[advanced drawing of single cell] 07. Display of KEGG enrichment results

The pictures drawn in this section are relatively new , What I marked with red arrows in the picture is pathway Class A Annotation information (big annotation, What I think , It's not a proper noun ), The colorful marks on the vertical axis are pathway Of second level notes (small annotation).「 How to get comments 」 Calculate a difficult point , I Last time I've already said ：KEGG The subordination of the path / How to get annotation information .

The whole picture reflects How many genes Fell into the corresponding classification .

「 Look at it dialectically 」, The whole picture is pathway notes , There is no specific pathway name , It is very similar to the usual enrichment analysis Dissimilarity . Replace the secondary notes in the figure with specific pathway Will be better . in addition , The abscissa in this figure is Gene number , But I don't think this value is important in enrichment analysis （ gene ratio More important than a single value ）, We can change it to p value .

---

Before you start drawing , Need to sort out a form , The table contains at least ：「pathway ID、pathway description、pathway notes /big annotation、 Several enrichment indicators 、 Enriched genes 」.

• 「 The first three messages 」, The forms of the previous lecture have been sorted out , It can be used ;
• 「 Several enrichment indicators 」 stay clusterprofiler There are also ;
• 「 Enriched genes 」 Looks like clusterprofiler do kegg Enrichment cannot directly show genes symbol, So the returned result needs to be processed a little .

---

The code name of the sorting table is run.R, As shown below ：

library(Seurat)
library(tidyverse)
library(xlsx)

testseu=readRDS("testseu.rds")
Idents(testseu)="anno_new"

###  Look for differential genes  #########################################################################
marker_celltype=FindAllMarkers(testseu,logfc.threshold = 0.8,only.pos = T)
#  Filter 
marker_celltype=marker_celltype%>%filter(p_val_adj < 0.01)
marker_celltype$d=marker_celltype$pct.1-marker_celltype$pct.2
marker_celltype=marker_celltype%>%filter(d > 0.2)
marker_celltype=marker_celltype%>%arrange(cluster,desc(avg_log2FC))
marker_celltype=as.data.frame(marker_celltype)
write.xlsx(marker_celltype,file = "markers_log2fc0.8_padj0.01_pctd0.2.xlsx",row.names = F,col.names = T)

###  Enrichment analysis  ###########################################################################
library(clusterProfiler)
library(org.Hs.eg.db)
R.utils::setOption("clusterProfiler.download.method","auto") #https://github.com/YuLab-SMU/clusterProfiler/issues/256

source("syEnrich.R")
syEnrich(marker_celltype,outpath = "markers_log2fc0.8_padj0.01_pctd0.2")

###  Select a type of cell as a demonstration  #######################################################
kegg.res=read.xlsx("markers_log2fc0.8_padj0.01_pctd0.2.KEGG.xls",sheetIndex = 1,as.data.frame = T,header = T)
kegg.res=kegg.res%>%filter(p.adjust < 0.05)
kegg.res=kegg.res%>%filter(cluster == "Endothelial")

#  Import the file mentioned above 
kegg_info=read.xlsx("kegg_info.xlsx",sheetIndex = 1,startRow = 3)
kegg_info=kegg_info[,c("ID","Pathway","big.annotion")]

#  Merge two tables 
kegg.res$ID=str_replace(kegg.res$ID,"hsa","")
kegg.res=kegg.res%>%inner_join(kegg_info,by = "ID")
write.table(kegg.res,file = "kegg.res.txt",quote = F,sep = "\t",row.names = F,col.names = T)

I take Single cell analysis Medium kegg Enrichment analysis as a demonstration , Just take one of them cluster Draw a picture based on the enrichment results of .

The enrichment code used in the above code is called syEnrich.R, This file only needs to input single cells seurat Object running FindAllMarkers The resulting differential genes , I can go back GO/KEGG Enrichment results , At the same time, genes enriched into a certain pathway symbol Will also be listed .

function run.R after , The final table is shown in the figure below ：

---

Then start drawing , The code name is 3.plot.R, I'm not going to do that here , The final figure is as follows ：

Get the code

Include all the code that will be used in this figure , Data sorting and drawing , Is it super sweet ！

This series will take 「 Time limited disclosure 」 Share code in a way ,24 Hours The inside is free . How to get beyond this time , The background to reply 2022A You know （ You may need to use your little hand to forward ）.

The network disk link of code and test data is as follows : link ：https://pan.baidu.com/s/1hebbeQH4DgYA8JqFWXO4dg Extraction code ：abo5

If you think the code is useful , You can give three / Two / One after another （ At the end of the article give the thumbs-up 、 Share 、 Click the small advertisement ） Help me :( A few times ago, there was a pitiful rise