How to make a bar plot for the following data?-CodePudding

I have this data that describes the occurrence of some cell types, in responding patients vs non responsive patients for some drug:

> dput(cellsdata)

structure(list(response = c(1, 7, 2, 3, 4, 2, 1, 2, 1, 5, 5, 
4, 2, 0, 0, 2, 0, 0, 1, 3, 1, 2, 1, 0, 4, 3, 2, 3, 1, 1, 5, 0, 
2, 2, 5, 4, 4, 2), no_response = c(1, 0, 0, 0, 0, 0, 1, 0, 1, 
0, 0, 0, 0, 3, 2, 3, 2, 0, 3, 0, 0, 0, 1, 1, 0, 1, 4, 0, 0, 0, 
0, 2, 1, 0, 0, 0, 0, 0)), class = "data.frame", row.names = c("Adipocytes", 
"B-cells", "Basophils", "CD4  memory T-cells", "CD4  naive T-cells", 
"CD4  T-cells", "CD4  Tcm", "CD4  Tem", "CD8  naive T-cells", 
"CD8  T-cells", "CD8  Tcm", "Class-switched memory B-cells", 
"DC", "Endothelial cells", "Eosinophils", "Epithelial cells", 
"Fibroblasts", "Hepatocytes", "ly Endothelial cells", "Macrophages", 
"Macrophages M1", "Macrophages M2", "Mast cells", "Melanocytes", 
"Memory B-cells", "Monocytes", "mv Endothelial cells", "naive B-cells", 
"Neutrophils", "NK cells", "pDC", "Pericytes", "Plasma cells", 
"pro B-cells", "Tgd cells", "Th1 cells", "Th2 cells", "Tregs"
))

I want a bar plot, to look something like this, and I also need it to be in order, from the highest bin on the left to the right.

CodePudding user response：

I guess something like this. First convert the row names to a column. Then turn them into a factor whose levels are ordered according to the sum of your response / no_response columns. Pivot to long format and use the new Response column as the fill variable. The rest is just styling.

library(tidyverse)

cellsdata %>%
  rownames_to_column("Cell") %>%
  mutate(n = response   no_response) %>%
  arrange(-n) %>%
  mutate(Cell = factor(Cell, Cell)) %>%
  select(-n) %>%
  pivot_longer(-Cell, names_to = "Response") %>%
  ggplot(aes(Cell, value, fill = Response))  
  geom_col(position = "stack")  
  theme_light(base_size = 16)  
  theme(axis.text.x = element_text(angle = 90, hjust = 1, vjust = 0.5))  
  scale_fill_manual(values = c("orange", "#3A43D8"),
                    labels = c("No response", "Response"), name = NULL)

CodePudding user response：

Using reshape2 with your data

df$row.names<-rownames(df)
long.df<-reshape2::melt(df,id=c("row.names"))

library(ggplot2)
ggplot(long.df, aes(
  fill = variable,
  y = value,
  x = reorder(row.names, -value)
))  
  geom_bar(position = "stack", stat = "identity")   labs(x = "row name")  
  theme(axis.text.x = element_text(
    angle = 90,
    vjust = 0.5,
    hjust = 1
  ))

Output: