用来自不同数据框ggplot R的误差线绘制点

Question

I have 3 dataframes one with actual data and others with upper and lower limits data respectively. How can I combine these dataframes and create a scatter plot on ggplot with error bars (where the upper and lower limit data serve as the bars)?

Here is the example of the data and plotting code

library(tidyverse)
library(dplyr)
library(ggplot2)

DT <- data.frame(Samples=c("D1", "D2", "D3", "D4"),
             ACE=c(0.184, -0.169, 0.054, 0.087),
             FLN=c(-0.257, -0.239, -0.009, 0.038),
             ANT=c(-0.166, -0.304, 0.238, "" ),
             stringsAsFactors=FALSE)



 LowerLim <- data.frame(Samples=c("D1", "D2", "D3", "D4"),
              ACE=c(-1.514, "", -1.644, -1.611),
              FLN=c("", "", "", -1.660),
              ANT=c(-1.865, -2.003, 1.937, ""),
              stringsAsFactors=FALSE)

 UpperLim <- data.frame(Samples=c("D1", "D2", "D3", "D4"),
                   ACE=c(-0.560, "", -0.689,-0.657),
                   FLN=c("", "", "", -0.706),
                   ANT=c(-0.911,-1.049,0.983, "" ),
                   stringsAsFactors=FALSE)

 DTlong <- gather(DT, Cds, Values, "ACE":"ANT", factor_key=TRUE)

DTlong$Samples <- as.character(DTlong$Samples)

DTlong$Samples <- factor(DTlong$Samples, levels=unique(DTlong$Samples))

ggplot(DTlong)+geom_point(aes(color=Samples, y=Values, x=Cds))+
     theme(axis.title.x = element_blank())+ 
     theme(axis.text.x=element_text(size=7, angle=90,hjust=0.95,vjust=0.2))+
    ggtitle("Points")

Answer 1

There are a couple of issues here. Firstly, some of your columns are in character format rather than numeric format because you have put missing values as "" rather than NA . I appreciate that this might look more natural when printing the data frame to the screen, but it means any column containing "" is no longer numeric, and can't be used for calculations or plotting the values the column appears to contain.

That means the first step should be converting the three value columns in each data frame to numeric values:

DT[-1] <- lapply(DT[-1], as.numeric)
LowerLim[-1] <- lapply(LowerLim[-1], as.numeric)
UpperLim[-1] <- lapply(UpperLim[-1], as.numeric)

The second issue is that your DT data frame and UpperLit data frame appear to be round the wrong way, since all of the values in DT are higher than the values in UpperLimit . For the purposes of this answer, I've had to assume this is a mistake. If your real data doesn't have this problem, you can omit this step:

tmp <- DT
DT <- UpperLim
UpperLim <- tmp
rm(tmp)

The third issue is that the data are in the wrong shape to be plotted. Since the row and column orders are all the same, we can achieve this by pivoting each data frame into long format, and just keeping the value column from the upper and lower lim data frames.

df <- cbind(pivot_longer(DT, -1, names_to = "Cds", values_to = "Values"),
            lower = pivot_longer(LowerLim, -1)$value,
            upper = pivot_longer(UpperLim, -1)$value)

Now we are ready to plot. We will need to use position_dodge on both the points and the error bars to spread them out:

ggplot(df, aes(Cds, Values, color = Samples)) +
  geom_point(position = position_dodge(width = 0.5), size = 3) +
  geom_errorbar(aes(ymin = lower, ymax = upper), width = 0.25, size = 1,
                position = position_dodge(width = 0.5)) +
  scale_color_brewer(palette = "Set1") +
  theme_minimal(base_size = 16) +
  theme(axis.title.x = element_blank()) +
  ggtitle("Points")