Talk:Chado New Users

Meeting Notes

20070405

• Tried to work through some of the Central Dogma stuff in Chado Best Practices. Zheng ran the sample query, which hung his laptop.
• Jim created the Problems template Template:Problems for adding notes to documentation pages in the wiki.

20070329

Jim was out of town and everyone else attended the Systems Biology meeting

20070322

Lincoln was able to attend today. We looked at Chado tables using a combination of phpPgAdmin and command line. To do direct SQL queries in phpPgAdmin, click at the schema level.

Example query:

select f.name, d.*, fd.* from dbxref as  d, feature_dbxref as fd, feature as f where d.dbxref_id = fd.dbxref_id and f.feature_id = fd.feature_id and accession !='SGD'

Gives:

name  | dbxref_id | db_id | accession | version | description | feature_dbxref_id | feature_id | dbxref_id | is_current 
-------+-----------+-------+-----------+---------+-------------+-------------------+------------+-----------+------------
COX2  |     42435 |     2 | landmark  | 1       |             |             16241 |      16241 |     42435 | t
COB   |     42435 |     2 | landmark  | 1       |             |             16176 |      16176 |     42435 | t
COX1  |     42435 |     2 | landmark  | 1       |             |             16129 |      16129 |     42435 | t
RPC40 |     42435 |     2 | landmark  | 1       |             |             15800 |      15800 |     42435 | t
HTS1  |     42435 |     2 | landmark  | 1       |             |             15623 |      15623 |     42435 | t
ERG10 |     42435 |     2 | landmark  | 1       |             |             15474 |      15474 |     42435 | t
PEP4  |     42435 |     2 | landmark  | 1       |             |             15173 |      15173 |     42435 | t
TPK2  |     42435 |     2 | landmark  | 1       |             |             15058 |      15058 |     42435 | t
GAL4  |     42435 |     2 | landmark  | 1       |             |             14960 |      14960 |     42435 | t
RAD17 |     42435 |     2 | landmark  | 1       |             |             14775 |      14775 |     42435 | t

Note that all have the same dbxref_id.

Conclusions

• bulk loaded sample yeast gff3 may not create canonical chado data.
  • features all have dbxref = NULL
    • feature.dbxref_id is nullable; the original design was to have the dbxref_id that appears in the feature table to be the 'primary' dbxref_id. A fairly recent addition was to add a command line flag to the bulk loader to allow the user to specify what (if anything) should be used as feature.dbxref_id. All others go in feature_dbxref.
  • feature_dbxref acts strange (see example above)
    • Not strange at all :-) In typical installations, db_id=2 is the GFF_source that comes from the GFF3 file. If you take a look at the GFF3 that SGD produces, most have a source of 'SGD', but a few have a source of 'landmark', which is used by SGD to put a few gene features in the overview in a GBrowse display.
      • OK, it was a little strange--the GFF bulk loader (and the underlying bioperl module, Bio::FeatureIO::gff) didn't recognize attribute tags of 'dbxref' as dbxref tags (the standard is to use 'Dbxref'. I've modified bioperl and the bulk loader to allow the lowercase, non-standard dbxref now.
  • other?
• should get XML from a known Chado db and import via XORT or other.
  • talk to Pei Li and Eric Just