Transcript underlying technology

MCM Website
The underlying technology
Dr Stephen Pascoe
British Atmospheric Data Centre
Evolving use of the Web



Use of the web has changed a lot since 1995.
Websites are increasingly interfaces to dynamic
applications
Database driven backends.
2 Tier design
MCM Server
Update
MySQL Database
Query
Query
Response
Developer
Web Server
URL
HTML
Web Browser
Requirements
1.
2.
3.
4.
5.
6.
Find reactions of a species
quickly.
Search for species by SMILES
string.
Extract sub-mechanisms.
Depict (draw) structures.
Attach context (metadata) to
reactions.
Interact easily with current
developers tools (Accord)
Indexing
Coordinates
Database design
Indexing

What can we index?
Unique  NOT Systematic
 SMILES string: Systematic  NOT Unique


MCM name:
We need a Unique Systematic Identifier.


Unique linear representation: USMILES, IChI
Canonical number. Accord, ICAMS.
Searching
Accord can do this
BUT not on the web server
Web form
Search:
C(O)CC
Calculate
Canonical No.
7438297
Database
Lookup
Name
SMILES
Canon
PROPANOL
C(O)CC
7438297
…
…
…
Finding Reactions
Relational approach
Raction id
Species
Type
1
NBUTANE
Reactant
1
OH
Reactant
MySQL
1
CCCO[O.]
Product
Full Text Search
2
…
…
Indexed words
Accord DB
Reaction id
Expression
Rate
1
NBUTANE + OH = CCCO[O.]
…
2
…
…
Developer interaction
MySQL Tables
Local Tables
Species
Name
Species
SMILES
Name
Structural information
Accord data
Canonical number
Accord
Reactions
Reaction expression
Rate expression
Reaction category
Access
ODBC
Parameters
Reactions
Reaction expression
Rate expression
Reaction category
Depicting

The hard part is generating the 2D coordinates.
Accord can do this
BUT not on the web server

We need a connection table.
 MDL Mol file
MCM Server
MySQL Database
Connection Table
Insert
Export
Accord Chemistry
MDL Molfile
Render
PNG image
Developer interaction
MySQL Tables
Local Tables
Export Mol
Name
Species
SMILES
Name
Structural information
Accord data
Canonical number
Accord
Reactions
Reaction expression
Rate expression
Reaction category
Species
Access
ODBC
Parameters
Reactions
Reaction expression
Rate expression
Reaction category
Calc canon.
Context





Every reaction belongs to a
category.
Categories are web template URL.
General categories can be refined
into more specific ones.
Rate parameters also have a
context URL.
By explicitly parameterising SAR
rates we could provide full context
for rates.
Future

Context for rate coefficients




Know the source of a rate.
Evaluate uncertainty
We would need to redo all SAR calculations.
New developments in Cheminformatics



CML: Could replace Mol files
IChI: The new SMILES?
OpenSource development tools
• CDK: Depiction & uninque id. Generation.
Integrated species editor.
The End
Thanks.