=Paper= {{Paper |id=None |storemode=property |title=ADO: A formal ontology for understanding structure of information and domain knowledge of Alzheimer’s disease |pdfUrl=https://ceur-ws.org/Vol-897/poster_2.pdf |volume=Vol-897 |dblpUrl=https://dblp.org/rec/conf/icbo/MalhotraGH12 }} ==ADO: A formal ontology for understanding structure of information and domain knowledge of Alzheimer’s disease== https://ceur-ws.org/Vol-897/poster_2.pdf
ADO: A formal ontology for understanding structure of information and
             domain knowledge of Alzheimer’s disease
                   Ashutosh Malhotra, Michaela Gündel and Martin Hofmann-Apitius
    Department of Bioinformatics, Fraunhofer Institute for Algorithms and Scientific Computing, Sankt Augustin
                                                    53754 Germany



Formal biomedical ontologies offer the capability to retrieve   Including depth, breadth, tangledness, and fan-outness.
domain-specific knowledge. Particularly for grievous dis-       Using our state-of-the-art text mining tool, ProMiner, which
eases like Alzheimer’s where latency of onset to clinical       takes the ADO dictionary as input, functional evaluation of
symptoms takes many years, automated structured mining          the ontology was done over a test set of 200 abstracts.
approaches can generate great value. No single system is        Results of the evaluation produced an F-score of 0.72
currently capable of covering a complete domain of Alz-         showing that our designed ontology accurately captures
heimer’s disease (AD) by itself and manual effort will be       relative good range of AD concepts in scientific text.
insufficient to retrieve all relevant information. This makes   The ADO dictionary was also used to mine 650 AD
it necessary to develop focused automated applications that     Electronic patient health record’s (EHR) for comorbidity
can deal with individual aspects of the disease in a reliable   analysis, systematically looking for other diseases or
manner. A formal Alzheimer’s disease ontology (ADO)             disorders that may exist simultaneously but independently
with potential to engineer knowledge can provide a good         in patients suffering from AD. Term frequency count
way to describe and organize the AD-specific knowledge          calculation shows that hypertension, diabetes and stroke are
sphere. We have developed an ADO representing                   the top three disorders which occur frequently in patients
knowledge about impact, treatment, risk factors, diagnostic,    already suffering from AD. When integrated into our
effects and phases of AD. This is the first example of an       information retrieval system SCAIView, ADO leveraged the
ontology covering different aspects of AD.                      efficiency of semantic information retrieval and knowledge
ADO is constructed in accordance to the ontology building       representation by providing the possibility to perform cross
life cycle and using the Basic Formal Ontology (BFO) as its     sectional searches with other dictionaries. Particularly its
formal upper ontology. Protégé OWL editor was used as a         intersection with ‘HypothesisFinder’ facilitated the retrieval
tool for building ADO in Ontology Web Language (OWL)            and aggregation of domain-specific speculations
format. Collection of terms related to AD was performed by      (hypotheses).Some of the hypotheses proposed for a
scanning various knowledge sources which includes review        particular stage of AD retrieved by using this combination
articles, content of online books, standard knowledge bases,    methodology are:
encyclopedias, glossaries, and informative online sources
and websites. ADO was developed around four main con-           We suggest that M-CSF, thus generated, contributes to the
cepts, namely clinical things, etiological things, molecular    pathogenesis of AD, and that M-CSF in cerebrospinal fluid
and cellular mechanisms and Non clinical things; these con-     might provide a means for monitoring neuronal perturbation
cepts and their hierarchies represent what we call “contextu-   at an early stage in AD.(PMID: 9144231)
al views” on the ontology. Concepts under each of these
views are inferred using reasoning and are a collection of      Our results show that modification of tubulin function may
classes that can be either continuants or occurrents; thus,     contribute to intermediate or late stages in the pathogenesis
BFO upper class of these views is “entity”. This solution       of sporadic and inherited AD as well as FTDP-
that aims at creating a contextual collection of classes can    17.(PMID:16155344)
aid in domain-specific retrieval of documents, producing
                                                                In a benchmark experiment, the combination methodology
desired search results. Further, ADO was enriched by add-
                                                                using ADO and HypothesisFinder for extracting AD stage
ing synonyms, useful annotations and references.
                                                                specific speculative sentences was compared to the human
ADO has been manually curated by an expert clinician, who
                                                                curated Alzswan knowledge base. Our approach displays a
added certain more clinically relevant concepts to ADO
                                                                much higher stage specific coverage of speculative
increasing its pragmatic usability. Also, the structural
                                                                statements about the etiology of AD than the human (expert)
features of ontologies reflecting topological and logical
                                                                curated database AlzSWAN.
properties were measured by means of context-free metrics


                                                                                                                            1