跟着Nature学作图：ggplot2包绘制森林图/误差棒图

 数据分析服务请访问以下链接：

文章发表技术服务，数据分析服务

数据和代码获取：请查看主页个人信息

大家好，今天我将介绍如何使用R语言ggplot2包绘制个性化误差棒图和森林图。

1 误差棒图（Error Bar Plot）：用于表示测量数据的中心趋势和误差范围。它通常包括一个表示中心趋势（如均值或中位数）的点，以及表示误差范围的线段或区域。误差范围可以表示标准差、标准误差、置信区间等。

2 森林图（Forest Plot）：用于可视化多个研究结果或效应估计及其置信区间。它通常以垂直线段表示每个效应估计的置信区间，并根据效应估计的大小进行排序。森林图常用于荟萃分析和医学研究中。

图片样式灵感来源于Nature杂志的一篇文章：

接下来我们用文章的实例数据来进行可视化展示：

Step1：数据载入

rm(list=ls())
library(tidyverse)
library(cowplot)

load('dnds.comp1.rda')
load('dnds.comp2.rda')
load('ALL_plot.rda')
load('shading.rda')

head(All_plot)

Step2：定义绘图元素

type1 <- "seeding"
type2 <- "primary_nonMetastatic"

col.palette  <- setNames(c("#762a83", "#5aae61"), c(type1, type2))
fill.palette <- setNames(adjustcolor(c("#762a83", "#c2a5cf", "#5aae61"), alpha.f = 0.75), c(paste0(type1, " favored"), paste0(type2, " and ", type1), paste0(type2, " favored")))
cols <- paste( "point", c( 'w', 'w_low', 'w_high'), sep = '_' )

Step3：绘制误差棒图

ggplot(All_plot) +
  geom_point(aes(x = gene_name, y = get(cols[1]), colour = type), position = position_dodge(0.5), size = 2) +
  geom_errorbar(aes(x = gene_name, y = get(cols[1]), ymin = get(cols[2]), 
                    ymax = get(cols[3]), colour = type), position = position_dodge(0.5), size = 0.75, width = 0.5) +
  scale_y_continuous(trans = 'log10', limits = c(0.1, 1300),
                     breaks = 10^(-1:3), labels = c('<0.1',10^(0:2), '>1000')) +
  geom_hline(yintercept = 1) +
  labs(x = "",
       y = paste0("dNdS point mutations")) +
  theme_cowplot(font_size = 16) +
  theme(legend.title = element_blank(), 
        legend.position = "none") +
  coord_flip() +
  ggsci::scale_color_aaas() +
  ggsci::scale_fill_aaas()

Step4：绘制森林图

ggplot(All_plot) +
  geom_pointrange(aes(x = gene_name, y = get(cols[1]), ymin = get(cols[2]), 
                      ymax = get(cols[3]), colour = type), position = position_dodge(0.5)) +
  geom_rect(data = shading, aes(xmin=x1, xmax=x2, ymin=y1, ymax=y2, fill=cat), alpha = 0.3) +
  geom_pointrange(aes(x = gene_name, y = get(cols[1]), ymin = get(cols[2]), 
                      ymax = get(cols[3]), colour = type), position = position_dodge(0.5), size = 0.75) +
  scale_color_manual(values = col.palette) +
  scale_fill_manual(values = fill.palette) +
  scale_y_continuous(trans = 'log10', limits = c(0.1, 1300),
                     breaks = 10^(-1:3), labels = c('<0.1',10^(0:2), '>1000')) +
  geom_hline(yintercept = 1) +
  labs(x = "",
       y = paste0("dNdS point mutations")) +
  theme_cowplot(font_size = 16) +
  theme(axis.text.x = element_text( angle = 45, hjust = 1),legend.title = element_blank())

划重点：上述代码【geom_pointrange】函数用来绘制森林图；【geom_errorbar】函数用于绘制误差棒图。【geom_rect】函数增加图层，用来映射三种不同的绘图背景，这也是一个不错的绘图技巧~

Step5：个性化：改变横坐标轴字体颜色

# 定义横坐标轴颜色
color = ifelse(as.character(All_plot %>%
                              arrange(gene_name) %>%
                              pull(gene_name) %>%
                              unique()) %in% as.character(dnds.comp1 %>%
                                                            mutate(qvec = p.adjust(pvec, "fdr")) %>% 
                                                            filter(qvec < 0.05) %>% 
                                                            pull(genestotest)), 
               col.palette[type1], ifelse(as.character(All_plot %>% 
                                                         arrange(gene_name) %>% 
                                                         pull(gene_name) %>% 
                                                         unique()) %in% as.character(dnds.comp2 %>% 
                                                                                       mutate(qvec = p.adjust(pvec,"fdr")) %>% 
                                                                                       filter(qvec < 0.05) %>% 
                                                                                       pull(genestotest)), col.palette[type2], "black"))
ggplot(All_plot) +
  geom_pointrange(aes(x = gene_name, y = get(cols[1]), ymin = get(cols[2]), 
                      ymax = get(cols[3]), colour = type), position = position_dodge(0.5)) +
  geom_rect(data = shading, aes(xmin=x1, xmax=x2, ymin=y1, ymax=y2, fill=cat), alpha = 0.3) +
  geom_pointrange(aes(x = gene_name, y = get(cols[1]), ymin = get(cols[2]), 
                      ymax = get(cols[3]), colour = type), position = position_dodge(0.5), size = 0.75) +
  scale_color_manual(values = col.palette) +
  scale_fill_manual(values = fill.palette) +
  scale_y_continuous(trans = 'log10', limits = c(0.1, 1300),
                     breaks = 10^(-1:3), labels = c('<0.1',10^(0:2), '>1000')) +
  geom_hline(yintercept = 1) +
  labs(x = "",
       y = paste0("dNdS point mutations")) +
  theme_cowplot(font_size = 16) +
  theme(axis.text.x = element_text( angle = 45, hjust = 1,color = color), legend.title = element_blank(), legend.position = "none")

 “森林图”获得本期代码和数据