Go Ontology Bar Plot by ggplot

Data

Before running the codes, I’d like to have a brief introduces about my data.

There are three tables in total and name as In_all.table, Li_all.table, Mu_all.table.
It looks like:

Name	InCK_30	InCK_75	In30_75	Cate
localization	114	149	94	Biological process
single-organism process	285	357	225	Biological process
metabolic process	375	413	261	Biological process
multi-organism process	4	5	1	Biological process
immune system process	15	13	9	Biological process
multicellular organismal process	16	13	11	Biological process
cellular process	411	473	296	Biological process
reproduction	0	1	1	Biological process
reproductive process	0	1	1	Biological process

Quick Start

library('ggplot2')
library('reshape2')

theme_go <- theme(axis.text.x=element_text(angle=60,hjust=1),
      axis.title.x=element_text(hjust=0),
      plot.title = element_text(hjust = 0.5),
      panel.grid.major =element_blank(),
      panel.grid.minor = element_blank(),
      panel.background = element_blank(),
      axis.line = element_line(colour = "black"))

### Intestine

all <- read.table("In_all.table",sep='\t',header=T)
all_gg <- melt(all)

p1 <- ggplot(data=all_gg)+
      geom_bar(aes(x=Name,y=value,fill=variable),stat="identity")+
      facet_grid(. ~ Cate,scales = "free_x",space = "free") +
      labs(title="Intestine", x = "A", y="Countes")+ theme_go+
      theme(axis.text.x = element_text(size=15))
ggsave('GP_Q.png',wi=12,hei=7.4)

It is definitely the best result of GO bar plot.
But, thing gonna change when you want to align two or more others.

Plot Three files

library('ggplot2')
library('reshape2')

theme_go <- theme(axis.text.x=element_text(angle=60,hjust=1,size=12),
      axis.title.x=element_text(hjust=0),
      plot.title = element_text(hjust = 0.5),
      panel.grid.major =element_blank(),
      panel.grid.minor = element_blank(),
      panel.background = element_blank(),
      axis.line = element_line(colour = "black"))

### Intestine

all <- read.table("In_all.table",sep='\t',header=T)
all_gg <- melt(all)

p1 <- ggplot(data=all_gg)+
      geom_bar(aes(x=Name,y=value,fill=variable),stat="identity")+
      facet_grid(. ~ Cate,scales = "free_x",space = "free") +
      labs(title="Intestine", x = "A", y="Countes")+ theme_go

### Liver
all <- read.table("Li_all.table",sep='\t',header=T)
all_gg <- melt(all)

p2 <- ggplot(data=all_gg)+
      geom_bar(aes(x=Name,y=value,fill=variable),stat="identity")+
      facet_grid(. ~ Cate,scales = "free_x",space = "free") +
      labs(title="Liver", x = "B", y="Countes")+ theme_go

### Muscle
all <- read.table("Mu_all.table",sep='\t',header=T)
all_gg <- melt(all)

p3 <- ggplot(data=all_gg)+
      geom_bar(aes(x=Name,y=value,fill=variable),stat="identity")+
      facet_grid(. ~ Cate,scales = "free_x",space = "free") +
      labs(title="Muscle", x = "C", y="Countes")+ theme_go


p1/p2/p3
ggsave("GO.png", wi=10,height = 15)

Due to the name of categories are long and dense, it is hard to get an idea distribution of the chart.

One resolution is to combine three tables and sharing one X axis. Another is to annotate the categories at the side of the chart.

Combining X axis

theme_go <- theme(axis.text.x=element_text(angle=60,hjust=1,size=12),
      #axis.title.x=element_text(hjust=0),
      plot.title = element_text(hjust = 0.5),
      panel.grid.major =element_line(colour='grey'),
      #panel.grid.major =element_blank(),
      panel.grid.minor = element_blank(),
      panel.background = element_blank(),
      axis.line = element_line(colour = "black"))


File_list = c("In_all.table", "Li_all.table", "Mu_all.table")
Samples_list = c("Intestine", "Liver", "Muscle")
Degree_list = c("A","B","C")
TB = data.frame()
## Combining tables

for(i in c(1:length(File_list))){
    tmp = read.table(File_list[i],sep='\t',header=T)
    tmp$Samples= Samples_list[i]
    colnames(tmp)[2:4]=c("CK vs 30", "CK vs 75", "30 vs 75")
    TB = rbind(TB,tmp)
}

TB_melt <- melt(TB)
p <- ggplot(TB_melt)+  
  geom_bar(aes(x=Name,y=value,fill=variable),stat="identity")+  
  facet_grid(Samples ~ Cate,scales = "free",space = "free") +  
  labs(title="Intestine", x = "Categories", y="Countes") + theme_go

p <- p + theme( strip.background = element_rect(fill = '#f2f2f2'), strip.text = element_text(face = 'bold',size=15))
## we using cowplot to extent the axis
library(cowplot)
ggdraw() + draw_plot(p,0.05,0,0.95)
ggsave("GO_2.png",wi=12.7,hei = 10)

You may also like using heatmap to display the difference of counts among samples.

Heatmap

library(pheatmap)
library(ggplotify)
library(ggplot2)
library(reshape2)
library(patchwork)

## reading from data
all_In <- read.table("In_all.table",sep='\t',header=T)
all_Li <- read.table("Li_all.table",sep='\t',header=T)
all_Mu <- read.table("Mu_all.table",sep='\t',header=T)

## Merge
In_Li = merge(x=all_In,y=all_Li,by="Name", all.x = T, all.y = T)
In_Li_Mu = merge(x=In_Li,y=all_Mu,by="Name", all.x = T, all.y = T)

## Group
In_Li_Mu = In_Li_Mu[c(1,2,3,4,6,7,8,10,11,12,5,9,13)]
In_Li_Mu[13][which(is.na(In_Li_Mu[13])==TRUE),]=In_Li_Mu[12][which(is.na(In_Li_Mu[13])==TRUE),]
In_Li_Mu[13][which(is.na(In_Li_Mu[13])==TRUE),]=In_Li_Mu[11][which(is.na(In_Li_Mu[13])==TRUE),]
In_Li_Mu = In_Li_Mu[c(1:10,13)]

row.names(In_Li_Mu) = In_Li_Mu[[1]]


## Order
In_Li_Mu = In_Li_Mu[order(In_Li_Mu[[11]]),]
## Using raw counts
p1 <- pheatmap(In_Li_Mu[c(2:10)],  annotation_row =In_Li_Mu[11],cluster_rows = F, cluster_col=F, labels_row = "", legend = F, annotation_legend = F)
## Normalizing it with Log2(x+1)
p2 <- pheatmap(log(1+In_Li_Mu[c(2:10)]),  annotation_row =In_Li_Mu[11],cluster_rows = F, cluster_col=F, labels_row = "", legend = F, annotation_legend = F)
## Centering the matrix after log them
p3 <- pheatmap( data.frame(t(scale(t(log2(1+In_Li_Mu[c(2:10)])), scale=F))),  annotation_row =In_Li_Mu[11],cluster_rows = F, cluster_col=F, annotation_legend = F)

g1 = as.ggplot(p1)
g2 = as.ggplot(p2)
g3 = as.ggplot(p3)

(g1|g2)/g3
##ggsave("GO_3.png",width = 6.54,height = 10.5)

For have a better layout and easy to comparation, I tossed some legends.

Dots plot

library(ggplot2)
library(cowplot)
library(ggdendro)

TB_tree <- In_Li_Mu[2:10]
TB_tree[is.na(TB_tree)]=0
Tree  <- hclust(dist(TB_tree))
dendr    <- dendro_data(Tree, type="rectangle")
## plot the dendrogram; note use of color=cluster in geom_text(...)
p1 <- ggplot() +
  geom_segment(data=segment(dendr), aes(x=x, y=y, xend=xend, yend=yend)) +
  coord_flip() + labs(y="Distance")+
  theme(axis.line.y=element_blank(),
        axis.ticks.y=element_blank(),
        axis.text.y=element_blank(),
        axis.title.y=element_blank(),
        axis.text.x=element_text(angle=45,hjust=1),
        panel.background=element_blank(),
        panel.grid=element_blank())



TB <- melt(In_Li_Mu)
TB$Name = factor(TB$Name,levels = unique(TB$Name[order(TB$Cate)]))
TB$SubG = "Intest"
TB$SubG[grep( "Mu", TB$variable)] = "Muscle"
TB$SubG[grep( "Li", TB$variable)]  = "Liver"
TB$Name = factor(TB$Name, levels= dendr$labels$label)

p2 <- ggplot(TB,aes(x=variable,y=Name)) +
      geom_point(aes(size=value,color=Cate))+ theme_light()+
      facet_grid(~SubG, scales ="free", space = 'free') +
      labs(x='Samples',y='Category',title = "GO Anotation of DEGs")+
      theme(axis.text.x = element_text(angle = 45,hjust=1),
            axis.text.y = element_text(size=12),
            strip.text.x = element_text(size=12, face='bold'),
            legend.position = 'left')

ggdraw()+ draw_plot(p1,0.784,0.0,0.1,0.972) +
          draw_plot(p2,0,0.007,0.8)

ggsave('GO_Dots222.png', width = 9.4, height = 8.07)