AN ONTOLOGY TO ORGANIZE DATA ON ALZHEIMER'S DISEASE FROM INTERNATIONAL DATABASES TO SUPPORT INTEGRATED ANALYSIS

Objectives Our goal, through an interdisciplinary Consortium, is to implement a new diagnostic mathematical model by mining large databases of clinical variables of MCI, AD, other demented patients and controls subjects for providing probabilistic diagnoses: oCreate a detailed Ontology for clinical multidimensional datasets from international projects and European clinical centres; oImplement the Logic mining methodology to extract knowledge and generate probabilistic diagnostic models from the ontologies; The ontology will cover entities and concepts extracted from public international databases, integrating numerous clinical and instrumental parameters in a large population of patients. This ontology will allow a better organization of data that will support query process to databases. Methods A prototype of this ontology, derived from analysis of ADNI database, will be represented through OWL formal language and Protégé tool. It will cover psychometric tests, biospecimen, MR and laboratory results. A comparison and armonization of a subset of the parameters collected from ADNI with real world datasets from excellence Italian and Greek clinical dementia centres, will allow the design and implementation of a first database populated by such data. Results The computational core and database were integrated into a very flexible platform. We were able to extract unexpected and highly informative new diagnostic knowledge through the ontology. Conclusions The ontology, a guide for researcher and clinician to query the associated database, to extract any collection of parameters, could be the framework for new clinical protocols and will allow to design a new faster and potentially cheaper diagnostic work-flow.