PSI-MI TAB 2.8 Format

Introduction

The MITAB28 format is an extention of the PSI-MI 2.5 (1), 2.6 and 2.7 standards (2) and is suited to capture causal interactions among biological entities. It was initially derived from the tabular format provided by BioGrid and the newest columns are derived from Signor's CausalTAB format (3). MITAB28 describes binary interactions, one pair of interactors per row. Columns are separated by tabulations. Tools allowing to manipulate this data format are available (4).

(1) http://www.pubmedcentral.nih.gov/articlerender.fcgi?tool=pubmed&pubmedid=17925023

(2) https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3977660

(3) https://signor.uniroma2.it/scripts/causalTabInfo.php

(4) http://www.psidev.info/groups/molecular-interactions (Tools section)

Column definitions

The column contents should be as follows:

  1. Unique identifier for interactor A, represented as databaseName:identifier, where databaseName is the name of the corresponding database as defined in the PSI-MI controlled vocabulary, and identifier is the unique primary identifier of the molecule in the database. Even though identifiers from multiple databases can be separated by "|", it is recommended to give only one identifier in this column. It is recommended that proteins be identified by stable identifiers such as their UniProtKB or RefSeq accession number. Small molecules should have Chebi identifiers, nucleic acids should have embl/ddbj/genbank identifiers and gene should have entrez gene/locuslink, ensembl, or ensemblGenome identifiers. This column should never be empty ('-') except for describing intra-molecular interactions or auto-catalysis. Ex: uniprotkb:P12346
  2. Unique identifier for interactor B.
  3. Alternative identifier for interactor A, represented as databaseName:ac, where databaseName is the name of the corresponding database as defined in the PSI-MI controlled vocabulary, and ac is the primary identifier of the molecule in the database. Multiple identifiers separated by "|". It is recommended to only give database identifiers in this column. Other cross references for interactor A such as GO xrefs should be moved to the column 'Interactor xrefs A' and interactor names such as gene names should be moved to the column 'Alias A'. Ex: refseq:NP_001013128|ensembl:ENSRNOP00000012946
  4. Alternative identifier for interactor B.
  5. Aliases for A, separated by "|". Representation as databaseName:name(alias type), where databaseName is the name of the corresponding database as defined in the PSI-MI controlled vocabulary, name is the alias name and alias type is the name of the corresponding alias type as defined in the PSI-MI controlled vocabulary. In the absence of databaseName, one can use unknown. Multiple names separated by "|". In parenthesis, 'display_short' and 'display_long' are used to describe what name can be used for network display. Ex: uniprotkb:Tf(gene name)|uniprotkb:Serotransferrin(recommended name)|uniprotkb:Tf(display_short)
  6. Aliases for B.
  7. Interaction detection methods, taken from the corresponding PSI-MI controlled Vocabulary, and represented as databaseName:identifier(methodName), separated by "|". As the detection methods are taken from the PSI-MI ontology, the database name is 'psi-mi'. Interaction detection method is recommended by MIMIx and can be used for scoring interactions so it is recommended to always give this information. It is also recommended to give one interaction detection method per MITAB line (for clustering purposes). Ex: psi-mi:"MI:0006"(anti bait coimmunoprecipitation)
  8. First author: surname(s) of the publication(s) followed by 'et al.' and the publication year in parenthesis, e.g. Ciferri et al.(2005). Separated by "|".
  9. Identifier of the publication in which this interaction has been shown. Database name taken from the PSI-MI controlled vocabulary, represented as databaseName:identifier. Multiple identifiers separated by "|". The publication identifier is used for clustering interactions from different data providers in PSICQUIC and can be used for scoring interactions so it is recommended to always give this information. It is recommended to give one pubmed id per MITAB line and IMEx ids can be added. Ex: pubmed:16980971|imex:IM-1
  10. NCBI Taxonomy identifier for interactor A, represented as taxid:identifier(organism name) where the identifier is the taxon id of the organism and organism name can either be the common name or scientific name. Even though multiple identifiers can be separated by "|", it is recommended to have one organism per interactor per MITAB line. If both scientific name and common name are given, they should be represented with : taxid:id1(common name1)|taxid:id1(scientific name1). Note: Currently no taxonomy identifiers other than NCBI taxid are anticipated, apart from the use of -2 to indicate "chemical synthesis" and -3 indicates "unknown". It is recommended to always give this information for proteins and genes. For small molecules and nucleic acids, this information should be given when available but can be left empty ('-') if not relevant.
  11. NCBI Taxonomy identifier for interactor B.
  12. Interaction types, taken from the corresponding PSI-MI controlled vocabulary, and represented as dataBaseName:identifier(interactionType), separated by "|". As the interaction types are taken from the PSI-MI ontology, the database name is 'psi-mi'. Interaction type can be used for scoring interactions so it is recommended to always give this information. It is also recommended to give one interaction type per MITAB line (for clustering purposes). Ex: psi-mi:"MI:0914"(association)
  13. Source databases and identifiers, taken from the corresponding PSI-MI controlled vocabulary, and represented as databaseName:identifier(sourceName). As the detection methods are taken from the PSI-MI ontology, the database name is 'psi-mi'. Multiple source databases can be separated by "|". When the interaction has been imported and reported by different sources, it is recommended to give the original source plus the source that currently reports the interaction. Ex: psi-mi:"MI:0469"(intact)|psi-mi:"MI:0923"(irefindex)
  14. Interaction identifier(s) in the corresponding source database, represented by databaseName:identifier. It is recommended to always give a unique identifier per interaction (binary and n-ary) so in case of complexes which have been expanded, it would be possible to retrieve and re-build the original complex based on the interaction identifier. IrefIndex IDs go into the interaction checksum field. Other interaction references such as GO references go into the interaction xref field. Ex: intact:EBI-1547321|imex:IM-11865-3
  15. Confidence score: Denoted as scoreType:value where scoreType is taken from the corresponding PSI-MI controlled vocabulary. Multiple scores separated by "|". Ex: author-score:0.60|author-score:high|intact-miscore:0.36784992
  16. Complex expansion: Model used to convert n-ary interactions into binary interactions for purpose of export in MITAB file. It should be represented as databaseName:identifier(expansion name) and taken from the corresponding PSI-MI controlled vocabulary. It is recommended to always give this information when n-ary interactions have been expanded. In case of true binary interactions, this column should be empty ('-'). Ex: psi-mi:"MI:1060"(spoke expansion)
  17. Biological role A, taken from the corresponding PSI-MI controlled vocabulary, and represented as dataBaseName:identifier(biological role name), separated by "|". As the biological roles are taken from the PSI-MI ontology, the database name is 'psi-mi'. Biological roles are recommended by MIMIx so it is recommended to always give this information. When the participant A does not have any specific biological roles, the term 'unspecified role' (MI:0499) should be used. Ex: psi-mi:"MI:0684"(ancillary)
  18. Biological role B.
  19. Experimental role A, taken from the corresponding PSI-MI controlled vocabulary, and represented as dataBaseName:identifier(experimental role name), separated by "|". As the experimental roles are taken from the PSI-MI ontology, the database name is 'psi-mi'. Experimental roles are recommended by MIMIx so it is recommended to always give this information. When the participant A does not have any specific experimental roles, the term 'unspecified role' (MI:0499) should be used. Ex:psi-mi:"MI:0496"(bait)
  20. Experimental role B .
  21. Interactor type A, taken from the corresponding PSI-MI controlled vocabulary, and represented as dataBaseName:identifier(interactor type name), separated by "|". As the interactor types are taken from the PSI-MI ontology, the database name is 'psi-mi'. It is recommended to always give this information as it can be useful for recognizing protein-protein, small molecule-proteins, nucleic acids-proteins and gene-proteins interactions. Ex: psi-mi:"MI:0326"(protein)
  22. Interactor type B.
  23. Xref for interactor A, represented as databaseName:ac(text), where databaseName is the name of the corresponding database as defined in the PSI-MI controlled vocabulary, and ac is the primary accession of the molecule in the database. For example the gene ontology cross references associated. The text can be used to describe the qualifier type of the cross reference (see the corresponding PSI-MI controlled vocabulary) or could be used to give the name of the GO term in case of cross references to ontology databases. Multiple cross references separated by "|". This column aims at adding more information to describe the interactor A but cannot be used to identify the interactor A. If some sequence database accessions are ambiguous (Ex : uniprot secondary accessions that are shared between different uniprot entries and so cannot be used as identifiers of interactor A), it is possible to report them in this column. Ex: go:"GO:0003824"(catalytic activity)
  24. Xref for interactor B.
  25. Xref for the interaction, represented as databaseName:ac(text), where databaseName is the name of the corresponding database as defined in the PSI-MI controlled vocabulary, and ac is the primary accession in the database. For example the gene ontology cross references associated (components, etc.) or OMIM cross references. Multiple cross references separated by "|". The text can be used to describe the qualifier type of the cross reference (see the corresponding PSI-MI controlled vocabulary) or could be used to give the name of the GO term in case of cross references to ontology databases. Note, that this field is not meant not store interaction accessions or interaction checksums in this field. Ex: go:"GO:0005643"(nuclear pore)
  26. Annotations for interactor A, represented as topic:"text", where topic is the name of the topic as defined in the PSI-MI controlled vocabulary and text is free text associated with the topic (linebreak and other MITAB reserved characters should be properly escaped, replaced and/or removed). For example comments about this interactor : comment:"sequence not available in uniprotKb". The text is optional and only a topic could be given, e.g. anti-bacterial. Multiple annotations separated by "|".
  27. Annotations for Interactor B.
  28. Annotations for the interaction, represented as topic:"text", where topic is the name of the topic as defined in the PSI-MI controlled vocabulary and text is free text associated with the topic (linebreak and other MITAB reserved characters should be properly escaped, replaced and/or removed). For example figure legends : figure legend:"Supp Tables 1 and 2". The text is optional and only a topic could be given. This column would also be used for tagging interactions and in such a case, topics for tags are also defined in the PSI-MI controlled vocabulary except for the complex expansion tags that have their own column. Ex: internally-curated Multiple annotations separated by "|".
  29. NCBI Taxonomy identifier for the host organism, represented as taxid:identifier(organism name) where the identifier is the taxon id of the organism and organism name can either be the common name or scientific name. Multiple identifiers can be separated by "|". Cells and tissues cannot be described in this column. If both scientific name and common name are given, they should be represented with : taxid:id1(common name1)|taxid:id1(scientific name1). Note: Currently no taxonomy identifiers other than NCBI taxid are anticipated, apart from the use of -1 to indicate "in vitro", -2 to indicate "chemical synthesis", -3 indicates "unknown", -4 indicates "in vivo" and -5 indicates "in silico".
  30. Parameters of the interaction, for example kinetics. Representation as type:value(text). The type can be taken from the corresponding PSI-MI controlled vocabulary. The value syntax format is factor x base ^ exponent ~uncertainty. Multiple parameters separated by "|". Ex: kd:"4.0x2^5 ~0.3"|kd:2.0
  31. Creation date: when the curation of the publication started. Representation as yyyy/mm/dd. This field is equivalent to the release date in the PSI-XML schema. Ex:2010/10/17
  32. Update date: when the interaction was updated for the last time. Ex:2011/12/13 Representation as yyyy/mm/dd. This field does not have any equivalence in PSI-XML (could be the release date as well) and helps to keep track of changes in curated data.
  33. Checksum for interactor A, for instance the ROGID of the interactor which takes into consideration both the sequence and the organism of the interactor. Representation as methodName:checksum where methodName is the name of the method used to create the checksum. It is recommended to give the ROGID and CROGID for proteins and the standard Inchi key for small molecules. Ex: rogid:UcdngwpTSS6hG/pvQGgpp40u67I9606|crogid:UcdngwpTSS6hG/pvQGgpp40u67I9606
  34. Checksum for interactor B.
  35. Checksum for interaction, for instance the RIGID of the interaction. Representation as methodName:checksum where methodName is the name of the method used to create the checksum. Ex: rigid:"+++94o2VtVJcuk6jD3H2JZXaVYc"
  36. negative: Boolean value to distinguish positive interactions (false) from negative interactions (true). A molecular interaction between A and B entities is considered as negative (true) when an isoform of A interacts/binds to B and not A itself. For that particular isoform of A that interacts with entity B, in the corresponding binary interaction row, the negative value should be set to false. By default, if the column is empty ('-'), the negative value is considered to be false (positive interaction). Ex: true
  37. Feature(s) for interactor A: describe features for participant A such as binding sites, PTMs, tags, etc. Represented as feature_type:range(text), where feature_type is the feature type as described in the PSI-MI controlled vocabulary. For the PTMs, the MI ontology terms are obsolete and the PSI-MOD ontology should be used instead. The text can be used for feature type names, feature names, interpro cross references, etc. For instance : sufficient to bind:27-195,201-133 (IPR000785). The use of the following characters is allowed to describe a range position : ‘?’ (undetermined position), ‘n’ (n terminal range), ‘c’ (c-terminal range), ‘>x’ (greater than x), ‘<’ (less than x), ‘x1..x1’ (fuzzy range position Ex : 5..5-9..10). The character '-' is used to separate start position(s) from end position(s). Multiple features separated by '|'. Multiple ranges per feature separated by ','. However, It is not possible to represent linked features/ranges. Ex: gst tag:n-n(n-terminal region)|sufficient to bind:23-45. or binding site:23..24-46,33-33
  38. Feature(s) for interactor B.
  39. Stoichiometry for interactor A: A numerical value describing the count of instance of the molecule participating in the interaction. If no stoichiometry is available for interactor A, the column should be empty ('-'), otherwise a positive Integer value should be given. Several specific cases should be taken into consideration : in case of auto-catalysis, only one interactor is given and the stoichiometry should be 1. In case of homodimers, homotrimers, etc., the stoichiometry of one interactor should be 0 and the stoichiometry of the other should be a valid positive Integer. In case of homo-oligomer, the stoichiometry of both interactors should be 0. Example: for self interactors e.g. a kinase occluding its kinase domain by an internal phospho-tyrosine/SH2 domain interaction, only Interactor A column will show the molecule accession number with the stoichiometry 1. The columns for Interactor B will be empty. Ex: 4
  40. Stoichiometry for interactor B.
  41. Participant identification method for interactor A: taken from the corresponding PSI-MI controlled Vocabulary, and represented as databaseName:identifier(methodName), separated by "|". As the identification methods are taken from the PSI-MI ontology, the database name is 'psi-mi'. Participant detection method is recommended by MIMIx so it is recommended to always give this information. Ex: psi-mi:"MI:0102"(sequence tag identification)
  42. Participant identification method for interactor B.
  43. Biological effect of interactor A: The GO term associated with the molecular function of interactor A that is responsible for its regulatory activity. Represented as dataBaseName:identifier(biological effect name), separated by "|". As the biological effect is taken from the GO ontology, the database name should be 'go'. Also, it is recommended that at most one GO xref is given as a biological effect. Ex: go:"GO:0016301"(kinase activity)
  44. Biological effect of interactor B.
  45. Causal regulatory mechanism: This column describes the regulatory mechanism of indirect causal interactions, where the entity A is not immediately upstream of the entity B. It is taken from the corresponding PSI-MI controlled vocabulary and represented as dataBaseName:identifier(regulatoryMechanism), separated by "|". As the causal regulatory mechanism is taken from the PSI-MI ontology, the database name is 'psi-mi'. When the ‘Interaction Types’ column is equal to the psi-mi term: psi-mi:"MI:2286"(functional association), then the causal regulatory mechanism must be absolutely filled - otherwise it can be empty. Ex: psi-mi:"MI:2247"(transcriptional regulation)
  46. Causal statement: This column describes the effect of modulator entity A on a modulated entity B. It is taken from the corresponding PSI-MI controlled vocabulary and represented as dataBaseName:identifier(causalStatement), separated by "|". As the causal statement is taken from the PSI-MI ontology, the database name is 'psi-mi'. Ex: psi-mi:"MI:2240"(down regulates)

Empty columns should be represented with '-' to keep track of the columns.

Syntax

Columns are normally formed by fields delimited by "|", with a structure like this one:

<XREF>:<VALUE>(<DESCRIPTION>)

Due to the unsafe use of reserved characters in the values, we have recently added the possibility to surround <XREF>, <VALUE> or <DESCRIPTION> with quotes if they contain a special symbol.

In MI-TAB, the reserved characters are:

|
(
)
:
\t (tabulation)

Whenever this happen in your data, surround the value with double quotes:

"<XREF_WITH_RESERVED_CHARS>":"<VALUE_WITH_RESERVED_CHARS>"("<DESCRIPTION>")

Note that the quotes are before and after each part. The escaped data should look like in the following examples:

psi-mi:"MI:0000"(a cv term)
psi-mi:"MI:0000"("I can now use braces ()()() or pipes ||| here and ::colons::")

If you want to use a quote within a quote, escape it:

uniprotkb:P12345("a \"nice\" protein")

results matching ""

    No results matching ""