Neglected tropical diseases - A challenge to biomedical ontology engineering

1. Fred Freitas Informatics Center - Federal University of Pernambuco, Brazil KR & KM group - University of Mannheim, Germany fred@cin.ufpe.br Stefan Schulz Institut für Medizinische Biometrie und Medizinische InformatikUniversitätsklinikum Freiburg, Germanystschulz@uni-freiburg.de Zulma Medeiros Aggeu Magalhães Research Center, Oswaldo Cruz Foundation (CPqAM/FIOCRUZ), Recife, Pernambuco, Brazil medeiros@cpqam.fiocruz.br Neglected tropical diseases -A challenge to biomedical ontology engineering

2. Domain description • Diseases, actions, institutions involved • Use cases envisaged • Ontologies and their connections • Challenges summary

3. They have been hardly heard of in richer countries … • … but cause severe disability in the world's poorest regions in over 1 billion people [WHO] • Lymphatic filariasis, • Onchocerciasis, • Schistosomiasis, • Leishmaniasis • Chagas disease(American trypanosomiasis) • Trachoma • Dengue • Malaria • … neglected tropical diseases (NTDs)

4. In most (if not all) of them, biological organisms play these different roles: • Pathogens: complicated organisms with different relevant lifecycles that cause disabilities in humans • Vectors: transmit the pathogens, if their habitat is comfortable for them to reproduce • Hosts: are also a means of transmission (e.g. dogs) • These organisms exhibit complicated lifecycles • They cause diseases with diverse sets of manifestations and symptoms • NTDs require different kinds of actions • prophylaxis, detection and treatment Common features of these diseases

5. An example: Schistosomiasis

6. Prophylaxis: To prevent their transmission • Improvement of basic sanitation (long term) • Educational programs • Field operations • In the environment, by avoiding a comfortable habitat for the organisms • E.g. cover river parts with small polystyrene balls • Against vectors, to reduce heir population • E.g., a chemical smoke to kill dengue’s mosquitoes • Detection: To check individuals’ and populations’ prevalence • Treatments ACTIONS against NTDs In Brazil

7. Municipalities, Federal and State Universities • Actions, treatment, registration • Federal States • Inter-municipality action coordination, policies and guidelines definitions, database analysis • Federal Government • State coordination of the actions, policy and guidelines definitions, database analysis • Oswaldo Cruz Foundation’s instances • Study, research on the disease as well as its actions, campaign planning and creation of treatment and diagnosis new methods Involved institutions

8. Domain description • Diseases, actions, institutions involved • Use cases envisaged • Ontologies and their connections • Challenges summary

9. Decision support systems (DSS) for neglected diseases • Stakeholders: governments on the 3 levels • Phase 1: ontology-based information integration that allows querying heterogeneous NTDs-related databases from different governmental sources (county, state and country). • Integration with OTICSSS [Bissolotti & Silva 2009], an emerging health information integration initiative in Brazil. • Phase 2: Diagnoses of the situation • Phase 3: Assessment of actions’ effectiveness Intended Use cases for the ontologies I

10. A document search engine for NTDs • Morphosemantic indexing developed by Freiburg partners • Ontology-enhanced text processing • take advantage not only of keywords, but also of the ontology structure • Intelligent agents/decision support systems that • assist on diagnosis and prognosis of neglected diseases in potential and actual patients. • serve as a basis for eLearning artifacts Intended Use cases for the ontologies Ii

11. Domain description • Diseases, actions, institutions involved • Use cases envisaged • Ontologies and their connections • Challenges summary

12. EXTERNAL ONTOLOGIES TO BE USED BioTop Core ontology for biomedicine SNOMED Procedures, treatments, pharmacological products, diseases, social context, body structure, event,… OBO, GO Malaria [Ceusters & Smith 2009]: guideline Issue: how to integrate

13. Top: Basic Formal Ontology Middle downwards: Biotop

14. MODELING PRINCIPLES Modeling with good ontological engagement Temporal aspects are sometimes needed [Ceusters & Smith 2009] If possible, adopting OWL DL, or its temporal variations [Artale & Franconi 1998, Lutz 2002] Reuse Since diseases share common features… Pathogens, hosts, habitat type and locations, disease, exams …a core ontology(ies) about them facilitates reuse Issue: a new core ontology vs extending BioTop Modularization

15. Environment/ Hosts / Vectors Prophylactic actions ONTOLOGIES Managementactions Pathogens Detection Treatment Diseases

16. A new NTD extend the core ontology on NTDs Vectors\pathogens\diseases: connects to Biotop and may reuse some vocabulary \ definitions from external sources (GO, SNOMED,…) Actions connect to BFO and the core ontology to be defined ONTOLOGY BRIDGING

17. On time t1, there was x1 new instances of the disease detected by exams in a population p and a certain prophylactic action pa1 was taken On time t9, the rate of the new instances x9 is falling down (x1>x9) for the same number of new exams in the same conditions If ok, this is due to pai The same for treatments TEMPORAL ASPECTS

18. Domain description • Diseases, actions, institutions involved • Use cases envisaged • Ontologies and their connections • Challenges summary

19. Administrative regions are arbitrarily delineated, but endemies don’t respect such divisions • They depend upon natural sources of the vectors/hosts, mostly geographical accidents • Actions can only be well succeeded / assessed if they focus to the endemic spaces • They vary according to the disease • Schistosomiasis: hydrographic basins • Dengue/Malaria/Filariasis: sources of still water • Bubonic plague: mountains, hills • … political x vectors’ habitats

20. Different granularities • individual disease vs. affected populations • Linking very different entity types (e.g., socioeconomic factors, housing, mobility,…) • Public health authorities and their roles in the process • Temporal management of data, according to what is defined in the ontology (phases, stages, action sequences, …) • Complicated organisms with different relevant lifecycles • Broad spectrum of disease manifestations • Benefit: the different standpoints (health researchers, managers, workers) hopefully can live in harmony Modeling challenges

21. Thank you for your attention! Questions , remarks?

22. Bissolotti, k. ; Tavares, J. L. . Integração de Tecnologias de Informação e Conhecimento no Projeto OTICSSS. In: XVII Encontro de Jovens Pesquisadores, 2009, Caxias do Sul. XVII Encontro de Jovens Pesquisadores, 2009. Elena Beißwanger, Stefan Schulz, Holger Stenzhorn and Udo HahnBioTop: An Upper Domain Ontology for the Life Sciences - A Description of its Current Structure, Contents, and Interfaces to OBO Ontologies in Applied Ontology, Volume 3, Issue 4, Pages 205-212, IOS Press, Amsterdam, Netherlands, December 2008 Freitas, F. ; Schulz, S. ; Moraes, E. . Survey of current terminologies and ontologies in biology and medicine. RECIIS. Electronic Journal of Communication, Information and Innovation in Health (English edition. Online), v. 3, p. 1-13, 2009 Alessandro Artale, Enrico Franconi. A Temporal Description Logic for Reasoning about Actions and Plans, Journal of Articial Intelligence Research 9 (1998) 463-506 12/98 Lutz, C. The Complexity of Description Logics with Concrete Domains, PhD Thesis. Werner Ceusters, Barry Smith, Malaria Diagnosis and the Plasmodium Life Cycle: the BFO Perspective, Nature, Nov 2009. references

23. Modeling challenges • Modularization of the big life science ontologies • BioTop, SNOMED,… • In order to guarantee more • Control, management • Better visualization • Reasoning performance