[Gmod-help] Re: GMOD Membership of the GNPAnnot Chado Controller ?

Wed Jan 11 00:27:21 EST 2012

Hi Stephanie,

So sorry that I neglected to answer this sooner.  I've been fairly
swamped since the beginning of the year.  I am very enthusiastic about
the idea of the Chado Controller.  It looks to me like it meets all of
the requirements for inclusion as a GMOD component (laid out here:
http://gmod.org/wiki/GMOD_Membership#Software).  What I'd like to
suggest moving forward is that I'll announce the nomination on the
schema (Chado) mailing list and ask for feedback, and it might be
worthwhile to specifically ask people who have a lot of experience
with Chado and editing tools like Apollo and Artemis.  Two groups that
come to mind are Chris Elsik's group (BeeBase and BovineBase) and
Adrian Tivey's group at Sanger (GeneDB).

Before I make that announcement, it might be good for you to create a
page on the gmod.org wiki for CC, giving background on what it is/does
and indicating that it's been nominated for inclusion as a GMOD
component that I could reference in the announcement.

Does this sound like an agreeable plan?

Thanks,
Scott

2012/1/6 Stéphanie Sidibé--Bocs <stephanie.sidibe-bocs at cirad.fr>:
> Sorry, the supplementary_file_20120103_13_VG_SSB_clean.pdf was not complete
> find here enclosed the complete version.
>
> Le 04/01/12 00:25, Stéphanie Sidibé--Bocs a écrit :
>
>> Dear Scott,
>>
>> In the framework of the GNPAnnot project (2008-2011), we have been
>> developing a component based on Chado PostgreSQL database schema: the
>> Chado Controller (CC). This PgSQL layer was presented at the January
>> 2010 GMOD Meeting:
>> http://gmod.org/w/images/5/54/Jan2010GnpAnnotURGI.pdf
>> Our manuscript is under review (Guignon, Droc et al. Revised in 2011,
>> BIOINF-2011-1444) (here enclosed and confidential).
>>
>> We would like to nominate the CC for a new GMOD software component
>> indeed we think it meets all the requirements.
>>
>> The CC meets common needs in community annotation systems (CAS). It
>> allows three types of genomic feature monitoring:
>> 1) Access restriction to ensure the confidentiality of predefined
>> scaffolds and features to the annotators;
>> 2) Manual annotation Inspector to take advantage of controlled
>> vocabularies and control the work of the curator when submitting
>> annotation modifications through a genome editor and
>> 3) Annotation history to ensure the monitoring of annotation history for
>> any modified feature.
>> It is used on the banana complete genome sequenced with NGS.
>>
>> The need to monitor genomic feature is a common and longstanding need
>> that will outlive any particular CAS. So, the CC is going to be useful
>> over time. Indeed, the amount of sequenced genomes increase
>> exponentially. Thus, we are going to manage more and more genomic
>> features that need to be kept of quality and up to date.
>>
>> The CC is configurable and extensible. Indeed it is made up of three
>> modules that can be independently installed, configured, deactivated
>> (compatibility mode).
>> 1) The access restriction module uses views, triggers, rules and
>> functions. It works actually on the Chado feature table but it can be
>> extended to other tables being inspired by the
>> “install_access_restriction.tmpl” PgSQL script.
>> 2) The annotation Inspector module calls checking PgSQL procedures
>> listed in the table “annotation_inspector_procedures” of the modified
>> Chado database. Each record of the table can be modified to enable or
>> disable a function (e.g. “check_te_structure”). New functions could be
>> added being inspired by already existing functions (e.g.
>> “check_gene_structure”) and knowing how the annotation history module
>> works (read the CC technical documentation here enclosed).
>> 3) The annotation history module is based on the Chado Audit module.
>> History module is installed using the current Chado schema of the CAS
>> and not a static given schema. This made it independent of Chado
>> versions. Configuration of the history detail is made as for the
>> “gbrowse_details” through the modification of the PERL script
>> “gbrowse_history”.
>>
>> The CC uses the same open source license for all users than GBrowse and
>> Chado, i.e. Artistic Licence 2.
>>
>> The CC is obviously interoperable with other GMOD components as it is
>> middleware between a Chado database and client Software (e.g. GBrowse,
>> Artemis, Apollo).
>> For the access restriction module working with GBrowse 1.70, a patch
>> needs to be installed. For GBrowse 2.40, an authentication module is now
>> available but has not been extensively tested yet. For Artemis 13.2.8 a
>> patch needs to be installed. For Apollo 1.11.4, a new data adapter has
>> been added. Concerning administrative tasks, a compatibility mode was
>> developed. For instance, two problems occurred when loading data in the
>> Chado database with the “gmod_bulk_load” utility: (i) the PostgreSQL
>> command "COPY ... FROM ..." does not work with views and the feature
>> view is used by the access restriction module and (ii) the triggers of
>> the annotation Inspector should not be called during data loading. Two
>> solutions were set up: (i) if you do not need to simultaneously do data
>> loading and manual curation, you switch on the compatibility mode
>> (“cc_compatibility.pl” restore the “feature” table instead of the
>> “feature” view of access restriction module and disable the triggers of
>> the annotation Inspector module) or (ii) if you need to simultaneously
>> do data loading and manual curation, you take advantage of the modified
>> version gmod_bulk_load that works on “feature_data” table instead of
>> feature and you disable the triggers of the Inspector (“SELECT
>> set_ai_compatibility(TRUE);”). In both cases, the compatibility mode is
>> switched off after data loading.
>> All GMOD components working with Chado were not tested, but in
>> principle, the CC architecture implied that it could work with all
>> components with only slowing down in the worst case. To avoid slowing
>> down, the GMOD component just has to call the optimization function
>> “init_access” at the beginning of each PostgreSQL session. However, the
>> CC could not be compatible with software intrinsically linked to the
>> Chado schema and which need the “feature” table (i.e. software that use
>> advance functions based on PostgreSQL system tables to obtain
>> information on the Chado table typically “SELECT * FROM pg_class WHERE
>> relname = 'feature' AND relkind = 'r';”).
>>
>> The developers of the Chado Controller (Valentin Guignon, Gaëtan Droc,
>> Mathieu Rouard and Stéphanie Bocs) are willing to make a good faith
>> commitment to support the Chado Controller for at least 2 years after
>> joining GMOD (setting up email lists, responding to user questions,
>> writing and maintaining documentation, and adding new features and bug
>> fixes to the code).
>> We are also open to new developers from the GMOD community contributing
>> code and other support to the new component.
>>
>> We have created the gnpannot-users at cirad.fr mailing list that we could
>> publicly archived if the CC become a part of the GMOD suite.
>>
>> We have created a public SVN repository
>>
>> http://gmod:@svn-southgreen.cirad.fr/svn/chadocontroller/branches/installer_package/
>> (login gmod no password) and documentation are available the GNPAnnot
>> Web site http://www.gnpannot.org/content/chado-controller-doc
>> We can also migrate our code on the GMOD repository if needed.
>>
>> Please could you evaluate the Chado Controller? If you decide that it is
>> suitable to become officially part of GMOD and you are interested in
>> integrating Chado schema modifications for the CC you can ask
>> suggestions directly to Valentin Guignon.
>>
>> Guignon, V., G. Droc, et al. (Revised in 2011, BIOINF-2011-1444). "Chado
>> Controller: advanced annotation management with a community annotation
>> system." Bioinformatics.
>>
>> Happy new year.
>>
>
>
> --
> UMR AGAP - Cirad TA A-96 / 03 (B‰t. 3, Bur. 12)
> Av. Agropolis 34398 Montpellier Cedex 5
> Phone +33 4 67 61 56 31 Fax +33 4 67 61 56 05 Cell phone +33 6 84 97 85 30
> skype: sibocs
> http://southgreen.cirad.fr/

-- 
------------------------------------------------------------------------
Scott Cain, Ph. D.                                   scott at scottcain dot net
GMOD Coordinator (http://gmod.org/)                     216-392-3087
Ontario Institute for Cancer Research