Hi,

I posted a question regarding using Gviz's GeneRegionTrack on http://stackoverflow.com/questions/41421785/customizing-gvizs-plottracks-function but since Gviz is not a keyword I'm posting this here as well.

Hope this is acceptable.

I think that there are a bunch of things in this worth commenting:

First of all, Gviz supports GTF files straight out of the box. Rather than importing those and trying to manipulate the ranges, you could simply provide the path to the file. Indeed there is a bunch of additional stuff in that annotation file, and you may want to clean this up prior to plotting:

options(ucscChromosomeNames=FALSE)
gm <- GeneRegionTrack("~/Downloads/gencode.v25.chr_patch_hapl_scaff.annotation.gtf", chromosome="chr12")
gm <- gm[gene(gm) == "ENSG00000161791.13" & feature(gm) %in% c("CDS", "UTR")]

You will also need to tell Gviz which of the regions are now UTRs by providing the thinBoxFeature display parameter:

plotTracks(gm, thinBoxFeature="UTR")

The reason why you do not see all UTR regions in your plot has to do with the limited horizontal space and the somewhat erratic behaviour of collapsing. If you would provide a somewhat wider plotting device, or turn of collapsing altogether you should get the desired result:

plotTracks(gm, thinBoxFeature="UTR", collapse=FALSE)

Hope that help,

Florian

PS: instead of loading your gene models from a GTF file, why don't you make use of the TxDB packages in Bioconductor, or load the gene model data from UCSC (using the UcscTrack class, they also give you the Gencode annotations) or from Ensembl (using the BiomartGeneRegionTrack class).   

bmTrack <- as(BiomartGeneRegionTrack(genome="hg38", symbol="FMNL3"), "GeneRegionTrack")
bmTrack <- bmTrack[symbol(bmTrack) == "FMNL3"]
seqlevels(ranges(bmTrack)) <- "chr12"
chromosome(bmTrack) <- "chr12"
plotTracks(list(GenomeAxisTrack(), gm, bmTrack), thinBoxFeature=c(Gviz:::.THIN_BOX_FEATURES, "UTR"), collapse=FALSE, transcriptAnnotation="symbol", chromosme=12)