求助：使用pdf()导出R绘图时图外图例无法显示的问题

我遇到的问题与多年前的一个Stack提问一致，但当时的所有解答都未能解决我的问题，因此重新发起提问。

需求是批量生成20张带外部图例的PDF图，代码如下：

for(i in 1:20){    
  pdf(paste0("Plot_BOTH_cM-Mb_CHR", i, ".pdf"), width = 10, height = 8) 
  par(mar = c(5,5,5,10), xpd = T)
  plot(cm_r10 ~ mb, data = subset(recomb_match[[i]], recomb_match[[i]][,"snp_both"]=="1_1"),
       las = 1, main = paste("chromosome",i, sep = " "),
       xlab = "Mb", ylab = "cM", ylim = c(0,120))
  points(cm_m10 ~ mb, data = subset(recomb_match[[i]], recomb_match[[i]][,"snp_both"]=="1_1"), 
  col = "red")
  legend("topright", inset=c(-0.81,-0.08), col = c("black","red"), 
         legend = c("RheMac10 (old)","Mmul_10 (new)"), bty="n", pch = 19,
         y.intersp=0.2, x.intersp = 0.2, xpd = T)
  dev.off()
}

当前现象：在RStudio中手动导出PDF时，图例可以正常显示；但使用pdf()函数批量导出时，图外的图例无法显示。

1. 调整边距与inset的匹配 #

当前右边距设置为10，但图例的x方向偏移量过大（inset=c(-0.81,-0.08)），可能超出PDF画布范围。尝试增大右边距并减小偏移量：

for(i in 1:20){    
  pdf(paste0("Plot_BOTH_cM-Mb_CHR", i, ".pdf"), width = 10, height = 8) 
  par(mar = c(5,5,5,15), xpd = TRUE) # 右边距从10调整为15
  plot(cm_r10 ~ mb, data = subset(recomb_match[[i]], recomb_match[[i]][,"snp_both"]=="1_1"),
       las = 1, main = paste("chromosome",i, sep = " "),
       xlab = "Mb", ylab = "cM", ylim = c(0,120))
  points(cm_m10 ~ mb, data = subset(recomb_match[[i]], recomb_match[[i]][,"snp_both"]=="1_1"), 
         col = "red")
  legend("topright", inset=c(-0.3, 0), # 减小x方向负偏移
         col = c("black","red"), 
         legend = c("RheMac10 (old)","Mmul_10 (new)"), bty="n", pch = 19,
         y.intersp=0.2, x.intersp = 0.2, xpd = TRUE)
  dev.off()
}

2. 使用xpd=NA替代xpd=T #

xpd=T仅允许内容在绘图区和边距内显示，xpd=NA则允许内容超出整个设备区域，可避免画布裁剪：

for(i in 1:20){    
  pdf(paste0("Plot_BOTH_cM-Mb_CHR", i, ".pdf"), width = 10, height = 8) 
  par(mar = c(5,5,5,10), xpd = NA) # 改为xpd=NA
  plot(cm_r10 ~ mb, data = subset(recomb_match[[i]], recomb_match[[i]][,"snp_both"]=="1_1"),
       las = 1, main = paste("chromosome",i, sep = " "),
       xlab = "Mb", ylab = "cM", ylim = c(0,120))
  points(cm_m10 ~ mb, data = subset(recomb_match[[i]], recomb_match[[i]][,"snp_both"]=="1_1"), 
         col = "red")
  legend("topright", inset=c(-0.81,-0.08), col = c("black","red"), 
         legend = c("RheMac10 (old)","Mmul_10 (new)"), bty="n", pch = 19,
         y.intersp=0.2, x.intersp = 0.2, xpd = NA)
  dev.off()
}

3. 用layout()分割画布，单独预留图例区域 #

这种方法更稳定，直接将画布分为绘图区和图例区，彻底避免偏移量问题：

for(i in 1:20){    
  pdf(paste0("Plot_BOTH_cM-Mb_CHR", i, ".pdf"), width = 12, height = 8) # 适当加宽画布
  # 布局：左侧80%为绘图区，右侧20%为图例区
  layout(matrix(c(1,2), ncol=2), widths=c(4,1))
  
  # 绘图区设置
  par(mar = c(5,5,5,2)) # 减小右边距
  plot(cm_r10 ~ mb, data = subset(recomb_match[[i]], recomb_match[[i]][,"snp_both"]=="1_1"),
       las = 1, main = paste("chromosome",i, sep = " "),
       xlab = "Mb", ylab = "cM", ylim = c(0,120))
  points(cm_m10 ~ mb, data = subset(recomb_match[[i]], recomb_match[[i]][,"snp_both"]=="1_1"), 
         col = "red")
  
  # 图例区设置
  par(mar = c(5,0,5,5)) # 左侧边距设为0，消除空隙
  plot.new() # 创建空绘图区
  legend("left", col = c("black","red"), 
         legend = c("RheMac10 (old)","Mmul_10 (new)"), bty="n", pch = 19,
         y.intersp=0.2, x.intersp = 0.2)
  dev.off()
}

4. 调整PDF设备参数 #

部分PDF阅读器对符号渲染支持不佳，可尝试关闭useDingbats：

pdf(paste0("Plot_BOTH_cM-Mb_CHR", i, ".pdf"), width = 10, height = 8, useDingbats = FALSE)

内容的提问来源于stack exchange，提问作者Cephalanthera