Hello,

I am trying to use GViz for plotting genes annotations from GTF/GFF files. 

It works nicely with files downloaded from UCSC table browser (refflat, refseq, gencode). However I could not find a way to download a file that provides simultaneous genes and transcripts information. All the lines share the same 'gene_id' and 'transcript_id'.

When I download annotations directly from gencode website, it seems that these file carry more information (genes/transcripts/symbols...) : http://www.gencodegenes.org/releases/current.html

However, while the latter ones are correctly loaded and parsed by the constructor 'GeneRegionTrack', the plotting step does not resolve the internal gene structure (just a big box, no exons or thin boxes for non coding...). I tried playin with the thinBoxFeature in vain.

Has anybody experience the same ?

Using the gencode files would be very helpful because it allows to use ensembl IDs or genes symbol directly for selecting genes/transcripts. My question probably reflects my misunderstanding of UCSC/Gencode or GFF/GTF file formats. Any help would be appreciated :)

On a side question, is GViz package still maintained ? Or should I ask my questions about it on a more adapted forum/list ?

Thank you for your help.

Romain.

I found this hack how to get the correct plot. I am no so sure why the direct import does not work so well but I guess the format of GFF3 from GenCode and one from UCSC is not exactly the same.

library(rtracklayer)
library(GenomicFeatures)
library(Gviz)
# using the rtracklayer import function I can read the GFF3 from GenCode and create conversion table between gene id and gene symbol
gff3 <- import.gff3("gencode.v24.annotation.gff3.gz")
gene2symbol <- mcols(gff3)[,c("gene_id","gene_name")]
gene2symbol <- unique(gene2symbol)
rownames(gene2symbol) <- gene2symbol$gene_id

# the GFF3 from GenCode can be parsed into TxDb object and after creating the GeneRegionTrack I could set the correct gene symbols using the table from previous step
txdb <- makeTxDbFromGFF("gencode.v24.annotation.gff3.gz")
geneTrack <- GeneRegionTrack(txdb, chromosome="chr7", from=5527151, to=5563784)
ranges(geneTrack)$symbol <- gene2symbol[ranges(geneTrack)$gene, "gene_name"]

# plotting with gene symbols
plotTracks(geneTrack, chromosome="chr7", from=5527151, to=5563784, showId=TRUE)

Hello,

yes there is indeed some minor differences between both formats. Thank you very much for the workaround, that is a great idea !

Romain.

Hello,

yes there is indeed some minor differences between both formats. Thank you very much for the workaround, that is a great idea !

Romain.