September 21, 2017, Thursday

From BIMIB

Jump to: navigation, search

SSFW09 Presentations

Contents

Speakers' Presentations

DAY 1: Upper Ontology

Nicola Guarino: Qualities in DOLCE

Download PDF

Riichiro Mizoguchi: Reconciliation between BFO and DOLCE leads to YATO

In order to realize better quality descrition, we need two reconciliations: One between philoshopy and engineering and the other between BFO and DOLCE. Especially they are critical to obtain a better phenotype description framework. Although BFO is beautifully designed as ontology, its quality description needs some adjustment for practical use, that is, from engineering point of view. How to capture change(evolution) and how to cover various requirements for quality descriptions out there. In order to cope with these issues, I designed a reference ontology named YATO by incorporating quality in BFO, qualia in DOLCE.

Download ppt

Barry Smith: BFO on Diseases, Signs and Symptoms

Download ppt

DAY 1: OGMS (Ontology for General Medical Science)

Richard Scheuermann: Representing Influenza in OGMS

In many existing medical terminologies there is often a conflation of terms used to represent the causes of disease, the manifestations of disease, the signs we use to recognize disease in the clinic and the representation of these signs in the medical record. We have developed a framework designed to bring clarity to these distinctions – the Ontology for General Medical Science (OGMS). In OGMS we propose that a disease involves in every case some physical basis (the disorder) within the organism that bears the disposition (the disease) toward the execution of pathological processes (the realizations), which can be recognized through physical exam assessments and laboratory tests in order to arrive at a diagnosis. In this presentation we will show how OGMS can be used to represent the natural and clinical aspects of the disease arising from the infection of a susceptible human by the influenza virus.

Download ppt

William Hogan: Towards an Ontology for General Medical Science

In response to last year's meeting in Dallas and the resulting paper Toward an Ontological Treatment of Disease and Diagnosis, we in coordination with the curators of the Ontology of Biomedical Investigations have initiated the Ontology for General Medical Science (OGMS). This ontology covers the top-level terms used in medical science and clinical medicine, including disease, diagnosis, symptom, finding, and sign. The goal of OGMS is to be a small, "umbrella" ontology from which we and others can derive various domain ontologies-such as a drug ontology and a laboratory test ontology. In this talk, I will review the current status of OGMS, current issues and proposed modifications, and proposed future work.

Download pptx

DAY 2: Ontology Development and Data Integration

Sivaram Arabandi: Developing a Sleep Domain Ontology

The NCRR-funded PhysioMIMI project aims at developing a federated data integration environment supporting collaborative clinical and translational research using physiological, clinical and genomic data across institutions and databases. Sleep Domain is used as an exemplar in this project which includes any studies that combine physiological monitoring (e.g., polysomnography, ECG, EEG, etc.) with outcome, clinical predictor, and genomic data. A Sleep Domain Ontology (SDO) is being developed to provide a common framework for data sharing and for the federated query interface called VISAGE. SDO makes use of a number of upper level and reference ontologies such as BFO, RO, FMA, CPR, OGMS to provide coverage of the terms necessary for sleep research. VISAGE relies on SDO to generate queries through an interactive user interface that dynamically creates sliders for continuous variables and multiple choices for categorical variables. To connect databases to the PhysioMIMI environment, SDO is used to create Ontology to Database maps to support the translation of abstract queries generated from VISAGE to database-specific SQL queries for the retrieval of patient cohorts and related data.

Download PDF

Gianluca Colombo et Daniele Merico: Representing Cerebrovascular Phenotypes in NEUROWEB

NEUROWEB is a EU-funded project to support association studies in the cerebrovascular domain, integrating clinical data from four European excellence centers. To these ends, we delevoped a reference ontology deconstructing core cerebrovascular phenotypes from the TOAST classification into more elementary units, which were eventually mapped on local databases. The NEUROWEB Reference Ontology meta-model relies on the following criteria: 1) distinguishing diseases and patho/physiological processes from the observation of their manifestations during the diagnostic process (clinical findings), 2) distinguishing between long-term pathological conditions (such as atherosclerosis or coronary artery disease) and traumatic point-events (such as ischemic or hemorrhagic stroke), 3) identifying minimal-granularity diagnostic indicators. We finally analyze the relations between the NEUROWEB meta-model and OGMS (Ontology for General Medical Science).

Download ppt

Leonardo Lezcano: A Model for Integrating Disease and Diagnosis Semantics in Clinical Archetypes

Developing and integrating ontologies into the EHR and decision support systems has become a key priority as well as a necessary condition to reach Semantic Interoperability (SIOp) between heterogeneous healthcare systems. This presentation briefly describes the integration of the Ontology for General Medical Science (OGMS) and the OpenEHR Information Model. These models are different in the levels of abstraction so they have different types of authors, representations and purposes. However they can and they should be integrated in order to allow for SIOp. A tentative mapping and two approaches are presented to address the ambiguity and incompleteness of current archetypes when describing the initiation, realization and recognition of diseases. Besides, the integration with clinical archetypes offers several advantages as the automatic translation to OWL has already been implemented (http://code.google.com/p/ehr2ont/)

Download ppt part I

Download ppt part II

Andrew James: Structured representation of Disorders of the Newborn Infant

The structured representation of disorders of the newborn infant is challenging because a neonatal disorder may present in the embryonic, fetal or neonatal periods; be the outcome of a disorder that exists only during the embryonic and/or fetal periods; be a transient disturbance of function that is unique to the neonatal period; persist into infancy, childhood and beyond. SNOMED CT® provides good representation for disorders of the newborn infant. 9% of terms in a SNOMED RefSet for respiratory disorders of the newborn infant are used more than once per month; 23% are used at least once every six months; 60% are used less than once per year. Congenital disorders of the respiratory system accounted for 74% of the infrequently used terms. Disorders of the newborn infant are typically classified by organ system with an emphasis upon pathology. Physicians usually focus on pathophysiology as manifest by clinical features (symptoms, signs, and investigations). The simultaneous representation of disorders from both a structural and functional perspective should provide a better representation of reality than representation from either perspective alone.

Download ppt

Peter Krawitz: Clinical Diagnostics in Human Genetics with Semantic Similarity Search in Ontologies

The differential diagnostic process attempts to identify candidate diseases that best explain a set of clinical features. This process can be complicated by the fact that the features can have varying degrees of specificity, and by the presence of features unrelated to the disease itself. Depending on the experience of the physician and the availability of laboratory tests, clinical abnormalities may be described in greater or lesser detail. We have adapted semantic similarity metrics to measure phenotypic similarity between queries and hereditary diseases annotated using the Human Phenotype Ontology (HPO) and have developed a statistical model to assign P-values to the resulting similarity scores, which can be used to rank the candidate diseases. We show that our approach outperforms simpler term-matching approaches that do not take the semantic interrelationships between terms into account. The advantage of our approach was greater for queries containing phenotypic noise or imprecise clinical descriptions. The semantic network de ned by the HPO can be used to refine the differential diagnosis by suggesting clinical features that if present best differentiate among the candidate diagnoses. Semantic similarity searches in ontologies represent a novel and useful way of harnessing the semantic structure of human phenotypic abnormalities to help with the differential diagnosis. We have implemented our methods in a freely available web-application for the eld of human Mendelian disorders available at http://compbio.charite.de/phenomizer

Download ppt

Download PDF

Alan Ruttenberg: Where do these 'disease states' go?

Collaborating with the OBI, the Ontology of Biomedical Investigations, the EFO (Experimental Factor Ontology) has submitted, for inclusion in OBI, a set of terms users have filled in to the "disease state" field when annotating Microarray experiments. Terms are cross-referenced to/acquired from a number of source ontologies and terminologies, including NCI Thesaurus, the OBO Disease Ontology (DO), UMLS, Gene-RIF, ICD9 and SNOMED.

Based on trying to understand and sort these terms, I present a selection of issues that arise, discussing what challenges they bring to building an ontology of medicine based on the BFO, with attention to the current draft Ontology for General Medicine (OGMS) and Infectious Disease Ontology (IDO). I discuss, as well, the challenges to building information systems that support users in making important distinctions that clarify their intentions.

Links: