WP2 Updates

Work Package 2 (WP2) of the I-DCC proposal is for the construction of a central repository for the storage of information on gene targeting constructs.  This is a quick update on our progress towards these goals and to ask for feedback on the system that we are implementing.

The Database Schema

This is the proposed format of the database schema.

As you can see it is extremely simple (and nowhere near a fully normalized database form), as it is meant to only record top-level construct information for reporting on, and not be a full product tracking system.

• The genes table is intended to record basic information on genes.
• The products table will house one row of data per ES cell produced, details of the design and vectors for this cell are also recorded in the same row.
• The genbank_files table will be used to store the genbank files related to each of the constructs recorded in the products table.  The rationale for this information to be in a separate table is that the genbank files are large and can be shared between constructs that would produce the same file.

Interacting With the Repository

As stated in the proposal, the main forms of interaction with this repository will be as follows...

For inputting data:

• Web Services - This will be the preferred method of loading data into the schema.  We are currently in the process of testing and validating a Ruby on Rails based application over this model that will supply a strict RESTful web services layer using XML (and in future builds, JSON) as the data transfer format.  Authentication will be HTTP based.
• Basic Web Pages - There will also be a very basic web site over the same model allowing basic loading and/or editing of existing records.  This will share the same user accounts and authentication model as the web services.

• Biomart - This will be the preferred way of reading data from the repository.  The simplistic design of the WP2 schema was deliberate - this will make it very easy to move the data from the repository into a Biomart server for reporting.  This shall be done on a daily basis and will supply users with a standardized interface for the data, and will also supply a standard Biomart web services layer over the data.  There will be no other website created for reporting on the WP2 schema.
• Web Services - A side effect of the RESTful architecture, will be that the same RESTful services can be used to view the data programmatically.  This may be available, but will not be supported - if you wish to use web services to interact with the data, use the ones provided through the Biomart.

Source Code

Comments and Discussion

We are currently in the process of putting this all together, and testing our first draft of the application, but we thought we’d post these plans here for comment and discussion.  Please let us know if you have any comments/reservations about the implementation, or have any suggestions for its improvement.