DrugBank MITAB2.6 File Format

From irefindex

Last edited: 2011-08-25

Description

This document describes usage of the PSI-MITAB2.6 format in order to represent the drug-protein and drug-drug interactions provided by DrugBank in a form suitable for consumption by MITAB-aware tools and services such as MITAB parsers and PSICQUIC Web services.

Although MITAB2.6 is used by iRefIndex (as described in the format documentation), certain elements of that extended format are not directly applicable to DrugBank data, whereas other elements (such as a general checksum for an interactor) are applicable but not directly equivalent to the kind of data provided by iRefIndex: although a drug may have an InChIKey - a form of checksum or hash of the chemical structure of the drug - it is not equivalent or directly comparable to the ROG (redundant object group) employed by iRefIndex.

Format Summary

Note that each record describes a drug-drug or drug-protein interaction where the second interactor (uidB) may be either a protein or a small molecule. Where uidB is a protein, it will always have a UniProt accession as an identifier: interactors not having a UniProt reference are excluded from the MITAB output.

Field Description Example Source
uidA A DrugBank identifier drugbank:DB00123 drug/drugbank-id
uidB For proteins: the given UniProt identifiers
For drugs: see uidA
UniProtKB:P18440 <partner>/external-identifiers/external-identifier/identifier (for proteins)
altA DrugBank secondary accessions having the form drugbank:accession (compatible with MI:2002)
InChIKey values of the form inchikey:key
External identifiers for the drug
drugbank:APRD00123
inchikey:MSTNYGQPCMXVAQ-KIYNQFGBSA-N
drug/secondary-accession-numbers/secondary-accession-number
drug/calculated-properties/calculated-property
drug/external-identifiers/external-identifier/identifier
altB For proteins: external identifiers (other than UniProt identifiers) of the form database:identifier
For drugs: see altA
GNC:7645 <partner>/external-identifiers/external-identifier/identifier (for proteins)
aliasA DrugBank synonyms and brand names using drugbank_synonym and drugbank_brand as qualifiers drugbank_synonym:Hirudin variant-1
drugbank_brand:Refludan
drug/synonyms/synonym
drug/brands/brand
aliasB For proteins: synonyms having the form drugbank_synonym:protein name
For drugs: see aliasA
drugbank_synonym:Arylamine N-acetyltransferase 1 <partner>/name (for proteins)
Method Not used
author Not used
pmids PubMed identifiers describing an interaction pubmed:10505536 drug/<partners>/<partner>/references (filtered)
taxA Not used
taxB Taxonomy identifier for protein of the form taxid:identifier taxid:9606 taken from UniProt
interactionType Not used
sourcedb Source database reference having the form ontology-term-code(ontology-term-name) (see MI:2002) MI:2002(drugbank) implicit
interactionIdentifier The DrugBank identifier used by uidA drugbank:DB00123 drug/drugbank-id
confidence Not used
expansion Not used
biological_role_A The action of the drug in the context of its partners antagonist drug/<partners>/<partner>/actions/action
biological_role_B Not used
experimental_role_A Not used
experimental_role_B Not used
interactor_type_A An ontology reference having the form ontology-term-code(ontology-term-name) where appropriate MI:0326(protein) derived from drug/protein-sequences
interactor_type_B See interactor_type_B derived from <partner>/protein-sequence or drug/protein-sequences
xrefs_A Not used
xrefs_B Not used
xrefs_Interaction Not used
Annotations_A Not used
Annotations_B Not used
Annotations_Interaction Not used
Host_organism_taxid Not used
parameters_Interaction Not used
Creation_date The DrugBank creation date 2005-06-13 07:24:05 -0600 drug/@created
Update_date The DrugBank update date 2011-01-04 14:50:20 -0700 drug/@updated
Checksum_A For drugs: the InChIKey prefixed with inchikey: if available
For proteins: the ROG identifier (rogid) prefixed with rogid:
inchikey:PAJMKGZZBBTTOY-YRIDSSQKSA-N
rogid:XgNg624m2wB07gcr/v+a02LvhNM6421
InChIKey references occur in drug/calculated-properties/property
ROG identifiers are derived from drug/protein-sequences
Checksum_B See Checksum_A InChIKey references occur in drug/calculated-properties/property
ROG identifiers are derived from <partner>/protein-sequence or drug/protein-sequences
Checksum_Interaction For combinations of identifiers from Checksum_A and Checksum_B: the RIG identifier (rigid) prefixed with rigid: rigid:mIwyIi4hME210rHllHmsQ5t3n9k derived from Checksum_A and Checksum_B