The Cybercell Database CCDB

1
The Cybercell Database (CCDB)
Presenter Shan Sundararaj
Provisional PhD. Candidate
Department of Pharmacy
University of Alberta
Supervisor Dr. D. Wishart
Date December 12, 2002.
2
Outline

• What is the Cybercell project?
• What is the CCDB?
• How was it made how does it work?
• What can it do?
• Conclusion Future plans

3
Project Cybercell

• Objective
• Create a virtual biological cell
• Whos involved
• Researchers across Canada internationally
• Both universities and private companies
• Dozens of structural biologists, computer
  scientists, bioinformaticists, etc.

4
Project Cybercell

• Objective Create a virtual cell
• simulate all elements and processes of a
  biological cell on a computer
• use Escherichia coli as a model
• use it to predict cellular phenomena
• e.g. response to drugs or environment
• Test simulations with the best quantitative data
  possible

5
Cybercell example
A model of a cell with a metabolite diffusing
into it through transport proteins and being
converted to a different molecule by an enzyme
http//129.128.166.250/research/simulation_gallery
/4d/cybercell4.mov
This simulation requires very specific
information about enzyme availability and rates
to be useful
6
Cybercell Database (CCDB)
7
Outline

• What is the Cybercell project?
• What is the CCDB?
• How was it made how does it work?
• What can it do?
• Conclusion Future plans

8
CCDB

• the CCDB is an evolving repository of
  quantitative data on proteins and genes compiled
  from a variety of data sources
• Includes browsable, searchable and sortable lists
  and links of

• protein names
• sequences
• functions
• structures
• physicochemical
• constants

• cofactors
• copy numbers
• products
• reactants
• binding partners
• More

9
CCDB Statistics
10
The Colicard

• Fundamental unit of the database (e.g. the
  record)
• One exists for all 4374 proteins identified in E.
  coli K12

11
Interesting Features

• Can view 3-D structures using WebMol
• Homology models created for proteins with
  appropriate template
• Contains lists of protein interacting partners
  and protein complexes
• Useful when looking for homologous interacting
  proteins
• Enzyme information from BRENDA

12
Related Databases

• CCRNA
• Equivalent database for tRNA and rRNA molecules
• CC3D
• Database that focuses on structural information
• CCMD
• Database of data on small metabolic compounds

13
Outline

• What is the Cybercell project?
• What is the CCDB?
• How was it made how does it work?
• What can it do?
• Conclusion Future plans

14
Data gathering

• There is a VAST amount of varied information
  available about proteins, nucleic acids and other
  molecules!!
• All stored nicely in the CCDB now, but where does
  it originate?

15
Information Sources
Manual Literature Searches
Other databases
16
Database structure

• CCDB is a flat-file database
• Individual text file for each Colicard
• Create an index file of all the pertinent
  information and use Perl cgi scripts to make a
  searchable, sortable web interface
• Contains many links to references and other
  databases for more detailed information
• Use other tools to perform search and data
  extraction functions for extra functionality

17
Data Gathering

• Would take years to do all manually!
• Use manual forms to fill in some information
• Make use of web robot to query databases and
  build as much of Colicards as possible

18
Robots Can Be Annoying

• Many sites do not like robots (e.g. EcoCyc)
• Some sites disallow robots altogether (e.g.
  BRENDA)
• Have to be polite, look at robots.txt file, dont
  query too fast or too often

19
Update robots (cont.)

• Weekly check to web-based databases to look for
  updates
• If update available, ALL updated files are
  downloaded
• Information is parsed and compared to last
  available local copy (e.g. line-by-line
  comparison of Swiss-Prot files)

20
Outline

• What is the Cybercell project?
• What is the CCDB?
• How was it made how does it work?
• What can it do?
• Conclusion Future plans

21
Search/Sort functions

• Main browsing page allows you to sort by any of
  11 characteristics, for example
• Protein Name
• Swiss Prot ID
• of amino acids
• Function
• Gene Name
• Good for broad overview, but what if you want to
  look at a specific protein?

http//redpoll.pharmacy.ualberta.ca/CCDB/cgi-bin/E
CARD_BROWS_NEWn.cgi?hits20browsn8pag1acco11

22
Data Extractor

• Simple to use JavaScript-based interface
• Can choose proteins based on ANY type of
  information in database
• Uses pre-parsed pairwise data sets and cgi
  scripts to extract and display data

http//redpoll.pharmacy.ualberta.ca/CCDB/CCDB_Ext.
html
23
WebGlimpse Search

• Even simpler, though less robust search method
• Uses program called WebGlimpse that indexes all
  the Colicard files and returns those that match
  the search criteria somewhere in the file
• Useful because it allows for misspellings, which
  the data extractor does not

24
The Whole Thing

• The entire CCDB database in text files is
  available for download at http//redpoll.pharmacy
  .ualberta.ca/bahram/cgi-bin/Pdownload.cgi
• This is for people who would like to have a local
  copy of all the information

25
Outline

• What is the Cybercell project?
• What is the CCDB?
• How was it made how does it work?
• What can it do?
• Conclusion Future plans

26
Conclusion

• The Cybercell project aims to simulate a
  biological cell on a computer in the next 5-10
  years
• The CCDB will act as a repository of all known
  and discovered information regarding E. coli
  molecules during the project
• CCDB is updated automatically as well as by
  manual forms to remain up-to-date
• CCDB can be queried in several ways (browsing,
  data extractor, search)

27
Thanks

• Dr. Wishart
• Bahram Habibi-Nazhad
• An Chi Guo
• Haiyan Zhang
• Melania
• Everyone in lab