A Question On Grouping Into Rna,Trna,Rrna,Sntna,Scrna,Snorna,Lincs.Etc

Question

[Hamilton]

Hi,

From ~28000 mm9 genes with refseq,ensembl,ucsc known id, gene symbol, and description for RNA-seq data, I am trying to get information of percentage for each group, RNA,tRNA,rRNA,snRNA, lincs. Since there are gene symbol and description for each gene, therefore, if i take a look at genes, roughly i am able to get where particular gene belongs to. But i would like to get an exact percentage how many genes belong to tRNA, how many genes to rRNA/snRNA/...etc

how can i get such information?

[Larry_Parnell]

What is mm9? Is this a build of the mouse genome?

[Hamilton]

the latest mouse genome

Answer 1

[Malachi Griffith]

If you have Ensembl gene ids you can get the 'gene_biotype' for each. There are many ways to access this information from Ensembl. For example:

This GTF file contains the biotypes for 37,681 genes: ftp://ftp.ensembl.org/pub/release-64/gtf/mus_musculus/Mus_musculus.NCBIM37.64.gtf.gz

NCBI Mouse Build 37 is 'mm9' according to UCSC: http://genome.ucsc.edu/FAQ/FAQreleases.html

Each gene is classified by one of the following in this Ensembl mouse version.

IG_C_gene
IG_D_gene
IG_J_gene
IG_V_gene
Mt_rRNA
Mt_tRNA
lincRNA
miRNA
misc_RNA
polymorphic_pseudogene
processed_transcript
protein_coding
pseudogene
rRNA
snRNA
snoRNA

Answer 2

[Eric Fournier]

Get your list of sequences in GenBank format, and look for ncRNA features. They should have an /ncRNA_class attribute which will tell you whether it represents tRNA, a snoRNA, etc.

[Larry_Parnell]

True +1. One need not get the sequences in GenBank format per se, but capture the GenBank features in flatfile or similar format.