【问题标题】：ggplot geom_errorbar not in correct locationggplot geom_errorbar 位置不正确
【发布时间】：2021-06-10 00:43:40
【问题描述】：

我对 R 还很陌生，我一直在尝试使用以下代码制作 barplot：

表格

这张表是我要绘制的。

library(tidyverse)
library(parallel)
library(devtools)
library(scales)

peak_intersect_rep_elements <-
  tribble(
    ~Sample,                ~CellLine, ~Rep,  ~Total_peaks, ~Alu_intersect, ~L1_intersect, ~Antibody,
    "ADAR062", "HEK293xT", "rep1",   4407,           3329,        201, "p110/p150\nAb1",
    "ADAR062", "HEK293xT", "rep1",   19103,          3481,        8737, "p150\nAb3",
    "ADAR062", "HEK293xT", "rep1",   1782,           836,         109, "p110/p150\nAb4",
    "ADAR112", "HEK293xT", "rep1",   2269,           1852,        61, "p110/p150\nAb1",
    "ADAR112", "HEK293xT", "rep1",   28573,          5725,        17037, "p150\nAb3",
    "ADAR112", "HEK293xT", "rep1",   5115,           4448,        213, "p110/p150\nAb4",
    "ADAR112", "K562",     "rep1",   1367,           770,         49, "p110/p150\nAb1",
    "ADAR112", "K562",     "rep1",   12195,          2889,        5323, "p150\nAb3",
    "ADAR112", "K562",     "rep1",   1178,           656,         58, "p110/p150\nAb4",
    "ADAR004", "HEK293xT", "rep1",   4130,           3289,        136, "p110/p150\nAb1",
    "ADAR004", "HEK293xT", "rep2",   3447,           2816,        135, "p110/p150\nAb1",
    "ADAR004", "HEK293xT", "rep3",   4607,           3697,        176, "p110/p150\nAb1",
    "ADAR004", "HEK293xT", "rep1",   9711,           8450,        373, "p110/p150\nAb4",
    "ADAR004", "HEK293xT", "rep2",   7275,           6163,        294, "p110/p150\nAb4",
    "ADAR004", "HEK293xT", "rep3",   6789,           5704,        256, "p110/p150\nAb4"
  )

分组方式

在这里，我将这些值平均在一起并计算上表的标准偏差。

total_peak_intersect_rep_elements_mean <- peak_intersect_rep_elements %>%
  group_by(Sample, CellLine, Antibody, Rep) %>%
  summarise(Mean_total_peaks = mean(Total_peaks), 
            Mean_alu_intersect = mean(Alu_intersect),
            Mean_l1_intersect = mean(L1_intersect),
            SD_total_peaks = sd(Total_peaks),
            SD_alu_intersect = sd(Alu_intersect),
            SD_l1_intersect = sd(L1_intersect)) %>%
  ungroup() %>%
  replace_na(list(SD_total_peaks = 0, SD_alu_intersect = 0, 
                  SD_l1_intersect = 0))

total_peak_intersect_rep_elements_mean

绘制条形图

在这里，我将 Rep 列分解为级别并为它们命名。

temp_plot <- total_peak_intersect_rep_elements_mean %>%
  mutate(Rep = factor(Rep, levels = c("rep1", "rep2", "rep3"),
                      labels = c("s1 (n = 1)", "s2 (n = 1)", 
                                 "s3 (n = 1)")))

我正在尝试填写此处的 Rep 列。

temp_plot %>%
  ggplot(aes(Antibody, Mean_total_peaks, fill=Rep)) +
  geom_bar(stat = "identity", position = "dodge", size = 0.3, color = "black") +
  geom_errorbar(aes(ymin=Mean_total_peaks - SD_total_peaks, 
                    ymax=Mean_total_peaks + SD_total_peaks), 
                width=0.6, size = 0.8,
                position=position_dodge(.9)) +
  facet_wrap(~CellLine, scales = 'free_x') +
  scale_fill_manual(values = c("s1 (n = 1)" = "#125863",
                               "s2 (n = 1)" = "#2BA8B3",
                               "s3 (n = 1)" = "#D5E9EB")) +
  scale_y_continuous(trans = "log10", labels=comma_format(accuracy = 1)) + 
  labs(y = "ADAR1 IP peaks", x = '', fill = "Replicate") +
  guides(color = F) +
  theme_minimal() +
  NULL

在这之后，我最终得到了一个看起来像这样的图表。

但是，如上所述，误差线没有正确对齐。谁能帮帮我？

【问题讨论】：

您检查过 temp_plot 数据框是否正常？
嗨，乔治，我的数据框看起来不错。
我一直运行代码，所有的 sd 都为零。也许看看为什么会这样......（编辑）你在“分组依据”中的变量似乎唯一地定义了父集中的观察结果，所以没有剩下的变化。
我猜你不想在 group_by 函数中使用 'sample'
谢谢，乔治！我的 groupby 中的“样本”似乎是罪魁祸首。它现在正在工作！

标签： r ggplot2

【解决方案1】：

修正了图表！我需要从我的 group_by 语句中删除 'Sample'。

【讨论】：