The PSI MI Common Query Interface

1
The PSI MI Common Query Interface

• PSICQUIC

Bruno Aranda - IntAct
2
Agenda

• Background and motivation
• Common service definition
• Common query language
• Future

3
Background and motivation

• Two PSI MI standards (PSIMI-TAB, PSIMI-XML) allow
  the exchange of molecular interaction (MI)
  information between databases.
• The IMEx Consortium member databases are getting
  ready to exchange data using the PSIMI-XML
  format.
• Tools to deal with these standards already exist.
• Even if standards are the same, the way to get
  the data from the different MI databases is
  different.
• It would be great if MI data could be provided in
  the same way by the different providers! -gt
  PSICQUIC

4
Background and motivation

• Currently, creating a client to retrieve MI data
  from differentdatabases can be tricky.

• We want to be able to createone client to
  retrieve data fromany MI database, in a
  commonand integrated way.

5
Background and motivation

• An early version of PSICQUIC was already proposed
  a few years ago, and a proof of concept exists.
• BioHackathon 2008 in Japan IMEx Consortium
  members started to work in a new and up to date
  version of PSICQUIC, with the direct feedback of
  MI data consumers (e.g. Cytoscape)
• Work in progress, but specification exists (in
  the form of a WSDL file)

6
Agenda

• Background and motivation
• Common service definition
• Common query language
• Future

7
Common Service Definition

• PSICQUIC defines a minimum list of standard
  methods to be implemented by every MI provider.
• Every MI database should provide a Web Service to
  provide the information.
• Once the user knows the methods, the same code
  can be reused to query the different MI
  databases.
• This methods will return results in both
  PSIMI-XML and PSIMI-TAB formats.
• A provider may choose to return additional
  formats, additionally.

8
Common Service Definition

• List of methods
• getByInteractor (dbRef)
• getByInteraction (dbRef)
• getByInteractorList (dbRefList)
• getByInteractionList (dbRefList)
• getByBetweenList (dbRefList)
• getByQuery (query)
• getSupportedReturnTypes
• getSupportedQueryTypes
• getSupportedDbRefTypes
• getCounts/Stats/Other service metadata (e.g. last
  update)
• All queries can be paginated, and exceptions have
  been defined too.
• Queries accept cross reference IDs and results
  are returned in both PSI-MI formats.

9
Agenda

• Background and motivation
• Common service definition
• Common query language
• Future

10
Common Query Language

• The MI Query Language (MIQL) allows more powerful
  and flexible queries.
• Designed for fast searches on PSI-MI TAB files.
• All fields can be searched.
• IntAct uses MIQL in its main search engine
  already, and provides a Web Service and Open
  Source libraries.
• Currently used by Cytoscape (and others).

11
Common Query Language

• The MIQL syntax is based on Lucene 1. A query
  is broken into terms and operators
• Terms single words or phrases (group of words
  surrounded by quotes). E.g. brca2 AND pull down
• Fields search in specific columns. E.g. idbrca2
  AND specieshuman
• Term modifiers wildcard searches, fuzzy
  searches, proximity and range searches. E.g. brc
  OR authbluno
• Operands OR (or space), AND, NOT, , -. E.g.
  brca2 AND rpa1 / brca2 NOT mouse / brca2 -mouse
• Grouping and field grouping brca2 AND (mouse "in
  vitro")

1 http//lucene.apache.org/java/docs/queryparser
syntax.html
12
Common Query Language

• type"direct interaction" type"enzymatic
  reaction" typephosphorylation typemethylation
  typeglycosylation typesumoylation typecleavage
  type"dna cleavage" type"mrna cleavage"
  type"protein cleavage" type"lipid cleavage"
  typedefarnesylation typedemyristoylation
  typedegeranylation typedepalmitoylation
  type"triphosphatase react" type"atpase
  reaction" type"gtpase reaction"
  typedeglycosylation typehydroxylation
  typedeformylation typedeubiquitination
  type"lipid addition" typegeranylgeranylation
  typemyristoylation typepalmitoylation
  typefarnesylation typeelongation
  typedeneddylation type"adp ribosylation"
  typedesumoylation typeformylation
  typedephosphorylation typedeacetylation
  typeamidation typeubiquitination
  typeacetylation typeneddylation
  typephosphotransfer typedemethylation
  type"covalent binding" type"disulfide bond"
  typetransglutamination

Generated by IntAct Advanced Searchhttp//www.ebi
.ac.uk/intact/binary-search/faces/search.xhtml?adv
Searchtrue
13
Overview

• Background and motivation
• Common service definition
• Common query language
• Future

14
Future

• Once the different MI data providers implement
  PSICQUIC, it will be possible to have the One
  (and simple!) Client to Rule them All.
• DIP, IntAct and MINT already working on an
  implementation of the Web Service. We expect
  others to follow.

15
Acknowledgements
INTACT Henning Hermjakob Sam Kerrien Sandra
Orchard Luisa Montecchi-Palazzi Jyoti
Khadake David Thorneycroft Achuthan Premanand
DIP Lukasz Salwinski MINT Arnaud
Ceol STRING Michael Kuhn MIPS Matthias
Oesterheld
CYTOSCAPE Keiichiro Ono
CELLDESIGNER Akira Funahashi
16
THANKS! Questions?
Bruno Aranda - IntAct