Compatability with Gencode GTF

[GuyAllard]

Hi,

I am trying to use pyensembl with the GTF for Gencode v19 as I find the pyensembl interface easy to work with, but I need a more highly curated set of transcripts than the ensembl data.

I am using the Gencode v19 GTF downloaded from ftp://ftp.sanger.ac.uk/pub/gencode/Gencode_human/release_19/gencode.v19.annotation.gtf.gz
Pyensembl imports the data without error and creates the sqlite databases.
The problem I am encountering is that certain loci are present twice in the file, due to their location in a genomic region common to both the X and Y chromosomes. The result of this is that the file contains multiple exons with the same exon_id but different coordinates. When trying to retrieve these exons, pyensembl chokes because these exon_ids are not unique.

Is there a way to overcome this?

[iskandr]

Hi @GuyAllard,

Do you know of any references explaining the differences in curation between Ensembl & Gencode? Also, I didn't realize that genes in the pseudoautosomal regions actually have the same IDs. Is that limited to exons or are there overlapping gene and transcript IDs?

[GuyAllard]

Hi,

The main information that I can find is in the Gencode FAQ http://www.gencodegenes.org/faq.html

The gencode version is a merge of the Ensembl data and the Havana manual annotations.

Re the pseudoautosomal regions, the gencode FAQ states "The gene annotation is the same in both files. The only exception is that the genes which are common to the human chromosome X and Y PAR regions can be found twice in the GENCODE GTF, while they are shown only for chromosome X in the Ensembl file."

As an interim solution I could probably identify the duplicated exon ids and give them a made-up unique id before feeding them to pyenseml.

[iskandr]

Closing as stale (2016, GENCODE v19 / Ensembl 75). GENCODE GTF compatibility has been substantially improved since — most recently #356 wiring gene_type/transcript_type aliases through the GTF reader. The specific pseudoautosomal-region duplicate-exon_id case isn't directly covered, but the original pairing is no longer canonical. Please file a fresh issue against a current GENCODE release if this still bites.