重新排序条形图 - R, ggplot, position = "dodge"

Question

我使用以下代码制作了一个数据框：

> p <- rep(c("5e-8", "0.05", "1"), 2)
> pgc1_rsq <- c(0.0037, 0.0726, 0.0847)
> meta_rsq <- c(0.0263, 0.1829, 0.1753)
> values <- c(pgc1_rsq, meta_rsq)
> Type <- c(rep("PGC1", 3), rep("PGC meta-analysis", 3))
> mydata <- data.frame(p, values)
> mydata$p <- factor(mydata$p, levels = c("5e-8", "0.05", "1"))

我使用以下代码创建了一个条形图：

> plot <-ggplot(mydata, aes(p, values))
> plot +geom_bar(stat = "identity", aes(fill = Type), position = "dodge") + xlab("P-value threshold") + ylab("Proportion of variance explained (Nagelkerke Rsq)")

产生了这个情节：

我现在想重新排序条形 - 以便在每对中“PGC1”位于“PGC 元分析”之前。我已经尝试使“值”和“类型”因素以及 PGC1 排在第一位的排序级别是这样的：

> mydata$value <- factor(mydata$value, levels = c("pgc1_rsq", "pgc_meta"))

但这给了我一个带有“类型”的错误消息，并且没有产生带有“值”的预期结果。

我们将不胜感激输入和建议。谢谢。

Answer 1

为什么不在dataframe 中包含Type：

p <- rep(c("5e-8", "0.05", "1"), 2)
pgc1_rsq <- c(0.0037, 0.0726, 0.0847)
meta_rsq <- c(0.0263, 0.1829, 0.1753)
values <- c(pgc1_rsq, meta_rsq)
Type <- c(rep("PGC1", 3), rep("PGC meta-analysis", 3))
mydata <- data.frame(p, values,Type)
mydata$Type <- factor(mydata$Type, levels = c("PGC1","PGC meta-analysis"))
mydata$p <- factor(mydata$p, levels = c("5e-8", "0.05", "1"))

plot <-ggplot(mydata, aes(p, values))
plot +geom_bar(stat = "identity", aes(fill = Type), position = "dodge") + xlab("P-value threshold") + ylab("Proportion of variance explained (Nagelkerke Rsq)")

Answer 2

尝试在 geom_bar 中分解填充：

plot <- ggplot(mydata, aes(p, values))
plot + geom_bar(stat = "identity", aes(fill = factor(Type, levels = c("PGC1", "PGC meta-analysis"))), position = "dodge") + 
xlab("P-value threshold") + 
ylab("Proportion of variance explained (Nagelkerke Rsq)") 

Answer 3

稍微清理一下你的数据框

mydata <- data.frame(p = factor(rep(c("5e-8", "0.05", "1"), 2), levels = c("5e-8", "0.05", "1")), 
                 values =c(0.0037, 0.0726, 0.0847, 0.0263, 0.1829, 0.1753), 
                 Type = factor(c( rep("PGC meta-analysis", 3),rep("PGC1", 3)), levels = c("PGC1", "PGC meta-analysis")))



ggplot(mydata, aes(p, values))+
  geom_bar(stat = "identity", aes(fill = Type), position = "dodge") + 
  xlab("P-value threshold") + 
  ylab("Proportion of variance explained (Nagelkerke Rsq)")

Answer 4

这些都是很好的答案。最优雅的方法是使用来自tidyr 的非常有用的gather。大多数 ggplots 必须的

library(dplyr)
library(tidyr)
library(ggplot2)

data.frame(p = c("5e-8", "0.05", "1"),
           `pgc1_rsq` = c(0.0037, 0.0726, 0.0847),                    # change names here
           `pgc_meta` = c(0.0263, 0.1829, 0.1753)) %>% 
  gather(Type, Value, 2:3) %>% 
  mutate(Type = factor(Type, levels = c("pgc1_rsq", "pgc_meta"))) %>% # and here if you need
  ggplot(aes(x = p, y = Value, fill = Type)) +
  geom_bar(stat = "identity", position = "dodge") +
  xlab("P-value threshold") + 
  ylab("Proportion of variance explained (Nagelkerke Rsq)")