Date post: | 10-May-2015 |
Category: |
Technology |
Upload: | michel-dumontier |
View: | 331 times |
Download: | 0 times |
Molecular symmetry and specialization of atomic connectivity by class-based
reasoning of chemical structure
1
Michel Dumontier, Ph.D.
Associate Professor of Bioinformatics Department of Biology, School of Computer Science, Institute of Biochemistry, Carleton
University Ottawa Institute of Systems Biology
Ottawa-Carleton Institute of Biomedical Engineering Professeur Associé, Université Laval
OWLED2012::Dumontier
chemical structure: molecules consist of atoms connected by bonds
caffeine
single bond
double bond
Carbon atom
Hydrogen atom
Nitrogen atom Oxygen atom
OWLED2012::Dumontier 2
First attempt: class-based representation of chemical functional groups
HydroxylGroup equivalentTo: CarbonGroup that (hasSingleBondWith some ( OxygenAtom that hasSingleBondWith some HydrogenAtom))
OWLED2012::Dumontier 3
Describing chemical functional groups in OWL-DL for the classification of chemical compounds. Natalia Villanueva-Rosales and Michel Dumontier. OWL: Experiences and Directions (OWLED 2007).
automatic classification of chemical functional groups
28 OC
OWLED2012::Dumontier 4
Problems
1. Descriptions started at an arbitrary central atom, so all descriptions needed to “specialize these”
2. Not possible to describe a chemical functional groups that are graph-like
e.g. contains a cycle
OWLED2012::Dumontier 5
OWL representation
We really need to represent and reason over structured objects
Without structure-based representation, all parts must be explicitly asserted
(combinatorial explosion for larger molecules)
But the structure of complex molecules breaks the OWL Tree Model requirement
does not have a model in the shape of a tree
OWLED2012::Dumontier 6
Description Graphs
• A decidable extension to OWL 2 allowing expression of complex structures as graphs within the ontology
• strong separation requirement: atomic properties used as graph edges have to be different to those used in axioms in the main OWL ontology
• Rules can be used to enhance OWL with the capacity to express if – then constructions
• Using OWL, Description Graphs and Rules we could represent and reason over (classify) chemical structures at the class level.
OWLED2012::Dumontier 7
Representing Chemicals using OWL, Description Graphs and Rules. J Hastings, M Dumontier, D Hull, M Horridge, C Steinbeck, U Sattler, R Stevens, T Horne, and K Britz. OWLED 2010.
OWL + DG + Rules = Chemical Classification
OWLED2012::Dumontier 8
Before After
So, what can we do with just OWL?
• generate connectivity descriptions for every atom to every other atom – overcome the central atom problem – exponential part list
• reason at different levels of granularity – we could describe atoms in terms of 1. the types of atoms they are connected to 2. the exact set of atoms they are connected to 3. the only atoms they are connected to
OWLED2012::Dumontier 9
Dataset
A) butane, B) pentane, C) iso-butane, D) iso-pentane, E) cyclobutane and F) cyclohexane
OWLED2012::Dumontier 10
HermiT
Method
OWLED2012::Dumontier 11
Protégé 4.2
SDF
PHP-based OWLAPI
SDF2OWL
OWL
Inference
formalization
reasoning
Explanation Workbench
Formalization separates the chemical graph from the molecule
OWLED2012::Dumontier 12
`fully connected atom M` equivalentTo `atom type` and `has bond with` exactly 1 `fully connected atom N` and ...
`atom X from molecule A` equivalentTo `fully connected atom M` and `is component part of` some `molecule Y`
Symmetry
OWLED2012::Dumontier 13
equivalence among 2,3 and 4 as every peripheral atom is connected to the central atom (1)
1
4 3
2
Symmetry
• For iso-pentane, we get equivalence between atoms 4 & 5 because they are both connected to atoms 1
• we get a different relationship – one of subsumption - between atoms 2 and 4 and atoms 2 and 5
OWLED2012::Dumontier 14
3
1
4 5
2
Atomic specialization
Basically, atom 2 has a bond to atom 1, as do atoms 4 and 5, but it also has a bond to atom 3
OWLED2012::Dumontier 15
1
4 5
2 3
Symmetry in butane
• Equivalence between atoms 1 & 3 as they both share connectivity to atoms 2 & 4, and vice versa.
• No equivalence among all atoms, however.
OWLED2012::Dumontier 16
1
2
3
4
But not in cyclohexane
• No 2 atoms are connected to the same pair of atoms.
OWLED2012::Dumontier 17
Conclusion
• We investigated class-based representation where class descriptions consisted of fully qualified cardinality restrictions to other fully-connected atoms.
• We found instances of equivalence (symmetry) and specialization (additional bonding), all within a single molecule
• Next, we’ll be looking at reasoning across different molecules, but this requires some equivalence between atoms of different molecules.
OWLED2012::Dumontier 18
dumontierlab.com [email protected]
19 EBI2011::Dumontier
Website: http://dumontierlab.com Presentations: http://slideshare.com/micheldumontier