The Red Pill

1
The Red Pill
Roger Sayle, Geoff Skillman, Matthew Stahl Robert
Tolbert OpenEye Scientific Software
2
Integration

• The process of computing an integral the inverse
  of differentiation.

3
Integration

• The organization of the psychological or social
  traits and tendencies of a personality into a
  harmonious whole.

4
Data Integration

• Merge (data) into a harmonious whole
• Chaining data generation
• Extensible data storage

5
OEChem

• Programming toolkit
• Python/C API's
• Public API
• Precise handling of chemistry
• Multiple models of chemistry
• Aromaticity
• Atom types
• Valence models
• Query semantics

6
Perception

• Kekule form
• Aromaticity (Daylight, Tripos, Merck, MDL,
  OpenEye)
• Atom types
• Topological symmetry
• Stereochemistry (tetrahedral, cis/trans)
• Partial charges
• Biomonomers recognition
• Bond orders from coordinates

7
Aromaticity Models
8
Data Integration

• Merge (data) into a harmonious whole
• Chaining data generation
• Extensible data storage

9
Chaining Data Generation
Software A
Software B
Data
-Challenging in a heterogeneous software
environment -Lossless data conversion -Feature
perception
10
Extensible Data Storage
Source
Source
Data
Source
Source
11
Question

• How often do people (mis)use SD files for
  attaching data to molecules?

12
Extensible Data Structures
Python atom.SetStringData(Spam,Eggs) atom.Get
StringData(Spam) C class Foo Foo
foo mol.SetData(VeryNiceData,foo) mol.GetDatalt
Foogt(VeryNiceData)
13
Chemical EXchange

• An interchange language to enable components to
  communicate
• Model similar to Unix pipes and single purpose
  commands
• CEX stream contains objects (molecule, message)
• Extensible named property/value pairs
• Each component in the CEX pipeline can read some
  objects and properties from the input stream and
  add new ones to the output stream

14
OEBinary V2

• Extensible tag/data format
• Heirarchical
• Persistent objects (automatic for POD types)
• Dynamic data parsing
• Efficient storage of conformers
• Ideal for storage as BLOB
• Lossless data storage possible
• Definition publicly available

15
Conclusions

• I have no idea what 'data integration' really
  means
• OEChem maintains the integrity of chemical data
• Extensible persistent data structures likely
  facilitate data integration
• OEChem provides extensible persistent data
  structures
• OEChem likely facilitates data integration

16
Acknowledgments

• Geoff Skillman
• Bob Tolbert
• Roger Sayle
• AstraZeneca Pharmaceuticals
• Vertex Pharmaceuticals