Extraction of structural information from ChemDraw CDX files: easy, or an underestimated, difficult challenge? Josef Eiblmaier (InfoChem, Germany) In the past decade various systems for the automatic identification and extraction of chemistry-related information from unstructured sources have emerged. They have opened up new possibilities for organizing, querying, and analyzing chemical content to support the research and development process. Patent authorities and scientific publishers make available, on a large scale, not only full text and images, but also ChemDraw CDX files for many sources. The chemical information contained in these CDX files is primarily intended for layout purposes for publications but it is often erroneously considered to be readily available as input for structure and reaction database building processes. Unfortunately, automatic work-up of chemical structures and reactions from these CDX files entails serious obstacles and problems and consequently the information produced is often incorrect or incomplete and thus not properly available to information professionals via structure and reaction searching. This talk will present different approaches to extracting reactions and structures correctly from CDX files and will describe the main difficulties and drawbacks encountered.

1. 1 / 23
Extraction of structural information from
ChemDraw CDX files: easy, or an
underestimated, difficult challenge?
Josef Eiblmaier, Hans Kraut, Sascha Hausberg, Peter Loew
ICIC 2013 Vienna, October 13 – 16
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

2. 2 / 23
Outline
» ChemDraw files:
Relevance and the Challenge
» Approach
» Projects
» InfoChem ChemProspector
© cora / PIXELIO, www.pixelio.de
» Wiley Smart Article
» Thieme Science of Synthesis Update / Pharmaceutical Substances
» Conclusion / Outlook
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

3. 3 / 23
Patents, Journal Articles and MRW‘s: a Buried Treasure?
Chemical structures
(images)
Chemical
names/fragments (text)
Markush
structures (text,
images, CDX)
Chemical structures
(CDX files)
InfoChem GmbH © 2013
Reactions (CDX files)
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

4. 4 / 23
Manuscript  Article  Database …
Publishing
Manuscript
submission
Manual
Indexing
Database production e.g.
SciFinder, Reaxys, SPRESI
eEROS, ...
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

5. 5 / 23
CDX Scheme vs. Database Record
ChemDraw file
Database
Purpose: presentation / publishing
no search
Purpose: search / retrieval
Unstructured
Structured
Structures: no strict rules
Structures: strict rules
General rules: none
Database rules: strict
Reactant
Product
Reagent
Solvent
Catalyst
LiOH
H2O, THF
Pd(OAc)2
Cl-Co2Et,
Et3N
Acetone,
H2O
SOCl2
Source: Thieme Pharmaceutical Substances, Ticagrelor (in production)
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

6. 6 / 23
CDX Scheme Processing,
what does that mean?
Chemical structures (SD files)
Reactions (RD files)
ICSchemeProcessor
Conditions (RD files)
Reagent
Solvent
Catalyst
LiOH
H2O, THF
Pd(OAc)2
Cl-Co2Et,
Et3N
Acetone,
H2O
SOCl2
Source: Thieme Pharmaceutical Substances, Ticagrelor (in production)
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

7. 7 / 23
But: CDX files, often an optical illusion!
Authors are very inventive for a ‚perfect‘ layout!
Appearences are deceiving!
» Usage of graphical symbols
• Polymer supports
• Heteroatoms
C Grid:
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

8. 8 / 23
Optical illusions 2
» Unresolvable labels
• Labels not defined
• Element symbols used as R-group labels
• Ambiguous fragment labels (e.g. molecular formula)
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

9. 9 / 23
Optical illusions 3
» Variable points of attachment
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

10. 10 / 23
Optical illusions 4
» Reaction arrows / forked arrows / brackets
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

11. 11 / 23
Approach
© Gerd Altmann / PIXELIO, www.pixelio.de
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

12. 12 / 23
Approach
» The algorithmic approach:
• Application of a set of rules in the software (generic, project unspecific). Software
should recognize all cases that might occur!
• project (title-) specific rules (drawing conventions must not change), otherwise
further development necessary
• manual post correction required (cost/time intensive)
• problem is infinite, unprecedented issues can not be handled
» The templating approach:
• software is developed to recognize a defined set of problems (PS)
• all content must be manually pre-templated (cost intensive) according to the
capabilities of the software
» The hybrid approach:
• depending on the source the focus can be laid on either approach
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

13. 13 / 23
Templating
» Templating: Guidelines for authors and typesetters
• Syntax definitions for tables, R-groups etc.
• Syntax rules for captions
• Reaction arrangement, forked arrows
• Rules for reaction conditions
(reactants, catalysts, solvents, yields, temperature)
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

14. 14 / 23
Examples:
» Algorithmic detection of features
» Resolution of repeating groups
» Enumeration of R-groups
» Resolution of aliases/labels
• source specific alias databases
• continuously extended
» Table Enumeration
• compound enumeration
• reaction factual data:
Caption/Yield
» Variable points of
attachment
» Forked arrows
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

15. 15 / 23
Projects
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

16. 16 / 23
Sucessful Application of CDX Processing:
Chemistry Enrichment Workflow*, (Wiley Smart Article)
*Reinhard Neudert: Enhancing the User Experience for Wiley Chemistry Content, ICIC 2012 14. – 17. October, Berlin
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

17. 17 / 23
Templating*
Author‘s CDX File
CDX Template
Templating
Enumerated structures
ICSchemeProcessor
CDX-Templating
Guidelines (Structures)
*Reinhard Neudert: Enhancing the User Experience for Wiley Chemistry Content, ICIC 2012 14. – 17. October, Berlin
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

18. 18 / 23
Workflow Science of Synthesis Update
R4
R4
O
R5
OH
HO
H2N
+
NH
HO
HN
 H2O
R5 NH
R4
R4
O
R5
OH
H2N
HO
R4
+
R4
NH
HO
N
HN
R
N
O
HO
R
HO
N
R5
O
N
R5
R4
R4
R5
R4
40
N
 H2O
 H2O
N
 NH3
N H2
4
••
R5 N H2
39
R4
R5
••
R5 N H2
39
N
 H2O
R5 NH
O
4
 NH3
N H2
R5
O
40
N
H
N
R5
ICSchemeProcessor
N
H
O
R5
OH
HO
H2N
+
HO
HN
Scheme
Error Report

NH
 H2O
R5 NH
R4
4
R
R4
N
N
O

Correct /
extend process
R4
R4


CDXTemplating
Guidelines
(Reactions)
R5
••
R5 N H2
39
HO
N H2
N
 NH3
R5
O
40
 H2O
R4
Scheme
correction not
possible
InfoChem GmbH © 2013
N
5
R
N
H
ICIC 2013 Vienna, October 13 – 16
Manual data
entry
Dr. Josef Eiblmaier

19. 19 / 23
Sample Pharmaceutical Substances Update
Source: Thieme Pharmaceutical Substances, Abiraterone
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

20. 20 / 23
Conclusion
» As much as possible algorithmic processing desirable
• generic: can be applied to other contents as well
• cheaper (humans cost!)
» 100% conversion (without human interaction) never possible
» Solutions are project / source specific
» Relevance of automatic extraction will continuously increase
» Authors / Publishers play an essential role in a successful conversion
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

21. 21 / 23
Acknowledgements
» Wiley
 Michael Forster
 Reinhard Neudert
» Thieme
 Guido Herrmann
 Rolf Hoppe
 Klaus Köberlein
» InfoChem
 Hans Kraut, Sascha Hausberg, Thomas Menke, Manuela Rauh
Fanny Irlinger, Huyen Ngyen, Dagmar Kunzmann
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

22. 22 / 23
© Thomas Link / Flickr
Thank you!
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier

23. 23 / 23
Questions?
InfoChem GmbH © 2013
ICIC 2013 Vienna, October 13 – 16
Dr. Josef Eiblmaier