The metabolic database was assembled to annotate metabolic pathways inferred with the Metabolic Pathfinder and Pathway extraction tools in NeAT. It unites data from KEGG (version 55), MetaCyc (version 14.1) and aMAZE, and contains in total 2,012 reference pathways.

The data model is described in this doctoral thesis (p. 191).

The metabolic database is provided as a postgres backup file and can be downloaded here.

To load the metabolic database into postgres, start postgres on command line:

  >psql -U postgres

In postgres, create a database user (here named 'metabolic') and the database:

  create user metabolic with password 'metabolic';
  create database "metabolicdb" with owner "metabolic" encoding='TF8';

Quit postgres and restore the database from the backup file on command line:

  >pg_restore -d metabolicdb metabolicdb_dump_03_10_2010.backup -U postgres

You can then enter the metabolic database with the following command line:

  >psql -U metabolic -d metabolicdb

To display some information on the tables and the data content, type:

  metabolicdb=> \dt
  metabolicdb=> \d databases;
  metabolicdb=> select * from databases; 

In the following, a few commands are given to get some statistics on the data content. The commands below provide the number of reactions (15,802), RPAIRs (12,287), compounds (13,195) and genes (3,981).

  metabolicdb=> select distinct transformation_type from transformations;
  metabolicdb=> select distinct bioentity_type from bioentities;
  metabolicdb=> select count(*) from transformations where transformation_type='REACTION';
  metabolicdb=> select count(*) from transformations where transformation_type='SUBREACTION';
  metabolicdb=> select count(*) from bioentities where bioentity_type='COMPOUND';
  metabolicdb=> select count(*) from bioentities where bioentity_type='GENE';

To obtain the number of pathways from KEGG (144), MetaCyc (1752) and aMAZE (116), the following commands can be used:

  metabolicdb=> select count(*) from pathway_versus_database where database_unid=3;
  metabolicdb=> select count(*) from pathways where name like '%map%';
  metabolicdb=> select count(*) from pathway_versus_database where database_unid=4;
  metabolicdb=> select count(*) from pathway_versus_database where database_unid=6;