[Genome] Question about extending genome graph (/cgi-bin/hgGenome) parameters

[Ross Lazarus]

Genome Graphs is a great way to see multiple custom tracks. I'm using it on an
internal private mirror here at the Channing lab to display hundreds of outputs
from analyses of large association studies for the investigators.

I have a proposal that would make it even greater with a relatively small
increase in complexity. I have a working Galaxy genome graphs datatype that
works exactly like the current Galaxy bed file viewer - it has a "view at ucsc"
link that opens a (in our case!) local mirror genome graphs display. Users
really like it, but they complain about 3 things - the lines joining points are
misleading for these statistics; they don't like the fact that the default is
all tracks on at 30 pixels or so; and they want the tracks scaled sensibly.

I was wondering if we could have some additional parameters - eg 1) a flag to
substitute bars when you don't want the apparent smoothing of joined dots, and
2) some ways to hint at each track's vertical size 3) and scale, 4) and default
state (full, packed etc)? I am constructing the URL in python code and it's easy
for me to comply with extensions to the current genome graphs call to
/cgi-bin/hgGenome

I quickly understood from reading the code that making these changes correctly
would be a non trivial, high risk venture for me. Does anyone with deep
knowledge have time for this please? I would be interested in discussing the
possibility of supporting this work via a small (!) subcontract from one of my
grants if that is a realistic option and would help get this done ?


> From ann at soe.ucsc.edu Mon Feb 5 05:42:17 2007
> From: ann at soe.ucsc.edu (Ann Zweig)
> Date: Mon, 05 Feb 2007 07:42:17 -0600
> Subject: [Genome-announce] New Genome Graphs Tool Available in Genome Browser
> Message-ID: <45C73439...@soe.ucsc.edu>
>
> We are pleased to announce the release of a new software tool in the
> Genome Browser collection, the Genome Graphs tool. Genome Graphs offers
> the ability to upload and display genome-wide data sets such as the
> results of genome-wide SNP association studies, linkage studies and
> homozygosity mapping. The Genome Graphs tool may be accessed from the
> menu on the UCSC Genome Bioinformatics home page, or from this link:
> http://genome.ucsc.edu/cgi-bin/hgGenome
>
> The initial release of Genome Graphs includes the following features:
>
> - upload several sets of genome-wide data and display them simultaneously
> - click on an area of interest and go directly to the genome browser at
> that position
> - set a significance threshold for your data and view only regions that
> meet that threshold
> - view the genes that exist in areas where your data meet your
> significance threshold
>
> For more information about the Genome Graphs tool, visit the Gateway
> page or consult the Getting Started on Genome Graphs section in the
> User's Guide:
> http://genome.ucsc.edu/goldenPath/help/hgTracksHelp.html#GenomeGraphs
>
> Genome Graphs was written by Jim Kent of the UCSC Genome Bioinformatics
> Group.
>
> -------
> Regards,
> Ann Zweig



--
Ross Lazarus MBBS MPH, Director of Bioinformatics
Channing Laboratory, 181 Longwood Ave., Boston MA 02115, USA.
Tel: +617 525 2730 Fax: +617 525 0958


The information transmitted in this electronic communication is intended only for the person or entity to whom it is addressed and may contain confidential and/or privileged material. Any review, retransmission, dissemination or other use of or taking of any action in reliance upon this information by persons or entities other than the intended recipient is prohibited. If you received this information in error, please contact the Compliance HelpLine at 800-856-1983 and properly dispose of this information.

[Robert Kuhn]

Hello, Ross,

We're glad you like the Genome Graphs tool. We will take your segguestions
to heart and see if we can implement them.

Bet wishes,

--b0b kuhn
ucsc genome bioinformatics group


> From genome-...@soe.ucsc.edu Mon Oct 8 02:17:48 2007
> To: gen...@soe.ucsc.edu
> Cc: ross.l...@channing.harvard.edu
> Subject: [Genome] Question about extending genome graph (/cgi-bin/hgGenome)
> parameters

> _______________________________________________
> Genome maillist - Gen...@soe.ucsc.edu
> http://www.soe.ucsc.edu/mailman/listinfo/genome
>

[Ross Lazarus]

Galt, thanks for asking - it helps me to clarify what I have in mind.

I basically want to make genome graphs more flexible and controllable - more
like custom tracks in their options. I'd like to be able to specify these new
options as parameters to /cgi-bin/hgGenome. I'm constructing URL's using a
python script, so it's easy for me to make really long, ugly ones!

Attached is an image showing 3 tracks - the 2 top ones are from loading a gg file using an
URL my code created -
http://hg/cgi-bin/hgGenome?db=hg18&hgGenome_dataSetName=hmYRI_CEU_tdt_TDTgg&hgGenome_dataSetDescription=GalaxyGG_data&hgGenome_formatType=best%20guess&hgGenome_markerType=best%20guess&hgGenome_columnLabels=first%20row&hgGenome_maxVal=&hgGenome_labelVals=&hgGenome_maxGapToFill=25000000&hgGenome_uploadFile=http%3A%2F%2Fgalaxy%2Fdisplay_as%3Fid%3D351%26display_app%3Ducsc&hgGenome_doSubmitUpload=submit

The lower track is a custom track on our local UCSC mirror site - a bar graph for similar
p value data.

I want to be able to create a GG call URL with additional parameters so the GG tracks end
up looking more like the custom track - bar graph, 90pixels,...etc

The specific options I'd like to control are:

1) a track height pixels for each column - eg as a comma separated value list of
values

2) a default display (packed, full..) as above, one per column

3) a scale hint for each column

4) an optional colour for each column

5) graph type for each column (points, lines, bars,...)


The genome graphs upload form would also need to be adjusted to reflect these
for interactive users. I know this is substantial, particularly since you have
to already know how to correctly amend the input form and input parameter
parsing, and how to correctly adjust track drawing - none of which seem easy to me!

My primary motivation is getting bar graphs, but I'm guessing that there are
plenty of other representations that users might want to have available. A lot
of the work we do involves contemplating test statistics from half a million or
a million markers on thousands of phenotypes, so alternative ways of summarizing
with something like the -log10 pvalue coarse heatmap karyotype at
http://www.ncbi.nlm.nih.gov/SNP/GaP.cgi?rm=genomeTraitView&test_id=13 would be
useful...



Galt Barber wrote:
> Hi, Ross.
> I work on hgGenome sometimes.
> Can you provide more detailed descriptions,
> or even some visuals of what you have/what you want to have,
> so I'm sure I understand it.
>
> No promises here, just wanting to communicate clearly
> about what features you had in mind.
>
> thanks!
>
> -Galt

>
>
> On Fri, 5 Oct 2007, Ross Lazarus wrote:
>
>> Genome Graphs is a great way to see multiple custom tracks. I'm using it on an
>> internal private mirror here at the Channing lab to display hundreds of outputs
>> from analyses of large association studies for the investigators.
>>
>> I have a proposal that would make it even greater with a relatively small
>> increase in complexity. I have a working Galaxy genome graphs datatype that
>> works exactly like the current Galaxy bed file viewer - it has a "view at ucsc"
>> link that opens a (in our case!) local mirror genome graphs display. Users
>> really like it, but they complain about 3 things - the lines joining points are
>> misleading for these statistics; they don't like the fact that the default is
>> all tracks on at 30 pixels or so; and they want the tracks scaled sensibly.
>>
>> I was wondering if we could have some additional parameters - eg 1) a flag to
>> substitute bars when you don't want the apparent smoothing of joined dots, and
>> 2) some ways to hint at each track's vertical size 3) and scale, 4) and default
>> state (full, packed etc)? I am constructing the URL in python code and it's easy
>> for me to comply with extensions to the current genome graphs call to
>> /cgi-bin/hgGenome
>>