如何在ggplot中对geom_col进行分组和标记？

Question

我正在尝试使用 R 和 ggplot2 从 GraphPad 重现图形。 在少数情况下，我有少量样本的值：

df <- data.frame(Sample = c("blank", "blank", "blank", "A","A","A","A","A","A","B","B","B","B","B", "B", "C","C","C","C","C","C"), 
                 Condition = c("control", "control", "control", "C1", "C1", "C1", "C2", "C2", "C2","C1", "C1", "C1", "C2", "C2", "C2","C1", "C1", "C1", "C2", "C2", "C2"),
                 Value = c(0.719, 1.25, 0.687, 8.19, 4.68, 3.53, 14.1, 7.11, 8.8, 6.48, 7.05, 4.82, 6.32, 4.97, 6.97, 5.5, 7.22, 6.89, 8.89, 6.83, 8.73))

df$Sample <- factor(df$Sample, levels = c("blank", "A", "B", "C")) 
df$Condition <- factor(df$Condition, levels = c("control", "C1", "C2"))

我正在努力将它们组织和绘制为分组（按条件）和标记（按样本），如下所示：

我尝试使用fill 和position="dodge"，但这完全不是我想要的：

df %>%
  ggplot + 
  aes(x = Sample, y = Value, fill = Condition) + 
  geom_col(position = "dodge")

请注意，我还想从图例中排除空白。

我正在尝试创建、绘制和标记另一列，例如：

df <- df %>%
  mutate(Sample.Condition = paste(Sample, Condition, sep = "."))

..但它变得（太？）复杂了。为了学习，我正在寻找关于如何做到这一点的简单而整洁的解决方案。 谢谢！

Answer 1

我不知道你是否成功得到你的阴谋，如果你有兴趣，我找到了一种方法：

首先，我计算Values 的Values 的sd 和@Gregor 的建议，我创建了一个新列，它是@987654330 的串联@和Condition

library(dplyr)
df2 = df%>%
  group_by(Sample,Condition) %>%
  summarise(Mean = mean(Value), Sd = sd(Value)) %>%
  mutate(New_Var = paste0(Sample,Condition))

然后，我们可以绘制数据：

library(ggplot2)
ggplot(df2, aes(x = New_Var, y = Mean, fill = Condition)) +
  geom_bar(stat = "identity", color = "black",position = position_dodge(), width = 0.7) +
  geom_errorbar(aes(ymin = Mean - Sd, ymax = Mean + Sd), width = .2, position = position_dodge(.9)) +
  scale_fill_manual(values = c("black","grey","red"),
                    labels = c("Control","Condition 1", "Condition 2")) +
  scale_x_discrete(limits = c("blankcontrol","AC1","BC1","CC1","AC2","BC2","CC2") , labels = c("Blank","A","B","C","A","B","C")) +
  theme(axis.text.x = element_text(face = "bold",angle = 45),
        legend.title = element_blank()) +
  xlab("") +
  scale_y_continuous(limits = c(0,15), breaks = c(0,5,10,15))

情节看起来与您从 GraphPad 获得的情节非常相似。我同意这不是一件容易的事，但如果你真的想要这个情节，你可以得到它。

编辑 - 在图表上添加单个值

library(dplyr)
dfX= df %>%
  mutate(New_Var2 = paste0(Sample,Condition))

library(ggplot2)
ggplot(df2, aes(x = New_Var, y = Mean, fill = Condition)) +
  geom_bar(stat = "identity", color = "black",position = position_dodge(), width = 0.7) +
  geom_errorbar(aes(ymin = Mean - Sd, ymax = Mean + Sd), width = .2, position = position_dodge(.9)) +
  scale_fill_manual(values = c("black","grey","red"),
                    labels = c("Control","Condition 1", "Condition 2")) +
  scale_x_discrete(limits = c("blankcontrol","AC1","BC1","CC1","AC2","BC2","CC2") , labels = c("Blank","A","B","C","A","B","C")) +
  theme(axis.text.x = element_text(face = "bold",angle = 45),
        legend.title = element_blank()) +
  xlab("") +
  scale_y_continuous(limits = c(0,15), breaks = c(0,5,10,15))+
  geom_jitter(data = dfX, aes(x = New_Var2, y = Value), position=position_jitter(0.3), show.legend = F)

您会得到以下图表：

但是，每个条件只有三个点，我宁愿将均值表示为单个点，并以 sd 作为误差线。类似的东西。

ggplot(df2, aes(x = New_Var, y = Mean, group = Condition)) +
  geom_point(aes(shape = Condition, color= Condition), stat = "identity", position = position_dodge(), size = 2) +
  geom_errorbar(aes(ymin = Mean - Sd, ymax = Mean + Sd, color = Condition), width = .2, position = position_dodge(.9)) +
  scale_shape_manual(values=c(15, 16, 17))+
  scale_color_manual(values = c("black","darkgrey","darkred"), labels = c("Control","Condition 1", "Condition 2")) +
  scale_x_discrete(limits = c("blankcontrol","AC1","BC1","CC1","AC2","BC2","CC2") , labels = c("Blank","A","B","C","A","B","C")) +
  theme(axis.text.x = element_text(face = "bold",angle = 45),
        legend.title = element_blank()) +
  xlab("") +
  scale_y_continuous(limits = c(0,15), breaks = c(0,5,10,15))+
  geom_jitter(data = dfX, aes(x = New_Var2, y = Value, shape = Condition), 
              position=position_jitter(0.3), color = adjustcolor("black",alpha.f = 0.6), 
              show.legend = F, size = 2)

结果图：

但这只是我的个人意见，这取决于你;）

Answer 2

最简洁的解决方案可能是引入构面。

df %>%
  ggplot + 
  aes(x = Sample, y = Value, fill = Condition) + 
  geom_col() +
  facet_wrap(~Condition)