在使用kallisto等基于cDNA数据库注释软件时，我们获得的gene expression matrix是transcript id，这需要进行id转换。

加载R包

  1. # devtools::install_github("BUStools/BUSpaRse")
  2. library(BUSpaRse)

获取数据

国内访问Ensembl较为困难，内置biomart包可以设置镜像，有时间可以修改一下。

  1. tr2g <- transcript2gene(
  2.             c("Homo sapiens", "Mus musculus"), 
  3.             type = "vertebrate",
  4.             ensembl_version = 100, 
  5.             kallisto_out_path = "./")

How to get the mapping relationship between transcriptid and geneid - 图1

biomart方法获取

上述方法也是基于biomart包访问Ensembl接口，直接用biomart可能更好。

获取基因组的基因list

  1. cat Homo_sapiens.GRCh38.101.gtf | awk -F'\t' '{if($3=="gene") {split($9,a,";"); print a[1]"\t"$5-$4};}' | sed 's/[gene_id |"|]//g' | sort -u > Homo_sapiens.GRCh38.101.genelength.tsv
  1. library(biomaRt)
  2. library(curl)
  4. genelist <- read.table("Homo_sapiens.GRCh38.101.genelength.tsv", header = T)
  5. human_mart <- useMart(host="www.ensembl.org", 
  6.                      biomart="ENSEMBL_MART_ENSEMBL", 
  7.                      dataset = "hsapiens_gene_ensembl")
  8. human_gene_all <- getBM(attributes=c("ensembl_gene_id", 
  9.                                          "entrezgene_id",
  10.                                          "external_gene_name", 
  11.                                          "ensembl_transcript_id", 
  12.                                          "ensembl_transcript_id_version",
  13.                                          "transcript_biotype", 
  14.                                          "description"),
  15.                             filters="ensembl_gene_id",
  16.                             values = genelist$Geneid,
  17.                             mart=human_mart)

Reference

  1. transcript2gene usage
  2. BUSpaRse github