A carregar...
Publicação
The ImmunoPeptidomics Ontology : design and evaluation
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 2.6 MB | Adobe PDF |
Resumo(s)
There is significant hardship in manipulating the complex and variable data generated
by biomedical subdomains such as immunopeptidomics. Therefore, ontologies are often
used to express knowledge in a domain to help establish a standard nomenclature to help
handle and integrate medical data.
As a new field, immunopeptidomics lacks standardization: there is no recognized vocabulary or explicitly defined meanings. Hundreds of ontologies cover the biomedical
domain, including neighboring subdomains like proteomics and immunology, but none
adequately cover immunopeptidomics. This shortcoming needed to be addressed so cancer research in this domain could start to affect clinical practice.
This dissertation details the development of the ImmunoPeptidomics Ontology, ImPO,
to allow later integration of data produced by immunopeptidomics in personalized oncology. For this purpose, ImPO was designed following a process that comprised: capturing
domain specialist knowledge in immunopeptidomics; iterative conceptual modeling of
the domain through an Entity-Relationship model, semantic modeling by OWL formalization of the ER, cross-referencing ImPO with 28 external ontologies, and evaluation
with competency questions and construction pitfalls.
Unlike most biomedical ontologies currently accessible, ImPO was created to be populated with data and function as the semantic backbone of a knowledge graph (KG).
ImPO was created as part of the KATY project, which aims to apply “AI-empowered
knowledge” to clinical practice in Clear Cell Renal Cell Carcinoma. ImPO is one of the
KATY KG components that will facilitate data integration in the project and give explainability to the AI techniques created in the project. Nonetheless, the ImPO ontology was
created to be utilized independently of the KATY KG as a stand-alone knowledge model
to aid in data integration and knowledge discovery in immunopeptidomics.
Descrição
Tese de Mestrado, Bioinformática e Biologia Computacional, 2022, Universidade de Lisboa, Faculdade de Ciências
Palavras-chave
Ontologia Imunopeptidómica ImPO Teses de mestrado - 2023