1. ChemAxon - Pipeline Pilot Integration
Ton van Daelen
Accelrys, Inc.

2. Outline
Pipeline Pilot
ChemAxon components
Integration technologies

3. Pipeline Pilot
Instructions: The purpose of this is to start with a cool protocol – NOT to explain PP at this point (don’t get into what components, protocols etc are yet) – show this running live or substitute a protocol tailored to the account opportunity.
Script: We will begin by showing you one application solution built using Pipeline Pilot. It shows a data pipeline that accesses two data sources, one chemistry and one biology, joins the data together, applies some business rules and then produces output that allows you to understand the analysis and make decisions based on the results.

4. Aggregate and Disseminate Information…
An Example of How To Step Forward In The Informatics Evolution
Your real challenge and investment needs to be in tying together information and provide rich scientific web services. Delivering information needs to be quick and easy. Web-based, zero deployment, and zero footprint technologies are the fastest and easiest way to get information in the hands of Scientists.
Creating content in Pipeline Pilot automatically allows you to:
Deploy web services
Deploy Portlets to your portal server
Deploy web applications
Power custom desktop clients
Provide scientific services to containers like Spotfire and Diva
Easily Configurable Apps
Into a Web Service Callable
By Any Type Of Interface

5. Release history - major changes
Version 1.6, August 2009
New component: "ChemAxon 3D Conformers"
Version 1.5, May 2009
New components: "ChemAxon MolConverter", "ChemAxon Tautomerization", "ChemAxon Markush Enumeration"
Version 1.4, November 2008
New components: "LibMCS Clustering", "Molecule to IUPAC Name", "Molecule from IUPAC Name“
Major upgrade of "ChemAxon Reactor" component
Version 1.3, July 2008
New component: “Chemical Terms Calculator”
Version 1.2, March 2008
New components: “ChemAxon Reactor”, “Drop JChem Base Table”, “Create JChem Base Table”
Several components upgraded

6. Calculator
Easy access for the most important calculations

7. Chemical Terms Calculator
Maximum freedom trough Chemical Terms Expressions for the expert user
New in 1.3
Use arbitrary Chemical Terms expressions
Results stored to arbitrary properties
A wide range of ChemAxon functionality can be accessed as Chemical Terms functions

8. Reactor Improvements: Upgraded in 1.4 Synthesis code generation
Output reaction mapping
Advanced options:
Unambiguous only
Ignore rules:
Reactivity and Exclude
Selectivity
Tolerance

9. IUPAC naming components
New in 1.4
IUPAC Name to Molecule
Molecule to IUPAC Name
Example “roundtrip” protocol:

10. Clustering with LibMCS
Maximum Common Substructure (MCS) based clustering
New in 1.4
Options:
Size of smallest common substructure to consider
Three levels of heuristics:
Exact (no heuristics)
Fast
Very Fast
Bond type, atom type, charge can optionally be ignored
Disallow “breaking” rings (default)

11. Markush Enumeration Enumeration of generic structures New in 1.5
File input
Enumeration type:
Sequential
Random
Number of enumerated structures can be limited (per input structure)
Valence filter
Scaffold alignment
Markush code generation. The scaffold ID can be:
fetched from data field
generated (prefix + number)

12. Tautomerization Component for tautomer generation New in 1.5
Calculation modes:
All tautomers
Canonical tautomer
Generic tautomer
Major tautomer
Dominant tautomer distribution
Options:
Protect aromaticity, charge, double bond stereo, tetrahedral stereo
Exclude antiaromatic compounds
Single fragment mode
Consider pH at specific value

13. MolConverter “Swiss army knife” for molecular format conversion
New in 1.5
Input and output can either be
File
Property
Pipeline Pilot Molecule
Specified input format or auto-detection
Various output formats or custom format string
Option to halt or continue on error, error messages put into property
2D cleaning (coordinate generation) only when needed (default). Unconditional 2D or 3D cleaning or no cleaning can also be selected

14. Conformer generation Component for 3D conformer generation New in 1.6
Calculation modes:
Multiple conformers
Lowes energy conformer
Options:
Maximum number of conformers
Diversity limit
Optimization limit, hyperfine option
Time limit
Generate with explicit H atoms
Energy unit kcal/mol or KJ/mol, into arbitrary property

15. Pipeline Pilot Integration
“Client side”
Browser Clients
Native and Java Clients
Various Clients
JavaScript
SDK
.NET
SDK
Java
SDK
SOAP
SDK
Web services
Pipeline Pilot
“Server side”
Java
API
Perl
API
SOAP
Cmd
Line
Telnet /
FTP
ODBC
JDBC
Application integration layer
Data access layer
Oracle
ISIS
Accord
IBM II
Application Services
Data Sources

16. Telnet/FTP Component
Execute command line applications on remote server (any platform)
No coding required
Use files to pass input and output data
Secure version (SSH)

17. Integration of SOAP Web Services
Simple method example
public double GetCLogP(string SMILES) {
// ... Do the calculation ...
return clogp; }
Configuring SOAP Component
SOAP Endpoint - HTTP URL resource
Method Name - Name of function (“GetCLogP”)
Method Namespace - Found in WSDL (often
SOAP Parameter Names - Name of function parameter (“Name”)
Method Parameter Values - Pipeline Pilot property names
Property Return Names - What to call the returned value (“Greeting”)

18. Java on Server component
Create new PP components
Filters, calculators, readers, writers
Single class with three methods
onInitialize
onProcess
onFinalize
Referencing Java classes on server
Interact with Molecular Toolkit

19. New Java class: ChemAxonComponent
Parent class for component classes, implements com.scitegic.pilot.Component
Error reporting: all exceptions thrown in onInitializeBody(), onProcessBody(), onFinalizeBody() are caught, logged and reported (re-thrown with added information)
Parameters and context always readily available
Utility functions for conveniently accessing parameters and properties, e.g.: findStringParameter(String name)
Reduced code redundancy, better readability

20. Java Molecular Weight Example
private static final double HYDROGEN_WEIGHT = Hydrogen.atomicWeight(); // get molecule and properties, skip data record if molecule is not found Molecule mol = Molecule.findMolecule(data); if (mol == null) return Component.State.ReadyForInputData; PropertyCollection props = mol.getProperties(); // calculate mol weight and exact weight (most common isotopes) double mw = 0.0; for (Atom atom : mol.getAtoms()) { int numImplH = atom.getNumImplicitHydrogens(); mw += atom.getAtomicWeight(); mw += numImplH*HYDROGEN_WEIGHT; } props.define("Java_MW", mw);

21. Conclusions Compatible architectures
Pipeline Pilot good prototyping environment for ChemAxon’s advanced science
Multiple deployment options

22. Future Work Integrate more ChemAxon functionality
Let us know what your priorities are
Acknowledgements
Szilard Dorant
Moises Hassan
Keith Burdick
Contact
Alex Allardyce:
Ton van Daelen: