Chado SO Views
Starting with the GMOD 1.1 release, Chado has a set of views to make querying the feature table for specific SO types (eg, gene, transcript, region) easier. These views are a schema called "so"; as a result, to using them in a query, you generally need to identify the schema in the query.
How the views are defined
For each cvterm in the Sequence Ontology, a view is created with that terms name (except when the length of the name would exceed the length of allowable relation names in PostgreSQL). The views form a join between the cvterm table and the feature table such that if a feature is identified as having a feature.type_id of the name of the view, or a type_id that IS_A type child of the name of the view, it is found.
For example, the following view is for all exons in the feature table:
CREATE VIEW exon AS SELECT feature_id AS exon_id, feature.* FROM feature INNER JOIN cvterm ON (feature.type_id = cvterm.cvterm_id) WHERE cvterm.name = 'coding_exon' OR cvterm.name = 'noncoding_exon' OR cvterm.name = 'interior_exon' OR cvterm.name = 'exon_of_single_exon_gene' OR cvterm.name = 'interior_coding_exon' OR cvterm.name = 'five_prime_coding_exon' OR cvterm.name = 'three_prime_coding_exon' OR cvterm.name = 'three_prime_noncoding_exon' OR cvterm.name = 'five_prime_noncoding_exon' OR cvterm.name = 'exon';
Essentially, it is a way of taking advantage of the structure of the ontology without calculating the closure. Of course, a downside of this approach is that it is hard coded for a particular version of the Sequence Ontology, and if it changes, the views will need to be updated as well.