Hahne, Florian | 2 May 09:51 2013

Re: Gviz: DataTrack heatmap

Hi Fabian,
thanks for your feedback. Please see below for comments.
--

From: Fabian Müller <fmueller@...<mailto:fmueller@...de>>
Date: Tuesday, April 30, 2013 11:21 AM
To: NIBR <florian.hahne@...<mailto:florian.hahne@...>>
Subject: Re: Gviz: DataTrack heatmap

Hi Florian,
I saw the new version of the Gviz package and was very please that the sample names and color scales for
heatmaps are now incorprated. Thanks a lot.
In the meantime I ran across a few issues:
- In R 3.0 and Gviz 1.4.1, If I specify a custom color scale for the heatmaps, Gviz still depicts the default
color gradient in the legend. The colors within the heatmap itself are fine. Can you deactivate the
gradient legend somehow?

This clearly is a bug. When sample groups are set in the heat map, the gradient colors are being picked up from
the 'col' display parameter, and I forgot to evaluate the 'gradient' parameter in the single sample case.
As a quick fix, you can set dummy groups in your heat map and use the 'col' parameter, even though this does
not give you as much flexibility for the colour scale as the gradient one does:
plotTracks(dtTrack, type="heatmap", col="red",  groups=c(1,1,1,1))
I will fix this asap and also include a flag to turn off the colour legend completely.

- In R 2.15 and Gviz 1.2.1 I ran accorss a problem when trying to create Ideagrams. I received the error in the
attachment of this email. I saw that a similar issue has been fixed in Gviz 1.4.1 and R 3.0
(http://permalink.gmane.org/gmane.science.biology.informatics.conductor/47780). Could you
also fix this for R 2.15 and Gviz 1.2.1 (which we are running on a couple of machines due to compatibility issues)?

I will take a look at this and get back to you once a fixed version is available.

- In R 3.0 and Gviz 1.4.1,  I tried to visualize CpG Islands using UCSC tracks and received the error in the
second attachment. It used to work with Gviz under R 2.15. Did something change here?

This seems to be a bug in the tracklayer package which has already been reported. I am hoping for a quick fix there.

Do you have any suggestions for potential solutions. Thanks!

Best regards,
Fabian

__________________________________________________
    Fabian Müller
    Max Planck Institute for Informatics
    Computational Biology and Applied Algorithmics Department

    Campus E 1.4
    Room 509
    66123 Saarbrücken
    Germany

    Phone: +49 (681) 9325 3009
    Email:  fmueller@...<mailto:fmueller@...>
    http://www.mpi-inf.mpg.de/~fmueller/
On 7 Mar, 2013, at 16:16, "Hahne, Florian"
<florian.hahne@...<mailto:florian.hahne@...>> wrote:

Hi Fabian,
thanks for the suggestion.
This has actually come up a number of times already and I have added it to the list of developments for the next
release. If everything works out as expected you should be able to do this in a couple of weeks when the new
Bioconductor version is available. I'll let you know whether or not this feature makes it into the release
or not. Other additions to the heat maps that should be coming up are optional axis indicators that map the
colour range to numerical values.
Cheers,
Florian
--

From: Fabian Müller <fmueller@...<mailto:fmueller@...de>>
Date: Thursday, March 7, 2013 10:10 AM
To: NIBR <florian.hahne@...<mailto:florian.hahne@...>>
Subject: Gviz: DataTrack heatmap

Hi Florian,
I ran accross your Gviz R-package recently. First of all, thank you! It's a great package.
I have one question/request:
I am using the DataTrack type with heatmap visualization. I create the track from a GRanges object with
multiple columns in elementMetadata. The visualization with Gviz works nicely. However, I would like to
include the column names of elementMetadata as rownames in the visualized heatmap. Is this somehow possible?

Best regards,
Fabian

__________________________________________________
    Fabian Müller
    Max Planck Institute for Informatics
    Computational Biology and Applied Algorithmics Department

    Campus E 1.4
    Room 509
    66123 Saarbrücken
    Germany

    Phone: +49 (681) 9325 3009
    Email:  fmueller@...<mailto:fmueller@...>
    http://www.mpi-inf.mpg.de/~fmueller/

	[[alternative HTML version deleted]]

_______________________________________________
Bioconductor mailing list
Bioconductor@...
https://stat.ethz.ch/mailman/listinfo/bioconductor
Search the archives: http://news.gmane.org/gmane.science.biology.informatics.conductor

Gmane