(Drug = recode(Drug,
"Surv1" = "SGL T2i",
"Surv2" = "GLP-1RA")) %>%
pivot_longer(
cols = -Drug,
names_to = "Time",
values_to = "Value"
) %>%
mutate(Value_num = as.numeric(Value)) %>%
group_by(Drug) %>%
mutate(
Diff = first(Value_num) - Value_num,
# table format
Value_formatted = sprintf(
"%s (%s)",
format(Value_num, big.mark = ",", trim = TRUE),
format(round(Diff), big.mark = ",", trim = TRUE)
)
) %>%
ungroup() %>%
select(Drug, Time, Value_formatted) %>%
pivot_wider(
names_from = Time,
values_from = Value_formatted
) %>%
column_to_rownames("Drug")
# Main plot----
main_plot <- ggplot(plot_data, aes(x = Time, y = Surv)) +
geom_ribbon(
aes(ymin = Lower, ymax = Upper, fill = Cohort),
alpha = 0.2,
linetype = "dashed",
size = 0.3,
color = "gray50"
) +
geom_step(aes(color = Cohort), linewidth = 1.2) +
# y axis scale
scale_y_continuous(
limits = c(0.969, 1.0),
breaks = seq(0.97, 1.0, by = 0.01), # identify y-axis range
expand = c(0, 0),
labels = scales::number_format(accuracy = 0.01)
) +
# x axis scale
scale_x_continuous(
expand = c(0, 0),
breaks = function(limits) {
# ensure 0 exist
breaks <- unique(c(0, seq(0, max(limits), by = 250))) # x-axis range: 250days
breaks[breaks <= max(limits)]
}
) +
labs(
x = "Days",
y = "Survival Probability",
title = "KM Curve(Mortality)"
) +
scale_color_manual(
name = NULL,
values = c("SGLT2i" = color_SGLT2i, "GLP1RA" = color_GLP1RA),
labels = c("SGLT2i" = "SGLT2i", "GLP1RA" = "GLP-1RA")
) +
scale_fill_manual(
name = NULL,
values = c("SGLT2i" = fill_SGLT2i, "GLP1RA" = fill_GLP1RA),
labels = c("SGLT2i" = "SGLT2i", "GLP1RA" = "GLP-1RA")
) +
annotate("text",
x = Inf, y = Inf,
label = "P = 0.0003",
hjust = 1.1,
vjust = 1.5,
size = 5) +
theme_bw(base_size = 14) +
theme(
legend.position = c(0.02, 0.02),
legend.justification = c(0, 0),
legend.direction = "vertical",
legend.title = element_blank(),
legend.key = element_blank(),
legend.key.size = unit(0.8, "cm"),
panel.grid = element_blank(),
panel.border = element_rect(color = "black", fill = NA, linewidth = 0.8),
plot.title = element_text(hjust = 0.5, face = "bold"),
axis.line = element_blank(),
# use scale line
axis.ticks = element_line(color = "black", size = 0.5), # scale line size
axis.ticks.length = unit(0.2, "cm"), # scale line length
axis.text = element_text(color = "black"),
axis.text.y = element_text(margin = margin(r = 10)),
plot.margin = margin(10, 20, 10, 20)
) +
coord_cartesian(clip = "off") +
# show 0 label
annotate("text",
x = 0, y = -Inf,
label = "",
vjust = 1.5,
hjust = 0.5,
size = 4)
main_plot
# Table plot----
table_data <- as.data.frame(Ready_for_plot)
x_min <- -150
x_max <- 1100
x_breaks <- c(0, 250, 500, 750, 1000)
table_plot <- ggplot() +
# title
ggtitle("All-cause mortality") +
scale_x_continuous(
breaks = x_breaks,
limits = c(x_min, x_max),
expand = c(0, 0)
) +
# Data
annotate("text",
x = x_breaks,
y = 1,
label = as.character(table_data["SGL T2i", ]),
size = 4.5,
fontface = "bold",
hjust = 0.5) +
annotate("text",
x = x_breaks,
y = 0,
label = as.character(table_data["GLP-1RA", ]),
size = 4.5,
fontface = "bold",
hjust = 0.5) +
# Drug name
annotate("text",
x = -100,
y = 1,
label = "SGL T2i",
size = 4.5,
fontface = "bold",
hjust = 1) +
annotate("text",
x = -100,
y = 0,
label = "GLP-1RA",
size = 4.5,
fontface = "bold",
hjust = 1) +
# Set y axis limit
ylim(-0.5, 1.5) +
coord_cartesian(clip = "off") +
theme_void() +
theme(
plot.margin = margin(20, 50, 10, 100),
plot.title = element_text(
size = 16,
face = "bold",
hjust = 0,
vjust = 1.5,
margin = margin(b = 3)
)
)
table_plot
# Final plot(merge main_plot & table_plot)
final_plot <- (
# adjust main plot
main_plot +
scale_x_continuous(
breaks = c(0, 250, 500, 750, 1000),
expand = c(0, 0)
) +
theme(
plot.margin = margin(0, 20, 0, 50) # Control margin
)
) / (
# adjust table plot
ggplot() +
# Table data
annotate("text",
x = c(0, 250, 500, 750, 1000),
y = 1,
label = as.character(table_data["SGL T2i", ]),
size = 4.5,
fontface = "bold",
hjust = 0.5) +
annotate("text",
x = c(0, 250, 500, 750, 1000),
y = 0,
label = as.character(table_data["GLP-1RA", ]),
size = 4.5,
fontface = "bold",
hjust = 0.5) +
# Drug name
annotate("text",
x = -73,
y = 1,
label = "SGLT2i",
size = 4.5,
fontface = "bold",
hjust = 1) +
annotate("text",
x = -74,
y = 0,
label = "GLP-1RA",
size = 4.5,
fontface = "bold",
hjust = 1) +
# Table title
annotate("text",
x = -160,
y = 1.5,
label = "All-cause mortality",
size = 6,
fontface = "bold",
hjust = 0) +
# Set location
ylim(-0.5, 1.5) +
scale_x_continuous(expand = c(0, 0)) +
coord_cartesian(clip = "off", xlim = c(-20, 1100)) + # Limit shown region
theme_void() +
theme(plot.margin = margin(0, 20, 10, 20)) # balance margin
) +
plot_layout(heights = c(3, 1))
# show final result
final_plot
# Output
formats <- c("pdf", "png")
for(fmt in formats) {
ggsave(paste0("KMFinal_plot_ver2.", fmt),
plot = final_plot,
dpi = ifelse(fmt=="png", 1200, 300),
width=8.8, height=6)
}
今天翻来覆去搞这个图真是耗费心神，今天就这样吧，只有再看看怎么给博客添加图片