Scott Cain's website

Bioinformatics, Perl, Cool Stuff
Powered by Drupal, an open source content management system


The past few days I've done lots of work on creating a template for a private chado schema.

  • Had the private primary key table use the public key's sequence so their would be no problem moving the data from private to public
  • Created indexes that mirror all of the normal indexes in the public schema
  • Added unique, foreign key and not null constraints that mirror the public schema as well
  • Added default values that mirror the public schema; there is currently a problem with this that, when the default value is zero and the field is the last in the list, it doesn't get created properly. I'm continuing to work on that.

Work done yesterday and today: 9 hours.

Worked on the template to create a private schema. The template now successfully creates ddl to:

  • Create tables and views for all the tables in the public chado schema
  • Create insert and update rules so that inserts and updates into the views will result in inserts and updates to the corresponding table.

What does not happen yet is deletes. There are a few problem with deletes:

  • While it is quite straight forward to write general rules for update and inserts, it is not so easy to do so for deletes. I'm not sure if it is possible (looking into it).
  • The tables created for the private schema do not have the constraints that the public schema tables do, which means there is no "ON DELETE CASCADE" in the private schema, which the GFF loaders delete function depends on to work.

As a result, deletes might be problematic. For the time being, I suggest we just don't delete.

2.5 hours.

The template isn't done yet, but it is working.

1.5 hours

Did some preliminary testing with multiple private schemas, showing that the approach has promise and has been used by others for similar problems.

1 hour

Several meetings through the end of the week:

  • LBNL, 2/25: 7 hours
  • UC Berkeley, 2/26: 3 hours
  • JGI, 2/26: 3 hours
  • Stanford/TAIR: 3 hours

Billable hours:

Last week: 1 hour, phone calls and trip planning
Yesterday (2/24): meeting at HHMI: 7 hours

This is the first post for the iPlant project I'm working on. Any work done for iPlant will be recorded here with an 'iPlant' tag.

41 hours

Spent three hours this morning writing user documentation for all of the code I've written. (3 hours)

Over the last week or so, I've done a lot more testing of the featureloc update. Only 5 features failed outright due to an insert that grew features from about 1.5kb to about 5kb. Another 10 or so features failed because one or both of their coordinates failed to map to the new genome assembly. (4 hours)