• Hiast2 比對
• samtools sam 轉bam
• stringtie組裝轉錄本
• gffcompare將stringtie輸出的gtf文件與參考基因組的注釋文件做比較得到一- 個merged.combine.gtf
• 使用merged.combine.gtf 這個文件對每個樣本計算表達量，輸出文件存儲到ballgown文件夾下，這一步用到的命令是     stringtie -e -B -p 8 -G merged.combined.gtf -o ballgown/L01/L01.gtf output_bam/L01.sorted.bam

 

• 接下來是R語言的ballgown包讀入數據獲取基因和轉錄本的表達量 代碼是

library(ballgown)
library(genefilter)
library(dplyr)

pheno_data<-read.csv("pheno_data.txt",header=T)
grape_expr <- ballgown(dataDir = "ballgown",
                    samplePattern = "L0",
                    pData = pheno_data)

 

 

這一步可以拿到gene_id還有gene_name ,FPKM的表達量，cov對用的應該是reads count吧。

接下來是差異表達分析

代碼是

grape_expr_filter<-subset(grape_expr,
                          "rowVars(texpr(grape_expr))>1",
                          genomesubset=T)
results_genes <- stattest(grape_expr_filter,
                          feature = "gene",
                          covariate = "time_point",
                          getFC = TRUE,
                          meas = "FPKM")
#results_genes <- arrange(results_genes,pval)
results_genes%>%
  filter(abs(fc)>=2&pval<0.05) -> results_genes_diff

dim(results_genes_diff)
head(results_genes_diff)

現在有了基因id

 

接下來是使用clusterProfiler做go注釋

這里參考 https://guangchuangyu.github.io/cn/2017/07/clusterprofiler-maize/#disqus_thread

首先把這個基因id對應著轉換成 ENTREZID ，這里需要借助對應的gtf注釋文件

這里只關注蛋白編碼基因

grep 'gene_biotype "protein_coding"' 12X_genomic.gtf > 12X_protein_coding.gtf
#python
from gtfparse import read_gtf
known_proteincoding = read_gtf("12X_protein_coding.gtf")
known_proteincoding.to_csv("all_protein_coding.csv")

GO富集分析的R語言代碼

require(AnnotationHub)
hub<-AnnotationHub() #這一步對網路有要求
# aa<-query(hub,'zea')
# aa$title
# query(hub,'Malus domestica')
bb<-query(hub,"Vitis vinifera")
#bb$title
grape<-hub[['AH85815']]
# length(keys(grape))
# columns(grape)
protein_coding_all<-read.csv("all_protein_coding.csv",header=T)
df<-merge(results_genes_diff,grape_expr_filter@expr$trans,by.x="id",by.y="gene_id")
df1<-merge(df,protein_coding_all,by.x="gene_name",by.y="gene_id")
dim(df1)
gene_ids<- 
  df1$db_xref[!duplicated(df1$db_xref)]
gene_ids<-stringr::str_replace(gene_ids,"GeneID:","")


library(clusterProfiler)
bitr(keys(grape)[2],'ENTREZID',c("REFSEQ","GO",
                                 "ONTOLOGY","GENENAME",
                                 "SYMBOL"),grape)
res = enrichGO(gene_ids, 
               OrgDb=grape, pvalueCutoff=0.05, qvalueCutoff=0.05)
help(package="clusterProfiler")
dotplot(res)

最后的結果是