I would like to color parts of a transcript. Using this code:

transcriptID <- "ENST00000421310"
breakpoint <- 132225374
start <- 132223104
end <- 132228670
chromosome <- "chr6"

mart <- biomaRt::useMart(
  biomart = "ENSEMBL_MART_ENSEMBL",
  dataset = "hsapiens_gene_ensembl",
  host = "dec2013.archive.ensembl.org")

# Create transcript track
biomartTrack <- Gviz::BiomartGeneRegionTrack(
  filters = list(
    ensembl_transcript_id = transcriptID),
  biomart = mart)

# Create axis track
axisTrack <- Gviz::GenomeAxisTrack()

# works as expected
Gviz::plotTracks(
  c(biomartTrack, axisTrack),
  chromosome = chromosome)

Gets me this plot:

I want control of the colors of each exon, so I try this:

interestcolor <- list("excluded"="red", "included"="green")
feature(biomartTrack) <- "excluded"
feature(biomartTrack)[start(biomartTrack) >= start &
                       end(biomartTrack) <= breakpoint] <- "included"
displayPars(biomartTrack) <- interestcolor

feature(biomartTrack)
# [1] "included" "included" "excluded"

# no change in color?
Gviz::plotTracks(
  c(biomartTrack, axisTrack),
  chromosome = chromosome)

But this gets me the same result. Am I doing something wrong?

This post is relevant: Change the features and colors in BiomartGeneRegionTrack

BiomartTracks will extract all their relevant data directly from calls to Biomart, including the feature type. So assigning thinks via the feature() method does not really do much, since data will be refetched at the next draw. If you want to use this, you will have to coerce your BiomartGeneRegionTrack to a regular GeneRegionTrack first:

 biomartTrack <- as(biomartTrack, "GeneRegionTrack")

However please do note that you will loose the ability to extend the ranges of that track in order to load more annotation data from Biomart. 

From the GeneRegionTrack class documentation:

'GeneRegionTrack' objects also know about coding regions and

     non-coding regions (e.g., UTRs) in a transcript, and will indicate

     those by using different shapes (wide boxes for all coding

     regions, thinner boxes for non-coding regions). This is archived

     by setting the 'feature' values of the object for non-coding

     elements to one of the options that are provided in the

     'thinBoxFeature' display parameters. All other elements are

     considered to be coding elements.

And:

'thinBoxFeature=c("utr", "ncRNA", "utr3", "utr5", "miRNA", "lincRNA")':

          Character vector. A listing of feature types that should be

          drawn with thin boxes. Typically those are non-coding

          elements.

So you could set up an additional feature type like "utr3Red", and add that to the thinBoxFeature array. For that you can now also go ahead and provide another color mapping.

Florian

You can overwrite the assignment for colors to features in the BiomartGeneRegionTrack objects. Those are regular DisplayParameters. They are all listed in the Display Parameters section of the class' help page.

Since UTRs are parts of exons this behaviour makes sense to me. There is currently no way to color them differently from the coding remainder of the exon, and I don't think there should be one. In many cases the UTR regions are even dynamically computed from the gene models. In the graphical representation within Gviz, composite coding/non-coding exons are plotted using polygons, so even technically different colors would not be possible.