Does anybody know of R libraries that extracts information from GTF files? For instance, I have DESeq results with ensembl transcript IDs. I want to annotate each transcript row with the Transcript information from the GTF file. Column 9 of the GTF file is quite complicated by the many fields deliminted by ;. I was wondering if there are tools out there to manipulate GTF files.
The rtracklayer package can import a GTF file, and then you can use GenomicFeatures to make a TxDb object from the GTF, which is probably where you want to go. But if this is a common model organism that is probably not necessary. There are existing TxDb and if you are using Ensembl based genome, EnsDb packages on the AnnotationHub that you could use. But you are being rather mysterious about your organism, so it's hard to say for sure.
Hi pcantalupo,
if you are familiar with tidyverse, you can work with plyranges
It helped me a lot in handling Genomic ranges from different sources.
You can use plyranges::filter in order to keep only the rows of your interest, or try
dplyr::separate(col=yourcolumn, into=c("tag1", "tag2",etc), sep=";")
You can read more about plyranges here
Use of this site constitutes acceptance of our User Agreement and Privacy Policy.
version 2.3.6
Oh, sorry I'm using the latest Gencode v38 GTF file for Human found here: http://ftp.ebi.ac.uk/pub/databases/gencode/Gencode_human/release_38/gencode.v38.annotation.gtf.gz. It has specific attributes that I don't find in the Ensembl GTFs (version 104). For instance, Ensembl GTF does not have information for the canonical transcript but the Gencode v38 version does.
Thank you for the information!
You could do something like
And then all those transcripts would be the canonical ones?
This is great thank you...it would have taken me hours to figure it out. However, it does not give me the info I need. For example the Gencode GTF file may have multiple
tagattributes for each row:For example I'm interested in the gene TMEM14C. Here is an example row in the GTF file:
When I select for TMEM14C using the
zobject fromimportand check thetagvalues, I get this:However, I expect this when extracting all the
tagattributes for TMEM14C on the command line:Any ideas for a workaround? Thank you again,
Could you just grep the gtf for lines containing "Ensembl_canonical"?
Here you go...
It might not be in the gtf, but you can get canonical transcript info from BioMart. It might be easier to query biomart than to parse the gtf.
Correct, I already tried that, thank you. But I'm interested in all the attributes in the GTF file. I was hoping for a simple way in Bioconductor to provide that in a neat little dataframe so I could join it to various differentially expressed transcript tables from DESeq2. However, it does not handle lines with multiple
tagattributes and only parses the last one. I submitted this as an Issue to rtracklayer.You can always load it as a data frame and manipulate it to GRanges after :)