I have some bam files with chromosome names like "1" instead of "chr1". I run into problems when I want to plot data from ensembl along with coverage data from these bam files. The solution this far has been to use the command below to edit the bamfile before using it in my scripts:

samtools view -H test.bam | sed -e 's/SN:\([0-9XY]\)/SN:chr\1/' -e 's/SN:MT/SN:chrM/' | samtools reheader - test.bam > test_chr.bam

I don't feel good doing that, so I would like to modify the function that reads data from the bam file somehow so that it returns the chromosome name prefixed with "chr". Is that possible? I tried editing the default import function, only editing the line giving seqnames, like this:

seqnames = if (is.null(reads$rname)) character() else paste("chr", reads$rname, sep="")

But that didn't lead me anywhere. Does anyone have suggestions?

I think you are on the right track. The .import.bam.alignments function has a selection parameter (Granges object) which is constructed from the plotTracks arguments and limits the reads which are read from the BAM file. I assume your other tracks are using the "chr" in sequence names and therefore this selection object will point to "wrong" chromosome. Simple fix for that would be to add one additional line into the modified .import.bam.alignments function. Add this line after the  BamFile command (before creating the ScanBamParam object). And keep the modification you made before.

seqlevels(selection) <- sub("chr", "", seqlevels(selection))

Just another suggestion - if you have Ensembl based data, why not use ensembldb? You could use the dedicated getGeneRegionTrackForGviz to extract GRanges to pass to Gviz GeneRegionTrack method. Don't forget then to call options(ucscChromosomeNames = FALSE) too so that Gviz is not crying about the chromosome names not starting with "chr". Check the ensembldb vignette for more details and description.

Also, using EnsDb databases, since they match certain Ensembl releases, would allow you to ensure that the annotation you used for alignment or counting matches the annotation to plot the data.

cheers, jo