[BioC] rtracklayer -- how to hide some default tracks? fine-grained control of track view mode (pack, full, dense, squish...)

[Paul Shannon]

Hi Michael,

Thanks for your reply.  Very helpful.

My first and main problem with learning rtracklayer was the complexity of the first example in the vignette.  It's a great example, but so hefty, and so much dependent upon data reduction unrelated to the browser tracks (limma, for instance), that it obscured for me the simple powerful commands and capabilities you provide.  There also seems to be rather a shortage of examples in the man pages. 

Another place I got confused was in creating a GRange object to set the range of the browserView.  I passed in the same GRange object I had used in my call to GRangesForUCSCGenome.  It took me a while to figure out why I get 21 browser tabs in response!  Then I woke up, doing this instead:

   gr.chr1 <- GRangesForUCSCGenome ('hg18', exon.chrom.names, exon.ranges, exon.strand, name, score)
     # note: range has just one element
   range.to.view <- GRanges  (seqnames='chr1', range = IRanges (min (start (ranges (gr.chr1))), max (end (ranges (gr.chr1)))), span='*')

   session <- browserSession ('UCSC')
   track (session, 'exonHits') <- gr.chr1
   view <- browserView (session, range.to.view, full='exonHits', dense=c ('chainNetPanTro2', 'chainNetMm9'), hide=trackNames (session))

I also scratched my head for a while over the 'seqnames' argument to GRanges.  I naively thought these would be the names of the features I was laying down in the track, only to eventually figure out that chromosome names were needed.
  
Is it possible to control the color of features displayed in a track?  

Does rtracklayer choose a bed tract when no metadata is supplied in the GRanges object?  And a wiggle track if there is numerical metadata?  Does a 'score' column have special meaning?  Can we set the max and min of a wiggle track?  Right now they are calculated from scores in my GRanges object.

I hope all this makes sense.

 - Paul


On Oct 25, 2010, at 10:16 AM, Michael Lawrence wrote:

> 
> 
> On Mon, Oct 25, 2010 at 9:37 AM, Paul Shannon <pshannon at systemsbiology.org> wrote:
> I have climbed what, for me, has been a difficult learning curve with rtracklayer.  Now that I understand it better, I really like it.  It's a very useful package.
> 
> 
> It would be good to hear what you found to be difficult.
>  
> I have a few remaining questions about controlling which tracks I see, and what view mode they are in (dense, squish, pack, or full).
> How do I implicitly or explicitly hide all tracks, except for a select few, when creating a BrowserView?  And how can I subsequently modify the view mode of selected tracks?
> 
> 1) When my UCSC browser session has some previously viewed tracks on by default, these remain visible when I create a new view:
> 
>    view <- browserView (session, range.to.view, full='exonHits', dense=c ('chainNetPanTro2', 'chainNetMm9'))
> 
> I see these three tracks, plus a half-dozen others left over from previous browsing.
> 
> Instead, I wish to tell the browser "hide all tracks except those I explicitly mention', perhaps like this:
> 
>   view <- browserView (session, range.to.view, full='exonHits', dense=c ('chainNetPanTro2', 'chainNetMm9'), hide='everythingElse')
> 
> 
> Everything in "..." is passed to the ucscTrackModes function. This function favors "full" over "hide" (sorry, this was not documented), so you could do something like:
> 
> view <- browserView (session, range.to.view, full='exonHits', dense=c ('chainNetPanTro2', 'chainNetMm9'), hide=trackNames(session))
> 
> Another option would be to use the 'track' parameter, of browserView, which lists the tracks that should be included in the view. This is merged with the mode parameters, with the mode parameters taking precedence, so this would also do the trick, I think:
> 
> view <- browserView (session, range.to.view, track = character(), full='exonHits', dense=c ('chainNetPanTro2', 'chainNetMm9'))
>   
> 2) My second question: how do I communicate the specific view mode when calling the 'trackNames' method?  trackNames will hide everything besides what I explicitly name (a workaround for question 1, above), like this:
> 
>      trackNames (view) <-  c ('exonHits', 'chainNetPanTro2', 'chainNetMm9')
> 
> But this brings it two problems.   I cannot control the mode of these tracks (dense, squish, pack, full).  
> 
> This is the ucscTrackModes function, e.g.:
> 
> ucscTrackModes(view)[c('exonHits', 'chainNetPanTro2', 'chainNetMm9')] <- c("full", "dense", "dense")
> 
>  
> And then this method creates a new tab in my browser every time I call it.  (Chrome on OS X 10.6)  So every time I want a particular view, with particular tracks, I need two calls:
> 
>    view <- browserView (....)
>    trackNames (view) <- (<some track names>)
> 
> and I get two tabs, the first of which I always delete.
> 
> 
> There's really no easy way to avoid this, given our lack of control over the web browser. The current solution is to combine everything into a single call, as with your first question.
>  
> I have pored over the browserView and trackNames man pages, peeked at the source, all to no avail.
> 
> 
> I think the rtracklayer documentation needs to be rewritten. At the time I was not aware of how to effectively document S4 APIs. Following the Biostrings/IRanges style would be better, I think.
>  
> Sorry,
> Michael
> 
> Any advice?
> 
>  - Paul
> 
> ---- SessionInfo (following a recent update.packages ())
> 
> 
> R version 2.12.0 (2010-10-15)
> Platform: x86_64-apple-darwin9.8.0/x86_64 (64-bit)
> 
> locale:
> [1] en_US.utf-8/en_US.utf-8/C/C/en_US.utf-8/en_US.utf-8
> 
> attached base packages:
> [1] stats     graphics  grDevices utils     datasets  methods   base
> 
> other attached packages:
>  [1] GenomicFeatures_1.1.11             rtracklayer_1.10.0                 RCurl_1.4-3                        bitops_1.0-4.1
>  [5] BSgenome.Hsapiens.UCSC.hg18_1.3.16 BSgenome.Mmusculus.UCSC.mm9_1.3.16 BSgenome_1.17.5                    Biostrings_2.18.0
>  [9] GenomicRanges_1.2.0                IRanges_1.8.0                      org.Pt.eg.db_2.4.1                 org.Hs.eg.db_2.4.6
> [13] org.Mm.eg.db_2.4.1                 RSQLite_0.9-2                      DBI_0.2-5                          AnnotationDbi_1.11.8
> [17] Biobase_2.10.0                     RUnit_0.4.26
> 
> loaded via a namespace (and not attached):
> [1] biomaRt_2.6.0 tools_2.12.0  XML_3.2-0
> _______________________________________________
> Bioconductor mailing list
> Bioconductor at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/bioconductor
> Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor
>