如何获取hg19的CDS、UTR、intergenic、intron等的位置信息？

参考手册：

Hg19 regions for Intergenic, Promoters, Enhancer, Exon, Intron, 5-UTR, 3-UTR

怎么从gtf文件获取genome feature的区间

The coding region of a gene, also known as the CDS (from coding sequence), is that portion of a gene's DNA or RNA that codes for protein. The region usually begins at the 5' end by a start codon and ends at the 3' end with a stop codon.

CDS就是所有exon的组合

有细微的差别：

Exons = gene - introns

CDS = gene - introns - UTRs

therefore also:

CDS = Exons - UTRs

就是UTR也算作了exon了。

也就是exon算是一个比较大的概念，所以文章里用的CDS区域来统计，而不是exon。

获取数据的方法：

1. UCSC - 最全，最个性化

How to download the most similar annotation file as the author required from UCSC browser directly.

Go to UCSC brower and Tool, Table caterogy. and pick you reference genome and select right version under clade/genome/assembly
Make sure the group is "Genes and Gene Predictions"
Choose your preferred track (RefSeq/RefGene or UCSC gene/KnownGene)
Choose the table that gives gene information (RefSeq or KnownGene)
Select your region or the entire genome to get coordinates for
Select BED format as your output format
Name your output file
Click "get output"
Be careful, the ouput files don't have exon, intron, integenic, 5-UTR, 3-UTR informatics if you save it as a single file. You can save them as separated files so that you know the information for each subset.

需要自己再sort一下：

cat raw.txt/Gene.txt | egrep "^chr([0-9]{1,2})" | grep -v random | bedtools sort -g ../genome.txt > Gene.bed
cat raw.txt/UTR3.txt | egrep "^chr([0-9]{1,2})" | grep -v random | bedtools sort -g ../genome.txt > UTR3.bed
cat raw.txt/UTR5.txt | egrep "^chr([0-9]{1,2})" | grep -v random | bedtools sort -g ../genome.txt > UTR5.bed
cat raw.txt/down2K.txt | egrep "^chr([0-9]{1,2})" | grep -v random | bedtools sort -g ../genome.txt > Down2K.bed
cat raw.txt/CDS.txt | egrep "^chr([0-9]{1,2})" | grep -v random | bedtools sort -g ../genome.txt > CDS.bed
cat raw.txt/exon.txt | egrep "^chr([0-9]{1,2})" | grep -v random | bedtools sort -g ../genome.txt > Exon.bed
cat raw.txt/up2K.txt | egrep "^chr([0-9]{1,2})" | grep -v random | bedtools sort -g ../genome.txt > Up2K.bed
cat raw.txt/intron.txt | egrep "^chr([0-9]{1,2})" | grep -v random | bedtools sort -g ../genome.txt > Intron.bed

这里得到的CDS是有overlap的，因为是按转录本算的，同一个基因有多个转录本。

2. genecode - 最自动化，全部用代码搞定

3. 其他

RSeQC有比较好的bed注释文件，但是不是完全的明文，不好利用。

附录：

intergenic的需要自己计算

wget http://ftp.ensembl.org/pub/release-75/gtf/homo_sapiens/Homo_sapiens.GRCh37.75.gtf.gz

grep -v "_" hg19.chrlen.bed | egrep "^chr([0-9]{1,2})" | cut -f1,3 > hg19.chrlen.1_22.bed

sort -k1,1 hg19.chrlen.1_22.bed > sorted.genome

bedtools complement -i UCSC.anno/Gene.bed -g hg19.chrlen.1_22.bed > intergenic.bed
sort -k1,1 -k2,2n intergenic.bed > sorted.intergenic.bed
bedtools merge -i sorted.intergenic.bed > merged.sorted.intergenic.bed