Session VI: Biomedical
Session Organizers:
- Anne Thessen, Oregon State University
- Ramona Walls, University of Arizona
- Chris Mungall, Lawrence Berkeley National Laboratory
- Pier Luigi Buttigieg, Max Planck Institute
Abstract:
Disciplines such as biodiversity science, environmental science, biomedicine and phylogenetics, need to make assertions about characteristics of entities and processes. These characteristics have been referred to as traits, phenotypes, diseases, and qualities, sometimes interchangeably or inconsistently. Each community has developed its own design patterns, classes, and properties for representing characteristics, sometimes in isolation. Clarity is needed on why and how communities of practice are representing characteristics to avoid and remove unnecessary silos. Examples of topics of interest in this session include:
- Utility of different character types in inferencing
- Strategies for representing characters and why these are appropriate
- Discipline-specific usage and modeling
- Modeling and integrating character derived from sensor data
- Aggregating and de-aggregating character data without loss of information
- Representation of Essential Biodiversity Variables (EBV)
Semantic Instance Anatomies: Phenotype Descriptions and their accompanying Metadata as Instance-Based Semantic Graphs that are organized into several Named Graphs
Lars Vogt
Currently, morphological data are still mostly published as unstructured free text descriptions, which lack semantic transparency and cannot be parsed by computers, thus hampering their re-use by non-experts and their integration across many fields in the life sciences. With an ever-increasing amount of available ontologies and the development of adequate semantic technology, however, a solution to this problem becomes available. Instead of free text descriptions, morphological data can be recorded, stored and communicated through the Web in form of highly formalized and structured directed graphs (semantic graphs) that use ontology terms and URIs as terminology. After introducing an instance-based approach of recording morphological descriptions as semantic graphs (i.e., Semantic Instance Anatomy graphs) and accompanying metadata graphs, I propose a general scheme of how to efficiently organize the resulting graphs in a tuple store framework based on instances of defined named graph ontology classes, with each individual named graph relating to a particular observation of a specific type. The use of such named graph resources allows meaningful fragmentation of the data, which in turn enables subsequent specification of all kinds of data views for managing and accessing morphological data. This scheme has been implemented in the description module of the prototype for semantic Morph∙D∙Base (https://proto.morphdbase.de/).
Modeling phenological data as observations of traits
Brian Stuckey
Observations in TaxonWorks
Matthew Yoder, University of Illinois, Illinois Natural History Survey, Species File Group
Michael Twidale, University of Illinois, School of Information Sciences - twidale@illinois.edu
Andrea Thomer, University of Michigan, School of Information - athomer@umich.edu
TaxonWorks is an open-source workbench built to facilitate biological taxonomy. Our overall goal in TaxonWorks is to faithfully represent historical methodologies and data representations while pre-adapting scientist’s work to richer semantic contexts. The very nature of the scientific work taxonomists do, and the interfaces they use, provides a context for how phenotypes (or whatever we call them) are represented, we touch on these inter-relationships. The models and approaches to representing phenotypes therein, overviewed here, are based on well over 10 years of work, and they are far from finished.
Using phenotype ontologies to incorporate domain knowledge into models of trait evolution
James P Balhoff
Comparative descriptions of how organisms vary in phenotype are core to the study of the process of evolution. Researchers use rich, descriptive free text to describe their observations on museum specimens. Free text, however, is challenging for non-experts to parse and reuse, and on a larger scale, opaque to integrative analyses by computational methods. The Phenoscape project (www.phenoscape.org) has demonstrated that by annotating phenotypes with semantic terms from ontologies, links can be made between novel species phenotypes and the candidate genes that may underlie them. We are now exploring how ontological annotations can support improvements to evolutionary analyses, by allowing the incorporation of background knowledge of anatomy and development into statistical models of evolution.
The Ontology Of Plant Stress
Marie Angélique Laporte
During this talk, we will see how the ontology has been built automatically from scraping the APS website content and using design patterns, linking plant diseases to their hosts and pathogens and to the environment they occur in.