A knowledge based approach to matching human neurodegenerative disease and associated animal models
-
1
University of California, San Diego, United States
-
2
Lawrence Berkeley National Laboratory, United States
Many ontologies have been developed covering domains such as anatomy, cell types, and molecular processes. Ontologies for complex entities such as disease have been more difficult to define, but are critical to our pursuit of the subtle relation between clinical disease findings and observations made in model systems. Neurodegenerative diseases have a wide and complex range of biological and clinical features. While diseases share pathological features, they have unique signatures, particularly in targeting cells and subcellular structures. Animal models are key to translational research, yet typically only replicate a subset of indirectly related disease features. In addition, the pathologies occur across multiple spatial and temporal scales, and are expressed using varied vocabularies. Thus, data mining approaches for comparing animal to human conditions has proven challenging.
We take a phenotype-based approach to developing a multi-scale ontology for neurodegenerative disease and model systems, and thereby facilitate comparisons between neurodegenerative disease and model systems. We are defining phenotype to include any observable or measurable feature associated with organism, and, due to the nervous system’s complexity, we require knowledge represented in ontologies to bridge the multiple structural scales and neuroanatomical systems in which alterations occur. We constructed a flexible, formal template for structural phenotypes that is amenable to computational logical inference. Rather than a complete specification of a disease process, we focus on measured phenotypes observed in organisms. Humans are treated the same as model systems, yet defined as bearing a disease. Our template draws from the Ontology of Phenotypic Qualities (http://purl.org/obo/owl/PATO) and Neuroscience Information Framework ontologies (NIFSTD; http://purl.org/nif/ontology/nif.owl). The resulting Neurodegenerative Disease Phenotype Ontology (NDPO; http://ccdb.ucsd.edu/NDPO/1.0/NDPO.owl) is encoded in OWL and contains 700 class level phenotypes derived from reviews. The companion Phenotype Knowledge Base (PKB; http://ccdb.ucsd.edu/PKB/1.0/PKB.owl) imports NDPO and contains instances of phenotypes (human and non-human) in primary articles. We loaded our phenotypes into the Ontology-Based Database (OBD; http://berkeleybop.org/pkb), an open access database for ontology-based descriptions where phenotypes, diseases, and models are matched using logical inference and semantic similarity statistical metrics.
The OBD interface performs queries such as “Find organisms containing cellular inclusions”, using NIFSTD definitions to connect entities in clinical descriptions of human disease to models, e.g. Lewy body and cellular inclusions. We use OBD to perform similarity comparisons across models and human disease at the level of single phenotypes, e.g., find organisms with aggregated alpha synuclein. Knowledge in the ontology provides phenotype common subsumers. For example, a human with Parkinson’s diseas with phenotype “midline nuclear group has extra parts Lewy Body” matches an animal overexpressing alpha synuclein with phenotype “paracentral nucleus has extra parts cellular inclusion” with their common subsumer: Thalamus has extra parts cellular inclusion. OBD uses information content measures to compare aggregate phenotypes to find overall best matches between organisms. Using these queries, we have performed a detailed comparison of organisms and disease related phenotypes.
By using a consistent phenotype model referenced to well-structured ontologies with defined classes, we can aggregate and bridge phenotypes made in animal models from basic research and descriptions of pathological features from clinical preparations. This provides the steps toward a temporal specification of the disease process. We continue to enrich the knowledge base and representations to explore different statistical methods for enhancing the relevancy of the matches.
References
1. Gupta A, Ludascher B, Grethe JS, Martone ME (2003) Towards a formalization of disease-specific ontologies for neu-roinformatics. Neural Networks 16:1277-1292.
2. Washington NL, Haendel MA, Mungall CJ, Ashburner M, Westerfield M, Lewis SE (2009) Linking human diseases to animal models using ontology-based phenotype annotation. PLoS Biol 7(11):e1000247.
Conference:
Neuroinformatics 2010 , Kobe, Japan, 30 Aug - 1 Sep, 2010.
Presentation Type:
Poster Presentation
Topic:
General neuroinformatics
Citation:
Maynard
SM,
Mungall
CJ,
Lewis
SE and
Martone
ME
(2010). A knowledge based approach to matching human neurodegenerative disease and associated animal models.
Front. Neurosci.
Conference Abstract:
Neuroinformatics 2010 .
doi: 10.3389/conf.fnins.2010.13.00054
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
11 Jun 2010;
Published Online:
11 Jun 2010.
*
Correspondence:
Sarah M Maynard, University of California, San Diego, La Jolla, United States, sarahmarkels@gmail.com