ChEBI: The story so far

1
ChEBI The story so far
2
Private Data
Public Data
3
The state of affairs of bioinformatics in 2002

• Bioinformatics is booming
• Human Genome sequence rough draft published June
  2000
• Free resources and free data

4
A different story for chemoinformatics

• Private data and private software

5
Too hard to solve lets put our head in the sand
6
Bioinformatics data too large to keep track of
chemical compounds

• 100000 Protein entries in SwissProt (2002)
• 20 million entries in EMBL Database (2002)
• Small databases unable to keep track
• ENZYME resources 3500 enzymatic reactions

7
New initiatives start up

• PubChem
• Chemical repository, millions of entries, focus
  on screening assays
• ChEBI
• Manually annotated database, nomenclature
  reference and compound database, tens of
  thousands of entries

8
Principles of foundation

• December 2002 email exchanges within the EBI to
  address the issue of chemistry
• Three principles outlined

9

• Nothing held in the database must be
  proprietary or derived from a proprietary source
  that would limit its free distribution/availabilit
  y to anyone.

10
Every data item in the database should be fully
traceable and explicitly referenced to the
original source/version.
11
Although the EBI will provide a web interface,
the entirety of the data should be available to
all without constraint as, for example, SQL table
dumps, ASCII tables, and XML (e.g. DAMLOIL)
12
We make a start using existing resources

• Integrate three resources
• KEGG Compound
• IntEnz
• Chemical Ontology
• Annotation starts summer 2003
• Focus on nomenclature

13
Our first release was modest but it was a start

• 21 July 2004
• 2783 annotated entities
• Data
• ChEBI Name, ChEBI Id
• IUPAC Names, Synonyms
• Formula
• Cross-references

14
We introduce structures - Sep 2005

• Molfiles
• InChI (IUPAC International Chemical Identifier)
• SMILES (Simplified Molecular Input Line Entry
  System)
• Image (PNG)

15
Marvin in ChEBI
16
We start editing the chemical ontology Dec 2005
17
Web Services - Oct 2006

• Programmatic access to a ChEBI entry
• SOAP based Java implementation
• Clients currently available in Java and perl
• Four methods with which to access data
• getLiteEntity
• getCompleteEntity
• getOntologyParents
• getOntologyChildren

18
Automated Cross References Aug 2007
Current Databases UniProtKB, Reactome,
BioModels, IntAct, SABIO-RK, PubChem and
ArrayExpress
19
Chemical Structure Searching May 2008
20
After all this, where are we?
21
(No Transcript)
22
(No Transcript)
23
Annotation is linear
24
Diversity of users

• Constant challenge of balancing our users' varied
  interests.

25
Our positives

• Nomenclature database
• Manually annotated data
• Attention to detail
• Free and accessible
• Loyal users

26
Our not so positives

• Size for some people
• Not well integrated into other bioinformatics
  resources
• Community interaction
• No software publicly available to manipulate the
  database

27
Involve the community

• Create a submission web based tool
• Users can easily submit their entities on a one
  to one basis
• Also allowing bulk submission from other
  resources.

28
Improvements to data depth

• Addition of more Xrefs PDB, MACIE ???
• Addition of more chemical attributes? What
  chemical attributes?
• Text mining projects to extract relevant chemical
  information from patents, journals
• European Patent Office

29
Going Open Source

• Commercial software packages will be replaced
  with Open Source
• Long term goal allow people to create a free
  local installation of ChEBI
• Distribution of data in useful formats CML, SDF

30
Acknowledgements

• IntEnz Team
• Rafael Alcantára, Volker Ast, Kristian Axelsen,
  Anne Morgat
• EPO Collaborators
• Hélène Courrier, Stephane Nauche, Jeremy Parsons
• Database supporters
• ArrayExpress, IntAct, Reactome, SABIO-RK, RSC,
  GO, RESID etc

• ChEBI Team
• Paula de Matos, Kirill Degtyarenko, Marcus Ennis,
  Janna Hastings, Christoph Steinbeck
• Alumni
• Michael Darsow, Mickael Guedj, Alan McNaught,
  Martin Zbinden
• ChEBI supporters
• Rolf Apweiler, Michael Ashburner, Henning
  Hermjakob, Janet Thornton

31
Requirement for submitting data to ChEBI

• Disclaimer this is only the summary of a chat I
  have had with the ChEBI coordinator last night.
  So no promises !
• Information needed to submit a compound
• Structure
• Name, synonyms
• Registry
• Database accession(s)
• Mapping to ChEBI Ontology
• ChEBI currently quite busy with ongoing projects,
  but would consider taking submissions.

ChEBI The story so far
31
32
What Could be done within APO-SYS

• From Pekkas talk, I gathered that there are
  about 5,000 to 10,000 compounds in these siRNA
  libraries.
• Question who else is dealing with compounds in
  APO-SYS?
• One could use the ChEBIs web service using InCHI
  to identify what is already in the database.
• ChEBI can do targeted curation provided funding
  for the curation team.

ChEBI The story so far
32