[Fwd: Re: [Gmod-help] Displaying wiggle tracks]

Wed Sep 16 17:10:31 EDT 2009

Hi Jeff,

On Wed, Sep 16, 2009 at 8:49 AM, Jeff Bowes <bowes at ucalgary.ca> wrote:

> Hi Scott, Dave
>
> About your suggestion to Kevin that we move to GBrowse 2.0.:
>
> 1) Is there a best practice for the user it should be installed under, do
> you usually give GBrowse its own user?,  This was not clear in the install
> Wiki.
>

The default practice is to use the web server user.  I hesitate to call this
a best practice, though.  I think it is just easier to set up.

2)  I'm a little leery about using Beta software in production.  How solid
> is Gbrowse 2.0 ?
>

I haven't played with it extensively since mid-summer.  It still had some
issues then.  However, I would get it, and test it with your data and see
how it runs.

> 3) Who is using it in production now?
>

There are several power users out there who are using it in production,
including CBRG at Oxford and Bayer CropScience (see
http://gmod.org/wiki/August_2009_GMOD_Meeting for more on them), and
modENCODE.

4) When is the expected release date for  gbrowse 2.0
>

I'm not touching that one.  :-)

>
> 5) If we install Gbrowse 2.0 beta what would be the process for installing
> fixes and the updating to the final release version. (Will it be
> pain-free?!)
>

I have not heard anything about major changes coming to GBrowse 2 between
now and release.  I think all expected work is bug fixes.

> Currently we run Gbrowse using the Chado adpator.  At the time we did this
> I only new of us and dictybase that were doing this. Like, dicty to make
> this work we had to rewrite a number of queries to speed it up.  This makes
> it a pain to keep up with upgrades.  Additionally, the methylation data
> Kevin mentioned would require 46 million feature entries- I'm not sure the
> Chado adaptor can handle this.  So I am thinking we should move to having
> Gbrowse have its own data store and batch updates from chado.So my questions
> are:
>

46 million feature entries!  What does the methylation data look like?  If
it can represented in a wiggle_xyplot then this will greatly reduce the
volume of the data.  The CBRG uses wiggle_xyplot very successfully to show a
lot of methylation data.

1) Do you agree that we should move away from the Chado adaptor
>

Yes.

> 2) What is the fastest back end data store choice  for Gbrowse
>

The preferred choice is SeqFeature::Store.  It is slow to load though.  I
think that at browser time, SeqFeature::Store in MySQL (with it's none to
bright optimizer) will clearly beat Chado in DB2 (with it's darn good
optimizer).

3) Is the process for keeping it up to data to just do batch GFF3 exports
> from chado and import them into GBrowse?
>

Yes, but ...

4) For whatever data store you recommend,  is it possible to make
> incremental updates to it or do updates have to pretty much work as blow
> away and replace it.
>

... I don't believe that SeqFeature::Store lends itself to incremental
updates.  I would create a parallel database and switch the config file once
it is done being created.

5) Any other advice about this?
>

No , but, do you have to deal with reference sequence updates?  MAKER has
now added a functionality to move annotations from one version to the next.

Hope all is well, and give my regards to Erik,

Dave C.

Thanks very much,
>
> Jeff
>
>
>
>
>
>
>
>
>
>
> Kevin Snyder wrote:
>
>>
>>
>> ------------------------------------------------------------------------
>>
>> Subject:
>> Re: [Gmod-help] Displaying wiggle tracks
>> From:
>> Scott Cain <scott at scottcain.net>
>> Date:
>> Thu, 10 Sep 2009 13:02:45 -0400
>> To:
>> scott at scottcain.net
>>
>> To:
>> scott at scottcain.net
>> CC:
>> Kevin Snyder <ksnyder at ucalgary.ca>, help at gmod.org
>>
>>
>>
>> Hi Kevin,
>>
>> Since you're likely doing an upgrade anyway, you might want to look at
>> GBrowse2 (not officially released yet, but available on CPAN), as it will
>> have many performance boosts over what you are currently using.
>>
>> Scott
>>
>> On Sep 10, 2009, at 12:31 PM, Scott Cain wrote:
>>
>>  Hi Kevin,
>>>
>>> You won't be able to display wiggle plots without a never version of
>>> Bio::Graphics (which is no longer part of BioPerl) and a newer version of
>>> BioPerl (1.6 or better).
>>>
>>> Scott
>>>
>>> On Sep 10, 2009, at 12:22 PM, Kevin Snyder wrote:
>>>
>>>  Hello,
>>>>
>>>> We are using a rather old version of gbrowse at the moment (1.64) along
>>>> with Perl 5.8.8 and BioPerl 1.5.1.  We have been give a large set of data in
>>>> wiggle BED format.  Will this version properly display the  data in wiggle
>>>> format
>>>> or do we need to upgrade?
>>>>
>>>> Kevin Snyder
>>>> Xenbase
>>>>
>>>
>>> -----------------------------------------------------------------------
>>> Scott Cain, Ph. D. scott at scottcain dot net
>>> GMOD Coordinator (http://gmod.org/) 216-392-3087
>>> Ontario Institute for Cancer Research
>>>
>>>
>>>
>>>
>>>
>> -----------------------------------------------------------------------
>> Scott Cain, Ph. D. scott at scottcain dot net
>> GMOD Coordinator (http://gmod.org/) 216-392-3087
>> Ontario Institute for Cancer Research
>>
>>
>>
>>
>>
>>
> --
> Jeff Bowes M.Sc.
> Technical Director, Xenbase
> Department of Biological Sciences
> University of Calgary
> Calgary, Alberta T2N 1N4
> CANADA
> Tel: (403) 220-2824 Fax: (403) 284-4707
>
>
>
>

-- 
GMOD News: http://gmod.org/wiki/GMOD_News
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://brie4.cshl.edu/pipermail/gmod-help/attachments/20090916/6ffbbe65/attachment.html>