Tripal Developer's Meeting 2015-03-03

VCF “non” loader.

• How does this compare to how we would store this data anyway. Can an indexed

Associating files with data.

• Use Cases:

  • BAM files for RNA-Seq / integrated with Expression Modules

  • VCF

  • Whole genomes: FASTA, GFF, Excel.

Multi-Chado and Chado Controller for Tripal: some updates (about the projects) for the community (by VG).

• View initial design: https://drive.google.com/open?id=1q4rWcncmqvDseV3Np6NldoYKFIikXMhNqvXAUQEzbTc&authuser=0

• Chado Controller

  • Sandbox: https://www.drupal.org/sandbox/guignonv/2428743

  • Enforces constraints at the PostgreSQL level (not Drupal)

    • Can then use artemis/apollo or other software that has direct connection to the database (including Tripal).

    • Works by using PLSQL and views/triggers/rules/procedures.

  • Chado history: extension of Chado Audit module.

    • Records a history of edits/updates/inserts for a table in a mirror table

    • differences with the audit module: more audit columns to include user and transaction identifiers

  • Access restrictions: limits read/write access at a granular level.

    • Works on feature table at the moment but will support other tables soon.

    • ‘feature’ table is renamed to ‘feature_data’, and a ‘feature’ view takes the place of the feature table. PostgreSQL user account are not allowed to query ‘feature_data’ table but can query the ‘feature view’.

  • Annotation Inspector

    • Allows for review of annotations made by reviewers.

    • Allows to automate some changes according to the data that has been modified by somebody. (ex. auto-change a color property according to a gene annotation state).

    • Still some redesign to be done and it will be renamed into “Chado Inspector”.

    • Use cases:

      • could be run by an annotator on his favorite genes to check his work.

      • could be run on the whole database by and admin to have statistics and see what’s good and not so good.

      • automate tasks.

      • force some changes regardless what the users asked; for instance the “owner” property of a gene so the annotator who changed the gene can’t use a different name that his user name (ie. sometimes “John Smith”, “JOHN”, “jsmith”, whatever... only his real user name will be used regardless what he entered as a value for the “owner” property).

  • Performance?

    • Audit module has no noticeable impact.

    • Annotation inspector is not optimized, and cause slowness depending on the number of rules and elements on which the rules are checked.

    • Access restrictions, not fully explored in the context of Tripal usage. May have noticeable impact. Impact might only be at the first connection time.

• Available as an “alpha” Drupal Sandbox module.  

  • Currently only for annotation history: https://www.drupal.org/sandbox/guignonv/2428743

• Multi-Chado access (separate from Chado Controller)

  • Sandbox: https://www.drupal.org/sandbox/guignonv/2429515

  • Needs updates to core module to support this.

  • Access restriction will support multiple chado instances.

  • Use Cases:

    • public and a private version of a same chado database in 2 different schema, “chado” and “chado_private”; the Tripal site is public; anonymous people will only see what’s in “chado” schema; users that are logged on will see what’s inside “chado_private”;

    • Different funding sources necessitate separating database.

  • Possible use for access control for web services as long as the access can be limited through PostgreSQL access layer (GRANT/REVOKE) at PostgreSQL user level (ie. “REVOKE ALL ON TABLE feature FROM ROLE some_user;”).

Improvements / Changes to standard modules.

• Instead of a TOC have all data in a single page

• Publication module:  Need viewers for all linking tables.

• Stock images attach to stock pages and link to GRIN.

  • Possible EImage image coming soon

• Need to accommodate genetic maps created from crosses of multiple species (some may be synthetic).  

  • multi-species and synthetic crosses in stock table

  • required organism field in stock table points to a genus organism record

  • individuals in stock table are linked to multiple organisms (perhaps via a new stock_organism linker table)

  • featuremap records are linked to stocks via a custom linker table, featuremap_stock.

  • features (e.g. markers, linkage groups) in the genetic map are assigned to the genus organism record since the organism field must be set in the feature table.

  • need perhaps an accommodation in Chado to support this.

  • Additional use case: interspecific crosses (ie: Lens culinaris by Lens ervoides) cannot currently be stored in chado since you have to pick a single organism for each stock.

Stable 2.0 version still planned for March.

Next meeting Apr. 7: a formal agenda meeting.