We describe a prototype that performs structure-based classification of molecular structures. The software we present implements a sound and com- plete reasoning procedure of a formalism that extends logic programming and builds upon the DLV deductive databases system. We capture a wide range of chemical classes that are not expressible with OWL-based formalisms such as cyclic molecules, saturated molecules and alkanes. In terms of performance, a no- ticeable improvement is observed in comparison with previous approaches. Our evaluation has discovered subsumptions that are missing from the the manually curated ChEBI ontology as well as discrepancies with respect to existing subclass relations. We illustrate thus the potential of an ontology language which is suit- able for the Life Sciences domain and exhibits an encouraging balance between expressive power and practical feasibility.

1. O NTOLOGY-BASED
C LASSIFICATION OF M OLECULES :
A L OGIC P ROGRAMMING A PPROACH
Despoina Magka
Department of Computer Science, University of Oxford
November 30, 2012

2. B IOINFORMATICS AND S EMANTIC T ECHNOLOGIES
Life sciences data deluge
1

3. B IOINFORMATICS AND S EMANTIC T ECHNOLOGIES
Life sciences data deluge
Hierarchical organisation of biochemical knowledge
1

4. B IOINFORMATICS AND S EMANTIC T ECHNOLOGIES
Life sciences data deluge
Hierarchical organisation of biochemical knowledge
1

5. B IOINFORMATICS AND S EMANTIC T ECHNOLOGIES
Life sciences data deluge
Hierarchical organisation of biochemical knowledge
1

6. B IOINFORMATICS AND S EMANTIC T ECHNOLOGIES
Life sciences data deluge
Hierarchical organisation of biochemical knowledge
Fast, automatic and repeatable classification driven by
Semantic technologies
1

7. B IOINFORMATICS AND S EMANTIC T ECHNOLOGIES
Life sciences data deluge
Hierarchical organisation of biochemical knowledge
Fast, automatic and repeatable classification driven by
Semantic technologies
Web Ontology Language, a W3C standard family
of logic-based formalisms
1

8. B IOINFORMATICS AND S EMANTIC T ECHNOLOGIES
Life sciences data deluge
Hierarchical organisation of biochemical knowledge
Fast, automatic and repeatable classification driven by
Semantic technologies
Web Ontology Language, a W3C standard family
of logic-based formalisms
OWL bio- and chemo-ontologies widely adopted
1

9. T HE C H EBI O NTOLOGY
OWL ontology Chemical Entities of Biological Interest
2

10. T HE C H EBI O NTOLOGY
OWL ontology Chemical Entities of Biological Interest
Dictionary of molecules with taxonomical information
2

11. T HE C H EBI O NTOLOGY
OWL ontology Chemical Entities of Biological Interest
Dictionary of molecules with taxonomical information
caffeine is a cyclic molecule
2

12. T HE C H EBI O NTOLOGY
OWL ontology Chemical Entities of Biological Interest
Dictionary of molecules with taxonomical information
serotonin is an organic molecule
2

13. T HE C H EBI O NTOLOGY
OWL ontology Chemical Entities of Biological Interest
Dictionary of molecules with taxonomical information
ascorbic acid is a carboxylic ester
2

14. T HE C H EBI O NTOLOGY
OWL ontology Chemical Entities of Biological Interest
Dictionary of molecules with taxonomical information
Pharmaceutical design and study of biological pathways
2

15. T HE C H EBI O NTOLOGY
OWL ontology Chemical Entities of Biological Interest
Dictionary of molecules with taxonomical information
Pharmaceutical design and study of biological pathways
ChEBI is manually incremented
2

16. T HE C H EBI O NTOLOGY
OWL ontology Chemical Entities of Biological Interest
Dictionary of molecules with taxonomical information
Pharmaceutical design and study of biological pathways
ChEBI is manually incremented
Currently ~30,000 chemical entities, expands at 3,500/yr
2

17. T HE C H EBI O NTOLOGY
OWL ontology Chemical Entities of Biological Interest
Dictionary of molecules with taxonomical information
Pharmaceutical design and study of biological pathways
ChEBI is manually incremented
Currently ~30,000 chemical entities, expands at 3,500/yr
Existing chemical databases describe millions of molecules
2

18. T HE C H EBI O NTOLOGY
OWL ontology Chemical Entities of Biological Interest
Dictionary of molecules with taxonomical information
Pharmaceutical design and study of biological pathways
ChEBI is manually incremented
Currently ~30,000 chemical entities, expands at 3,500/yr
Existing chemical databases describe millions of molecules
Speed up growth by automating chemical classification
2

19. E XPRESSIVITY L IMITATIONS OF OWL
1 At least one tree-shaped model for each consistent OWL
ontology problematic representation of cycles
3

20. E XPRESSIVITY L IMITATIONS OF OWL
1 At least one tree-shaped model for each consistent OWL
ontology problematic representation of cycles
E XAMPLE
C C
C C
3

21. E XPRESSIVITY L IMITATIONS OF OWL
1 At least one tree-shaped model for each consistent OWL
ontology problematic representation of cycles
E XAMPLE
Cyclobutane ∃(= 4)hasAtom.(Carbon ∃(= 2)hasBond.Carbon)
C C
C C
3

22. E XPRESSIVITY L IMITATIONS OF OWL
1 At least one tree-shaped model for each consistent OWL
ontology problematic representation of cycles
E XAMPLE
Cyclobutane ∃(= 4)hasAtom.(Carbon ∃(= 2)hasBond.Carbon)
C C
C C
3

23. E XPRESSIVITY L IMITATIONS OF OWL
1 At least one tree-shaped model for each consistent OWL
ontology problematic representation of cycles
E XAMPLE
Cyclobutane ∃(= 4)hasAtom.(Carbon ∃(= 2)hasBond.Carbon)
C C
C C
OWL-based reasoning support
1 Is cyclobutane a cyclic molecule?
3

24. E XPRESSIVITY L IMITATIONS OF OWL
1 At least one tree-shaped model for each consistent OWL
ontology problematic representation of cycles
2 No minimality condition on the models hard to axiomatise
classes based on the absence of attributes
E XAMPLE
Cyclobutane ∃(= 4)hasAtom.(Carbon ∃(= 2)hasBond.Carbon)
C C
C C
OWL-based reasoning support
1 Is cyclobutane a cyclic molecule?
3

25. E XPRESSIVITY L IMITATIONS OF OWL
1 At least one tree-shaped model for each consistent OWL
ontology problematic representation of cycles
2 No minimality condition on the models hard to axiomatise
classes based on the absence of attributes
E XAMPLE
Cyclobutane ∃(= 4)hasAtom.(Carbon ∃(= 2)hasBond.Carbon)
Oxygen
C C
C C
OWL-based reasoning support
1 Is cyclobutane a cyclic molecule?
3

26. E XPRESSIVITY L IMITATIONS OF OWL
1 At least one tree-shaped model for each consistent OWL
ontology problematic representation of cycles
2 No minimality condition on the models hard to axiomatise
classes based on the absence of attributes
E XAMPLE
Cyclobutane ∃(= 4)hasAtom.(Carbon ∃(= 2)hasBond.Carbon)
Oxygen
C C
C C
OWL-based reasoning support
1 Is cyclobutane a cyclic molecule?
2 Is cyclobutane a hydrocarbon?
3

27. E XPRESSIVITY L IMITATIONS OF OWL
1 At least one tree-shaped model for each consistent OWL
ontology problematic representation of cycles
2 No minimality condition on the models hard to axiomatise
classes based on the absence of attributes
E XAMPLE
Cyclobutane ∃(= 4)hasAtom.(Carbon ∃(= 2)hasBond.Carbon)
Oxygen
C C
C C
3

28. E XPRESSIVITY L IMITATIONS OF OWL
1 At least one tree-shaped model for each consistent OWL
ontology problematic representation of cycles
2 No minimality condition on the models hard to axiomatise
classes based on the absence of attributes
E XAMPLE
Cyclobutane ∃(= 4)hasAtom.(Carbon ∃(= 2)hasBond.Carbon)
Oxygen
C C
C C
Required reasoning support
1 Is cyclobutane a cyclic molecule?
2 Is cyclobutane a hydrocarbon?
3

29. E XPRESSIVITY L IMITATIONS OF OWL
1 At least one tree-shaped model for each consistent OWL
ontology problematic representation of cycles
2 No minimality condition on the models hard to axiomatise
classes based on the absence of attributes
E XAMPLE
Cyclobutane ∃(= 4)hasAtom.(Carbon ∃(= 2)hasBond.Carbon)
Oxygen
C C
C C
Required reasoning support
1 Is cyclobutane a cyclic molecule?
2 Is cyclobutane a hydrocarbon?
3

30. R ESULTS OVERVIEW
1 Expressive and decidable formalism for modelling complex
objects: Description Graphs Logic Programs
4

31. R ESULTS OVERVIEW
1 Expressive and decidable formalism for modelling complex
objects: Description Graphs Logic Programs
2 Modelling that spans a wide range of structure-dependent
classes of molecules
4

32. R ESULTS OVERVIEW
1 Expressive and decidable formalism for modelling complex
objects: Description Graphs Logic Programs
2 Modelling that spans a wide range of structure-dependent
classes of molecules
3 Implementation that draws upon DLV and performs
structure-based classification with a significant speedup
4

33. R ESULTS OVERVIEW
1 Expressive and decidable formalism for modelling complex
objects: Description Graphs Logic Programs
2 Modelling that spans a wide range of structure-dependent
classes of molecules
3 Implementation that draws upon DLV and performs
structure-based classification with a significant speedup
4 Evaluation over part of the manually curated ChEBI
ontology revealed modelling errors
4

34. R ESULTS OVERVIEW
1 Expressive and decidable formalism for modelling complex
objects: Description Graphs Logic Programs
2 Modelling that spans a wide range of structure-dependent
classes of molecules
3 Implementation that draws upon DLV and performs
structure-based classification with a significant speedup
4 Evaluation over part of the manually curated ChEBI
ontology revealed modelling errors
Language for representing biochemical structures with a
favourable performance/expressivity trade-off
4

35. C LASSIFYING S TRUCTURED O BJECTS
5

36. C LASSIFYING S TRUCTURED O BJECTS
ascorbicAcid : 0
o
6
o c o
o
5 c 11 c 1 c
hasAtom h
2
12 10 7
single c c
13
double 9 8
4 o 3 o
5

37. C LASSIFYING S TRUCTURED O BJECTS
ascorbicAcid : 0
o
6
o c o
o
5 c 11 c 1 c
hasAtom h
2
12 10 7
single c c
13
double 9 8
4 o 3 o
ascorbicAcid(x) →hasAtom(x, f1 (x)) ∧ . . . ∧ hasAtom(x, f13 (x))
o(f1 (x)) ∧ . . . ∧ c(f7 (x)) ∧ . . . ∧
single(f1 (x), f7 (x)) ∧ double(f7 (x), f2 (x)) ∧ . . .
5

38. C LASSIFYING S TRUCTURED O BJECTS
ascorbicAcid : 0
o
6
o c o
o
5 c 11 c 1 c
hasAtom h
2
12 10 7
single c c
13
double 9 8
4 o 3 o
ascorbicAcid(x) →hasAtom(x, f1 (x)) ∧ . . . ∧ hasAtom(x, f13 (x))
o(f1 (x)) ∧ . . . ∧ c(f7 (x)) ∧ . . . ∧
single(f1 (x), f7 (x)) ∧ double(f7 (x), f2 (x)) ∧ . . .
hasAtom(x, y1 ) ∧ hasAtom(x, y2 ) ∧ y1 = y2 → polyatomicEntity(x)
∧5 hasAtom(x, yi ) ∧ c(y1 ) ∧ o(y2 ) ∧ o(y3 )∧
i=1
c(y4 ) ∧ horc(y5 ) ∧ double(y1 , y2 )∧
single(y1 , y3 ) ∧ single(y3 , y4 ) ∧ single(y1 , y5 ) → carboxylicEster(x)
5

39. C LASSIFYING S TRUCTURED O BJECTS
ascorbicAcid : 0
o
6
o c o
o
5 c 11 c 1 c
hasAtom h
2
12 10 7
single c c
13
double 9 8
4 o 3 o
Input fact: ascorbicAcid(a)
Stable model: ascorbicAcid(a), hasAtom(a, af ) for 1 ≤ i ≤ 13,
i
o(af ) for 1 ≤ i ≤ 6, c(af ) for 7 ≤ i ≤ 12, h(af ), single(af , af ),
i i 13 8 3
single(af , af ), single(af , af ) for i ∈ {5, 11}, single(af , af ),
9 4 12 i 11 6
single(af , af ) for i ∈ {1, 9, 11, 13}, single(af , af ) for i ∈ {1, 8},
10 i 7 i
double(af , af ), double(af , af ), horc(af ) for 7 ≤ i ≤ 13,
2 7 8 9 i
polyatomicEntity(a), carboxylicEster(a), cyclic(a)
5

40. C LASSIFYING S TRUCTURED O BJECTS
ascorbicAcid : 0
o
6
o c o
o
5 c 11 c 1 c
hasAtom h
2
12 10 7
single c c
13
double 9 8
4 o 3 o
Input fact: ascorbicAcid(a)
Stable model: ascorbicAcid(a), hasAtom(a, af ) for 1 ≤ i ≤ 13,
i
o(af ) for 1 ≤ i ≤ 6, c(af ) for 7 ≤ i ≤ 12, h(af ), single(af , af ),
i i 13 8 3
single(af , af ), single(af , af ) for i ∈ {5, 11}, single(af , af ),
9 4 12 i 11 6
single(af , af ) for i ∈ {1, 9, 11, 13}, single(af , af ) for i ∈ {1, 8},
10 i 7 i
double(af , af ), double(af , af ), horc(af ) for 7 ≤ i ≤ 13,
2 7 8 9 i
polyatomicEntity(a), carboxylicEster(a), cyclic(a)
Ascorbic acid is a cyclic polyatomic entity and a carboxylic ester
5

41. C HEMICAL C LASSES W E C OVERED
1 Existence of subcomponents
6

42. C HEMICAL C LASSES W E C OVERED
1 Existence of subcomponents
Carbon molecules
6

43. C HEMICAL C LASSES W E C OVERED
1 Existence of subcomponents
Carbon molecules
Carboxylic acids and carboxylic esters
6

44. C HEMICAL C LASSES W E C OVERED
1 Existence of subcomponents
Carbon molecules
Carboxylic acids and carboxylic esters
Ketones and aldehydes
6

45. C HEMICAL C LASSES W E C OVERED
1 Existence of subcomponents
Carbon molecules
Carboxylic acids and carboxylic esters
Ketones and aldehydes
2 Exact cardinality of parts
6

46. C HEMICAL C LASSES W E C OVERED
1 Existence of subcomponents
Carbon molecules
Carboxylic acids and carboxylic esters
Ketones and aldehydes
2 Exact cardinality of parts
Exactly two carbons
6

47. C HEMICAL C LASSES W E C OVERED
1 Existence of subcomponents
Carbon molecules
Carboxylic acids and carboxylic esters
Ketones and aldehydes
2 Exact cardinality of parts
Exactly two carbons
Dicarboxylic acid
6

48. C HEMICAL C LASSES W E C OVERED
1 Existence of subcomponents
Carbon molecules
Carboxylic acids and carboxylic esters
Ketones and aldehydes
2 Exact cardinality of parts
Exactly two carbons
Dicarboxylic acid
3 Exclusive composition
6

49. C HEMICAL C LASSES W E C OVERED
1 Existence of subcomponents
Carbon molecules
Carboxylic acids and carboxylic esters
Ketones and aldehydes
2 Exact cardinality of parts
Exactly two carbons
Dicarboxylic acid
3 Exclusive composition
Inorganic molecules
6

50. C HEMICAL C LASSES W E C OVERED
1 Existence of subcomponents
Carbon molecules
Carboxylic acids and carboxylic esters
Ketones and aldehydes
2 Exact cardinality of parts
Exactly two carbons
Dicarboxylic acid
3 Exclusive composition
Inorganic molecules
Hydrocarbons
6

51. C HEMICAL C LASSES W E C OVERED
1 Existence of subcomponents
Carbon molecules
Carboxylic acids and carboxylic esters
Ketones and aldehydes
2 Exact cardinality of parts
Exactly two carbons
Dicarboxylic acid
3 Exclusive composition
Inorganic molecules
Hydrocarbons
Saturated molecules
6

52. C HEMICAL C LASSES W E C OVERED
1 Existence of subcomponents
Carbon molecules
Carboxylic acids and carboxylic esters
Ketones and aldehydes
2 Exact cardinality of parts
Exactly two carbons
Dicarboxylic acid
3 Exclusive composition
Inorganic molecules
Hydrocarbons
Saturated molecules
4 Cyclicity-related classes
6

53. C HEMICAL C LASSES W E C OVERED
1 Existence of subcomponents
Carbon molecules
Carboxylic acids and carboxylic esters
Ketones and aldehydes
2 Exact cardinality of parts
Exactly two carbons
Dicarboxylic acid
3 Exclusive composition
Inorganic molecules
Hydrocarbons
Saturated molecules
4 Cyclicity-related classes
Benzenes
6

54. C HEMICAL C LASSES W E C OVERED
1 Existence of subcomponents
Carbon molecules
Carboxylic acids and carboxylic esters
Ketones and aldehydes
2 Exact cardinality of parts
Exactly two carbons
Dicarboxylic acid
3 Exclusive composition
Inorganic molecules
Hydrocarbons
Saturated molecules
4 Cyclicity-related classes
Benzenes
Cyclic molecules
6

55. C HEMICAL C LASSES W E C OVERED
1 Existence of subcomponents
Carbon molecules
Carboxylic acids and carboxylic esters
Ketones and aldehydes
2 Exact cardinality of parts
Exactly two carbons
Dicarboxylic acid
3 Exclusive composition
Inorganic molecules
Hydrocarbons
Saturated molecules
4 Cyclicity-related classes
Benzenes
Cyclic molecules
Alkanes
6

56. E MPIRICAL E VALUATION
Draws upon DLV, a deductive databases engine
7

57. E MPIRICAL E VALUATION
Draws upon DLV, a deductive databases engine
Evaluation with data extracted from ChEBI
7

58. E MPIRICAL E VALUATION
Draws upon DLV, a deductive databases engine
Evaluation with data extracted from ChEBI
500 molecules under 51 chemical classes in 40 secs
7

59. E MPIRICAL E VALUATION
Draws upon DLV, a deductive databases engine
Evaluation with data extracted from ChEBI
500 molecules under 51 chemical classes in 40 secs
Quicker than other approaches:
7

60. E MPIRICAL E VALUATION
Draws upon DLV, a deductive databases engine
Evaluation with data extracted from ChEBI
500 molecules under 51 chemical classes in 40 secs
Quicker than other approaches:
[Hastings et al., 2010] 140 molecules in 4 hours
[Magka et al., 2012] 70 molecules in 450 secs
7

61. E MPIRICAL E VALUATION
Draws upon DLV, a deductive databases engine
Evaluation with data extracted from ChEBI
500 molecules under 51 chemical classes in 40 secs
Quicker than other approaches:
[Hastings et al., 2010] 140 molecules in 4 hours
[Magka et al., 2012] 70 molecules in 450 secs
Subsumptions exposed by our prototype:
7

62. E MPIRICAL E VALUATION
Draws upon DLV, a deductive databases engine
Evaluation with data extracted from ChEBI
500 molecules under 51 chemical classes in 40 secs
Quicker than other approaches:
[Hastings et al., 2010] 140 molecules in 4 hours
[Magka et al., 2012] 70 molecules in 450 secs
Subsumptions exposed by our prototype:
ascorbic acid is a polyatomic entity, a carboxylic ester and a
cyclic molecule
missing from the ChEBI OWL ontology
7

63. E MPIRICAL E VALUATION
Draws upon DLV, a deductive databases engine
Evaluation with data extracted from ChEBI
500 molecules under 51 chemical classes in 40 secs
Quicker than other approaches:
[Hastings et al., 2010] 140 molecules in 4 hours
[Magka et al., 2012] 70 molecules in 450 secs
Subsumptions exposed by our prototype:
ascorbic acid is a polyatomic entity, a carboxylic ester and a
cyclic molecule
missing from the ChEBI OWL ontology
Contradictory subclass relation from ChEBI:
7

64. E MPIRICAL E VALUATION
Draws upon DLV, a deductive databases engine
Evaluation with data extracted from ChEBI
500 molecules under 51 chemical classes in 40 secs
Quicker than other approaches:
[Hastings et al., 2010] 140 molecules in 4 hours
[Magka et al., 2012] 70 molecules in 450 secs
Subsumptions exposed by our prototype:
ascorbic acid is a polyatomic entity, a carboxylic ester and a
cyclic molecule
missing from the ChEBI OWL ontology
Contradictory subclass relation from ChEBI:
Ascorbic acid is asserted to be a carboxylic acid (release 95)
Not listed among the subsumptions derived by our prototype
7

65. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
8

66. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
2 Wide range of structure-based classes
8

67. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
2 Wide range of structure-based classes
3 DLV-based implementation exhibits a significant speedup
8

68. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
2 Wide range of structure-based classes
3 DLV-based implementation exhibits a significant speedup
4 Evaluation over ChEBI ontology revealed modelling errors
8

69. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
2 Wide range of structure-based classes
3 DLV-based implementation exhibits a significant speedup
4 Evaluation over ChEBI ontology revealed modelling errors
Language for representing biochemical structures with a
favourable performance/expressivity trade-off
8

70. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
2 Wide range of structure-based classes
3 DLV-based implementation exhibits a significant speedup
4 Evaluation over ChEBI ontology revealed modelling errors
Language for representing biochemical structures with a
favourable performance/expressivity trade-off
Future directions
SMILES-based surface syntax
8

71. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
2 Wide range of structure-based classes
3 DLV-based implementation exhibits a significant speedup
4 Evaluation over ChEBI ontology revealed modelling errors
Language for representing biochemical structures with a
favourable performance/expressivity trade-off
Future directions
SMILES-based surface syntax
∧5 hasAtom(x, yi ) ∧ c(y1 ) ∧ o(y2 ) ∧ o(y3 ) ∧ c(y4 )∧
i=1
double(y1 , y2 ) ∧ single(y1 , y3 ) ∧ single(y3 , y4 ) ∧ single(y1 , y5 )
→ carboxylicEster(x)
8

72. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
2 Wide range of structure-based classes
3 DLV-based implementation exhibits a significant speedup
4 Evaluation over ChEBI ontology revealed modelling errors
Language for representing biochemical structures with a
favourable performance/expressivity trade-off
Future directions
SMILES-based surface syntax
define carboxylicEster
some hasAtom SMILES(COC(= O)[∗])
end.
8

73. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
2 Wide range of structure-based classes
3 DLV-based implementation exhibits a significant speedup
4 Evaluation over ChEBI ontology revealed modelling errors
Language for representing biochemical structures with a
favourable performance/expressivity trade-off
Future directions
SMILES-based surface syntax
Detect subsumptions between classes
8

74. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
2 Wide range of structure-based classes
3 DLV-based implementation exhibits a significant speedup
4 Evaluation over ChEBI ontology revealed modelling errors
Language for representing biochemical structures with a
favourable performance/expressivity trade-off
Future directions
SMILES-based surface syntax
Detect subsumptions between classes
E.g., Carboxylic ester is an organic molecular entity
8

75. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
2 Wide range of structure-based classes
3 DLV-based implementation exhibits a significant speedup
4 Evaluation over ChEBI ontology revealed modelling errors
Language for representing biochemical structures with a
favourable performance/expressivity trade-off
Future directions
SMILES-based surface syntax
Detect subsumptions between classes
Extensions with numerical datatypes
8

76. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
2 Wide range of structure-based classes
3 DLV-based implementation exhibits a significant speedup
4 Evaluation over ChEBI ontology revealed modelling errors
Language for representing biochemical structures with a
favourable performance/expressivity trade-off
Future directions
SMILES-based surface syntax
Detect subsumptions between classes
Extensions with numerical datatypes
E.g., Small molecules if they weigh less than 800 daltons
8

77. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
2 Wide range of structure-based classes
3 DLV-based implementation exhibits a significant speedup
4 Evaluation over ChEBI ontology revealed modelling errors
Language for representing biochemical structures with a
favourable performance/expressivity trade-off
Future directions
SMILES-based surface syntax
Detect subsumptions between classes
Extensions with numerical datatypes
Classification of complex biological objects
8

78. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
2 Wide range of structure-based classes
3 DLV-based implementation exhibits a significant speedup
4 Evaluation over ChEBI ontology revealed modelling errors
Language for representing biochemical structures with a
favourable performance/expressivity trade-off
Future directions
SMILES-based surface syntax
Detect subsumptions between classes
Extensions with numerical datatypes
Classification of complex biological objects
Integration with Protégé, Bioclipse, JChemPaint,. . .
8

79. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
2 Wide range of structure-based classes
3 DLV-based implementation exhibits a significant speedup
4 Evaluation over ChEBI ontology revealed modelling errors
Language for representing biochemical structures with a
favourable performance/expressivity trade-off
Future directions
SMILES-based surface syntax
Detect subsumptions between classes
Extensions with numerical datatypes
Classification of complex biological objects
Integration with Protégé, Bioclipse, JChemPaint,. . .
Mapping from our logic to RDF
8

80. C ONCLUSION AND F URTHER R ESEARCH
Results
1 Expressive and decidable formalism for complex objects
2 Wide range of structure-based classes
3 DLV-based implementation exhibits a significant speedup
4 Evaluation over ChEBI ontology revealed modelling errors
Language for representing biochemical structures with a
favourable performance/expressivity trade-off
Future directions
SMILES-based surface syntax
Detect subsumptions between classes
Extensions with numerical datatypes
Classification of complex biological objects
Integration with Protégé, Bioclipse, JChemPaint,. . .
Mapping from our logic to RDF
Thank you! Questions?!?
8