Ontology and RadLex

1
Ontology and RadLex
Robert Arp, Ph.D. Ontology Research Group
(ORG) www.org.buffalo.edu National Center for
Biomedical Ontology (NCBO) www.bioontology.org Th
is work was funded by the National Institutes of
Health through the NIH Roadmap for Medical
Research, Grant 1 U 54 HG004028. Information on
the National Centers for Biomedical Computing can
be found at http//nihroadmap.nih.gov/bioinformati
cs.
2
Problem radiologist wondering if radiation
therapy in chest is linked to Cauda Equina
Syndrome in lower back. she needs data and
information to find a solution - calls
colleagues - sends emails - queries RadLex
3
Problem radiologist wondering if radiation
therapy in chest is linked to Cauda Equina
Syndrome in lower back. She needs data and
information to find a solution - calls
colleagues - sends emails - queries RadLex
4
RadLexA Lexicon for Uniform Indexing and
Retrieval of Radiology Information Resources
5
RadLexA Lexicon for Uniform Indexing and
Retrieval of Radiology Information Resources
direct result of the information age and
information science
6
Informaticsthe science of data and information
collection, categorization, management, storage,
processing, retrieval, and dissemination.
7
Informaticsthe science of data and information
collection, categorization, management, storage,
processing, retrieval, and dissemination.
retrieval and dissemination the goal of
information science
8
Informaticsthe science of data and information
collection, categorization, management, storage,
processing, retrieval, and dissemination.
but goal depends upon accurate categorization and
management
9
Computers are dumb beasts At this point, they
can mostly only accurately output what we
accurately input. Not ruling out the
possibility of human-like thinking in the future
10
So, a central informatics problem that hampers
retrieval and dissemination is incorrect
thinking (IT) associated with categorization and
management of data and information
11
Cant get the data and/or info - radiologist
gives up the search in frustration- cure is not
found- suffering is not relieved moral
obligation to retrieve and disseminate data and
information?
12
To combat incorrect thinking (IT) and assist in
the accurate categorization and management of
data and information so that retrieval and
dissemination is maximally possible ontologies
are used
13

• What is an ontology?
• Philosophical Ontology
• Domain Ontology
• Formal Ontology

14

• Philosophical Ontology
• ...I can fit wholesale evolution and a creating
  god into my ontology without contradiction.
• ...just because it has mental existence doesnt
  mean it has ontological existence.

15

• - Ontos (being, existence) Logos (word, account,
  explanation)
• - The study of what is, of the kinds and
  structures of objects, properties, events,
  processes, and relations in every area of reality
• - The branch of Metaphysics that studies the
  nature of existence. Random House College
  Dictionary

16
(No Transcript)
17
PORPHYRIAN TREE
Compare Linnean Taxonomy and Periodic Table
18
Sub-types (children) should inherit all the
features of the relevant type (parent)
PORPHYRIAN TREE
Compare Linnean Taxonomy and Periodic Table
19
Sub-types (children) should inherit all the
features of the relevant type (parent)
Sub-types (children) should inherit all the
features of the relevant type (parent) along with
some differentia (distinguishing feature)
A is_a B that/which Cs A is_a B that/which has
Cs - rational animal (A) is_a living entity with
sensation (B) that has the capacity for conscious
thinking (C)
PORPHYRIAN TREE
Compare Linnean Taxonomy and Periodic Table
20
A is_a B that/which Cs A is_a B that/which has Cs

• radiology (A) is_a science (B) that deals with
  medical imaging so as to understand, diagnose
  (C)
• domain ontology is_a ontology that deals with a
  specific area, sphere, or portion of reality
• - mouse trap is_a rodent trap which captures mice

21
To a certain extent, all of us are Philosophical
Ontologists in that we naturally and
automatically categorize any and all things in
reality so as to understand, explain, control,
dominate, and navigate reality.
22

• What is an ontology?
• Philosophical Ontology
• Domain Ontology
• Formal Ontology

23
(2) Domain Ontology ...Im working on an
ontology for annelids. ...the Gene Ontology
has data on that HOX gene.
24

• Representation of the entities and relations
  existing within a particular domain of reality
  such as biology, medicine, geography, ecology, or
  law
• - Gene Ontology (GO)
• - Foundational Model of Anatomy (FMA)
• Environment Ontology (EnvO)
• See http//www.obofoundry.org/

25
- Opposed to ontology in the philosophical sense,
which has all of reality as its subject matter -
Ideally, provides a controlled, structured
vocabulary to annotate data in order to make it
more easily searchable by human beings and
processable by computers
26
ONTOLOGY a representational artifact,
comprising a taxonomy as its main part, whose
representational units are intended to designate
some combination of universals, defined classes,
and certain relations between them. Smith,
B., Kusnierczyk, W., Schober, D., Ceusters, W.
(2006). Towards a reference terminology for
ontology research and development in the
biomedical domain. Proceedings of KR-MED 2006, 1,
1-14.
27
Domain ontology contrasted with- Database-
Rule-Based Language - Thesaurus- Glossary-
Catalogue- Inventory- Axiomatic Theory- Simple
Taxonomy

• Ontology hybrid of
• - Taxonomy
• Various Relations
• Axiomatic Rules of Inference

28
Domain ontology contrasted with- Database-
Rule-Based Language - Thesaurus- Glossary-
Catalogue- Inventory- Axiomatic Theory- Simple
Taxonomy
RadLex, as of now, is a thesaurus of terms,
although confuses terms and entities those terms
refer to
29
Basic Mouse Trap Classification
Simple Taxonomy is_a, sub-type relation
30
Beginnings of Mouse Trap Domain Ontology
31
Beginnings of MRI Test Ontology
32
A Gene Ontology Example Cytokinesis
33
is_a - part_of -
A Gene Ontology Example
34
Part of a Scientific Experiment Ontology
35
Part of a Scientific Experiment Ontology
p/o part_of
? is_a
36
Now, its fine to have all of these domain
ontologies attempting to classify and categorize
data and informationBut, there is so much data
and information
37
(No Transcript)
38
(No Transcript)
39
(No Transcript)
40
(No Transcript)
41
(No Transcript)
42
(No Transcript)
43
Crazy and overwhelming, isnt it?
44
So, instead of using a STANDARD like the Periodic
Table of Elements, people start classifying
things their own different waysthis results in
45
SILO EFFECT
46
SILO EFFECT
47
PROBLEMDE-SILOING all of this domain data and
information so that it may be queried
effectively, shared, and re-usedINTEROPERABILITY
48
PROBLEMDE-SILOING all of this domain data and
information so that it may be queried
effectively, shared, and re-usedINTEROPERABILITY
49
SOLUTION
50
FORMAL ONTOLOGY- Upper-level- Applicable to
any domain
51

• What is an ontology?
• Philosophical Ontology
• Domain Ontology
• Formal Ontology

52
Assists in making communication between and among
domain ontologies possible by providing -
Common language- Common formal framework for
reasoning
53
So, just as ENGLISH is a common world language
enabling different people to communicate in
common So too, FORMAL ONTOLOGY is a common
language enabling different domain ontologies
to communicate in common
54
Formal ontology is like a backbone or spine
making communication, interoperability, and
optimal dissemination of data and information
possible between and among domain ontologies
55
Formal ontology is like a backbone or spine
making communication, interoperability, and
optimal dissemination of data and information
possible between and among domain ontologies
56
From this To this
57
From this To this
From this To this
58
Basic Formal Ontology (BFO)
59
REALISM-BASED ONTOLOGY Universals (1) real
objects, substances, endurants, or
continuants - SNAP shots of reality (2) real
processes, activities, perdurants, or
occurrents - SPAN of time Relations is_a,
part_of, has_participant...
60
continuants vs. occurrents
continuant (substance, object)

• In classifying parts of reality, we keep track
  of these two different kinds of entities in two
  different ways

61

• continuant entities
• - have continuous existence in time
• - preserve their identity through change
• - exist in toto, if they exist at all
• occurrent entities
• - have temporal parts
• - unfold themselves phase by phase
• - exist only in their phases/stages

62

• continuant entities
• - have continuous existence in time
• - preserve their identity through change
• - exist in toto, if they exist at all
• occurrent entities
• - have temporal parts
• - unfold themselves phase by phase
• - exist only in their phases/stages

objects, qualities, functions
processes, activities
63
continuant
Relation is_a
Sub-types (children) should inherit all the
features of the relevant type (parent)
64
HUMAN HEART
continuant
human heart
surface of the heart
pink, smooth
all hearts in this room
stops if no circulation
a biopsy of the heart
pumps blood
chest cavity
prop in a display
65
occurrent
Relation is_a
66
occurrent
ECG/EKG TEST
ECG/EKG TEST
ECG (EKG) test
start/end of ECG
s/t ECG began
moment ECG began
all ECGs in clinic
s/t region of ECG
time occupied
2nd lead attached
activities in clinic
67

1. Philosophical Ontology
2. Domain Ontology
3. Formal Ontology

68
Basic principles from philosophical ontology and
formal ontology can assist a domain like
radiology and a domain ontology like RadLex
(although, RadLex is at the beginning stages of a
domain ontology)
69
Fundamental Step Clear up the incorrect
thinking (IT) in Radlex
70
Incorrect Thinking (IT)
- simply getting the facts wrong-
perception/reality confusions- using examples
instead of definitions when defining
something- circular definitions- equivocation
of terms- use/mention confusions- unclear or
incoherent definitions
71
Incorrect Thinking (IT)

• these are just a few
• many more problems
• problems are legion in databases
• well look at just a few problems in RadLex

72
(No Transcript)
73
IT simply getting the facts wrong True, these
are all terms, but the terms are supposed to be
referring to the actual entities in reality
confusing classification of terms with a taxonomy
of things
74
IT simply getting the facts wrong Is a
substance a subtype of term? Is an imaging
procedure attribute a sub-type of term? None of
these are terms, they are things.
75

• IT unclear or incoherent definitions
• imaging procedure attribute
• is not really a child/sub-type of RadLex term and
• (b) it is the only child when, on this account,
  treatment, substance, etc., should be subtypes as
  well

76
IT use/mention confusion At the same time these
terms are being used as a taxonomy that refers to
reality, and mentioned in a lexical
classification. Terms are representations that
refer to things, but are being treated as the
things themselves confusion!
77
Where is the definition of a RadLex term? Where
are some examples? How am I supposed to know what
these things are?
?
78
Is a biopsy an imaging purpose? Is an infusion a
purpose? Is an imaging purpose an imaging
procedure attribute? Where are the definitions?
79
Is a biopsy a RadLex term? Is an infusion a
RadLex term?
Sub-types (children) should inherit all the
features of the relevant type (parent)
80
IT using examples instead of definitions when
defining something - I still dont know what an
imaging service request is. - Also, isnt a
request something different from exams? - I know
what an example of it might be
81
Also, isnt a request something different from an
exam or exams? Compare Drinking service
request Def. The drinks that can be ordered by
patrons at a bar
A synonym is a RadLex child, too?
82
A foreign body is a child of finding? A missile,
by definition, is a foreign body? I can see it
playing a role as a foreign body, but its essence
is to be a foreign body? Same with personal
item, surgical implement, and others
83
Three Levels to Keep Straight

• Level 1 The entities in reality such as
  universals and relations
• Level 2 Cognitive representations of this
  reality on the part of scientists
• Level 3 Publicly accessible concretizations of
  these cognitive representations in textual,
  graphical, or computational representational
  artifacts (like RadLex composed of terms)

84
Three Levels to Keep Straight

• Cognitive representations
• Representational artifacts
• Reality

85
- Bertaud, V., Belhadj, I., Dameron, O.,
Garcelon, N., Hendaoui, L., Marin, F.,
Duvauferrier, R. (2007). Computerizing the
radiological sign. Journal of Radiology, 88,
27-37.- Fielding, J., Marwede, D. (2006). Four
ontological models for radiological diagnostics.
Studies in Health Technology and Informatics,
124, 761-766.- Kahn, C., Channin, D., Rubin,
D. (2006). An ontology for PACS integration.
Journal of Digital Imaging, 19, 316-327.-
Marwede, D., Fielding, M. (2005). The
epistemological-ontological divide in clinical
radiology. Studies in Health Technology and
Informatics, 116, 749-754.- Pommert, A., Höhne,
K., Pflesser, B., Richter, E., Riemer, M.,
Schiemann, T., Schubert, R., Schumacher, U.,
Tiede, U. (2001). Creating a high-resolution
spatial/symbolic model of the inner organs based
on the visible human. Medical Image Analysis, 5,
221-228.- Rubin, D. (2007). Creating and
curating a terminology for radiology Ontology
modeling and analysis. Journal of Digital
Imaging. Available at http//www.springerlink.com
/content/ 978708n776738132/fulltext.pdf.- Rubin,
D., Dameron, O., Bashir, Y., Grossman, D., Dev,
P., Musen, M. (2006). Using ontologies linked
with geometric models to reason about penetrating
injuries. Artificial Intelligence in Medicine,
37, 167-176.
Radiology and Ontology Literature
86
Arp, R., Romagnoli, C., Chhem, R., Overton, J.
(2008). Radiological and Biomedical Knowledge
Integration The Ontological Way. In R. Chhem,
K. Hibbert, T. Van Deven (Eds.), Radiology
Education (Chapter 8, pp. 87-104). Berlin
Springer-Verlag.
87
Biomedical Informatics Literature

• Baxevanis, A., Ouellette, B. (2005).
  Bioinformatics A practical guide to the analysis
  of genes and proteins. Hoboken, NJ Wiley.
• Berman, J. (2006). Biomedical informatics.
  London Jones Bartlett Publishers.
• Chen, H., Fuller, S., Friedman, C., Hersh, W.
  (2005). Medical informatics Knowledge management
  and data mining in biomedicine. The Netherlands
  Springer.
• Goldstein, D., Groen, P., Ponkshe, S., Wine, M.
  (2007). Medical informatics 20/20 Quality and
  electronic health records through collaboration,
  open solutions, and innovation. New York Jones
  Bartlett.
• Polanski, A., Kimmel, M. (2007).
  Bioinformatics. London Springer.
• Shortliffe, E., Cimino, J. (Eds.). (2006).
  Biomedical informatics Computer applications in
  health care and biomedicine. London Springer.
• van Bemmel, J., Musen, M. (Eds.). (1997).
  Handbook of medical informatics. The Netherlands
  Springer.
• Xiong, J. (2006). Essential bioinformatics.
  Cambridge Cambridge University Press.

88
Domain and Formal Ontology Literature

• Arp, R. (2007). Philosophical ontology, domain
  ontology, formal ontology. The Reasoner, 1,
  12-13.
• Bittner, T., Donnelly, M., Winter, S. (2006).
  Ontology and semantic operability. In S.
  Zlatanova D. Prosperi (Eds.), Large-scale 3D
  data integration Challenges and opportunities
  (pp. 139-160). Boca Raton, FL CRC Press.
• Ceusters, W., Smith, B., van Mol, M. (2003).
  Using ontology in query answering systems
  Scenarios, requirements and challenges.
  Proceedings of the 2nd CoLogNET-ElsNET Symposium,
  Amsterdam, 2, 5-15.
• Grenon, P., Smith, B. (2004). SNAP and SPAN
  Towards dynamic spatial ontology. Spatial
  Cognition and Computation, 1, 1-10.
• Grenon, P., Smith, B. (2004a). A formal theory
  of substances, qualities and universals. In A.
  Varzi and L. Vieu (Eds.), Proceedings of FOIS
  2004. International Conference on Formal
  Ontology and Information Systems (pp. 49-59).
  Amsterdam IOS Press.
• Mars, N. (Ed.). (1995). Towards very large
  knowledge bases Knowledge building and knowledge
  sharing. Amsterdam IOS Press.
• Menzel, C. (2003). Ontology theory. In J.
  Euzenat, A. Gomez-Perez, N. Guarino, H.
  Stuckenschmidt (Eds.), Ontologies and semantic
  interoperability (pp. 13-30). Hamburg IOS Press.
• Smith, B. (2003). Ontology. In L. Floridi (Ed.),
  Blackwell guide to the philosophy of computing
  and information (pp. 155-166). Malden, MA
  Blackwell.
• Smith, B., Ceusters, W. (2007). Ontology as the
  core discipline of biomedical informatics
  Legacies of the past and recommendations for the
  future direction of research. In G. Crnkovic S.
  Stuart (Eds.), Computing, philosophy, and
  cognitive science (pp. 121-145). Cambridge
  Cambridge Scholars Press.
• Smith, B., Kumar, A., Bittner, T. (2004). Basic
  Formal Ontology for bioinformatics. Available at
  http//www.uni-leipzig.de/akumar/JAIS.pdf.

89
Thank You
Robert Arp, Ph.D. Ontology Research Group
(ORG) www.org.buffalo.edu National Center for
Biomedical Ontology (NCBO) www.bioontology.org Th
is work was funded by the National Institutes of
Health through the NIH Roadmap for Medical
Research, Grant 1 U 54 HG004028. Information on
the National Centers for Biomedical Computing can
be found at http//nihroadmap.nih.gov/bioinformati
cs.