Exploring the Unison Schema

For a thorough introduction to psql, you should consult the psql manual (try here). To get you started, here are a few of the most important commands:

Most of the precomputed data in Unison are stored in "Protein Feature" and "Protein Alignment" tables, which are prefixed with "pf" and "pa" respectively. The example below shows a listing of protein feature tables, and inspection of the BIG-PI results table in detail.

unison=> \dt+ unison.pf*
                                               List of relations
 Schema |     Name      | Type  | Owner  |                             Description
--------+---------------+-------+--------+----------------------------------------------------------------------
 unison | pfantigenic   | table | unison | EMBOSS antigenic features
 unison | pfbigpi       | table | unison | BIG-PI GPI predictions; see also bigpi_v
 unison | pfeature      | table | unison | parent table of all features; inherited by "subclass" feature tables
 unison | pfregexp      | table | unison | regular expression features based on pmregexp motifs
 unison | pfsigcleave   | table | unison | protein feature from EMBOSS sigcleave
 unison | pfsignalphmm  | table | unison | HMM output from SignalP
 unison | pfsignalpnn   | table | unison | neural network output from SignalP
 unison | pfpsipred     | table | unison | Psipred predicted secondary structure elements
 unison | pftmdetect    | table | unison | signal and transmembrane predictions
 unison | pftmhmm       | table | unison | TMHMM output
 unison | pftype        | table | unison | enumeration of feature types
(11 rows)


unison=> \d+ unison.pfbigpi
                       Table "unison.pfbigpi" / /
   Column    |     Type     | Modifiers       / /  |                 Description
-------------+--------------+-----------------/ / -+----------------------------------------------
 pfeature_id | integer      | not null default/ /  |
 pseq_id     | integer      | not null        /B/  |
 pftype_id   | integer      | not null default/R/  |
 start       | integer      | not null        /E/  | omega site
 stop        | integer      | not null default/A/  | 0
 confidence  | real         | not null        /K/  | BIG-PI Pv value
 params_id   | integer      | not null        / /  |
 site_no     | smallint     | not null        / /  |
 quality     | character(1) | not null        / /  | A-D are good sites; S,N,I are rejected sites
 score       | real         | not null        / /  |
Indexes:
    "pfbigpi_redundant_feature" UNIQUE, btree (pseq_id, "start", params_id)
    "pfbigpi_gpi_proteins" btree (pseq_id) WHERE quality = 'A'::bpchar OR
                                         quality = 'B'::bpchar OR quality = 'C'::bpchar OR quality = 'D'::bpchar
    "pfbigpi_pseq_id" btree (pseq_id)
Check constraints:
    "big_pi_identifies_only_two_sites" CHECK (site_no = 1 OR site_no = 2)
    "quality_must_be_one_of_abcdsni" CHECK (strpos('ABCDSNI'::text, quality::text) <> 0)