如何让ggplot2中置信区间线条延伸至面板边框？

我用以下R代码生成了漏斗图的95%和99.7%置信区间：

########################
## create the dataset ##
########################

t=seq(0,1200)
lambda=0.02
sigma=sqrt(lambda/t)

alpha_95=0.05
z_both_95=qnorm(1-alpha_95/2)
lower_bound_95=lambda-z_both_95*sigma
upper_bound_95=lambda+z_both_95*sigma

alpha_997=0.003
z_both_997=qnorm(1-alpha_997/2)
lower_bound_997=lambda-z_both_997*sigma
upper_bound_997=lambda+z_both_997*sigma

df=data.frame(t,lambda,sigma,lower_bound_95,upper_bound_95,lower_bound_997,upper_bound_997)

#############################
## draw the CI on the plot ##
#############################

fp <- ggplot()+
  geom_line(aes(x = t, y = lower_bound_95), linetype = "dashed", color="orange", size=0.6, data = df) +
  geom_line(aes(x = t, y = upper_bound_95), linetype = "dashed", color="orange", size=0.6, data = df) +
  geom_line(aes(x = t, y = lower_bound_997), linetype = "dashed", color="red", size=0.6, data = df) +
  geom_line(aes(x = t, y = upper_bound_997), linetype = "dashed", color="red", size=0.6, data = df) +
  geom_hline(aes(yintercept = lambda), color="limegreen", size=0.6, data = df) +
  scale_x_continuous(limits = c(0,1200), breaks = seq(0,1200,by=150)) +
  scale_y_continuous(limits = c(-0.1,0.3),breaks = seq(-0.1,0.3,by=0.05)) +
  xlab("Total Exposure (Days)") + ylab("Exposure-Adjusted Incicence Rate") +
  theme_bw()
fp

当前效果：置信区间的橙色、红色虚线在x=0附近未延伸至面板左侧边框，仅从x>0的位置开始绘制。
目标效果：让橙色、红色的置信区间虚线像绿色的lambda水平线一样，左右两端都延伸至面板边框。

核心问题是原代码中t=seq(0,1200)包含t=0值，计算sigma=sqrt(lambda/t)时会得到无穷大，导致对应置信区间边界值无效，ggplot自动跳过这些无效点，所以虚线无法延伸到左侧边框。下面提供两种可行的修改方式：

方法1：修正数据集，补充边界有效点 #

从t=1开始生成数据，再手动补充x=0处的边界点（用y轴上下限替代无穷值），确保线条能连接到面板边框：

########################
## create the dataset ##
########################

t=seq(1,1200)  # 避免t=0的无效值
lambda=0.02
sigma=sqrt(lambda/t)

alpha_95=0.05
z_both_95=qnorm(1-alpha_95/2)
lower_bound_95=lambda-z_both_95*sigma
upper_bound_95=lambda+z_both_95*sigma

alpha_997=0.003
z_both_997=qnorm(1-alpha_997/2)
lower_bound_997=lambda-z_both_997*sigma
upper_bound_997=lambda+z_both_997*sigma

df=data.frame(t,lambda,sigma,lower_bound_95,upper_bound_95,lower_bound_997,upper_bound_997)

# 补充x=0处的点，匹配y轴极限值
df <- rbind(
  data.frame(t=0, lambda=lambda, sigma=NA,
             lower_bound_95=-0.1, upper_bound_95=0.3,
             lower_bound_997=-0.1, upper_bound_997=0.3),
  df
)

#############################
## draw the CI on the plot ##
#############################

fp <- ggplot()+
  geom_line(aes(x = t, y = lower_bound_95), linetype = "dashed", color="orange", size=0.6, data = df) +
  geom_line(aes(x = t, y = upper_bound_95), linetype = "dashed", color="orange", size=0.6, data = df) +
  geom_line(aes(x = t, y = lower_bound_997), linetype = "dashed", color="red", size=0.6, data = df) +
  geom_line(aes(x = t, y = upper_bound_997), linetype = "dashed", color="red", size=0.6, data = df) +
  geom_hline(aes(yintercept = lambda), color="limegreen", size=0.6, data = df) +
  scale_x_continuous(limits = c(0,1200), breaks = seq(0,1200,by=150)) +
  scale_y_continuous(limits = c(-0.1,0.3),breaks = seq(-0.1,0.3,by=0.05)) +
  xlab("Total Exposure (Days)") + ylab("Exposure-Adjusted Incicence Rate") +
  theme_bw()
fp

方法2：用geom_segment手动补全边框连接 #

直接用geom_segment绘制x=0到首个有效点的线段，配合原有的geom_line，实现完整的延伸效果：

fp <- ggplot(df, aes(x=t))+
  # 补全95%置信区间左侧边框连接
  geom_segment(aes(x=0, xend=df$t[1], y=-0.1, yend=df$lower_bound_95[1]), 
               linetype="dashed", color="orange", size=0.6) +
  geom_line(aes(y=lower_bound_95), linetype="dashed", color="orange", size=0.6) +
  geom_segment(aes(x=0, xend=df$t[1], y=0.3, yend=df$upper_bound_95[1]), 
               linetype="dashed", color="orange", size=0.6) +
  geom_line(aes(y=upper_bound_95), linetype="dashed", color="orange", size=0.6) +
  # 补全99.7%置信区间左侧边框连接
  geom_segment(aes(x=0, xend=df$t[1], y=-0.1, yend=df$lower_bound_997[1]), 
               linetype="dashed", color="red", size=0.6) +
  geom_line(aes(y=lower_bound_997), linetype="dashed", color="red", size=0.6) +
  geom_segment(aes(x=0, xend=df$t[1], y=0.3, yend=df$upper_bound_997[1]), 
               linetype="dashed", color="red", size=0.6) +
  geom_line(aes(y=upper_bound_997), linetype="dashed", color="red", size=0.6) +
  geom_hline(aes(yintercept = lambda), color="limegreen", size=0.6) +
  scale_x_continuous(limits = c(0,1200), breaks = seq(0,1200,by=150)) +
  scale_y_continuous(limits = c(-0.1,0.3),breaks = seq(-0.1,0.3,by=0.05)) +
  xlab("Total Exposure (Days)") + ylab("Exposure-Adjusted Incicence Rate") +
  theme_bw()
fp

内容的提问来源于stack exchange，提问作者Robin