ONTOLOGY DEVELOPMENT FOR A PHARMACOGENETICS KNOWLEDGE BASE

Diane E. Oliver, Daniel L. Rubin, Joshua M. Stuart, Micheal Hewett, Teri E. Klein, Russ B. Altman2001 · DOI

TL;DR

This paper introduces the PharmGKB, a pharmacogenetics knowledge base designed to integrate complex biological, pharmacological, and clinical data. By developing a formal ontology, PharmGKB aims to facilitate the discovery of genotype-phenotype correlations, accelerating research into how genetic variations influence drug responses.

What Problem Does This Paper Address?

Understanding how an individual's genetic makeup affects their response to medications is a critical but complex challenge. This field, pharmacogenetics, generates vast amounts of data from diverse sources – laboratory experiments, computational analyses, and clinical studies. Coordinating and integrating this data to identify meaningful genotype-phenotype correlations is difficult due to a lack of standardized terminologies and data representations. The paper addresses this by proposing a structured knowledge base, PharmGKB, to organize and make accessible this complex information, thereby enabling hypothesis generation and advancing research.

Methods

The core of PharmGKB's approach is the development of a formal ontology using a frame-based knowledge-representation system (Protégé). This ontology has two main components:

  • Knowledge-base schema for experimental data: This defines the structure for storing experimental results, including classes, slots, and facets that dictate data models and input constraints. Experimental data itself is treated as instances within this schema.
  • Domain conceptual knowledge: This component models the biological, pharmacological, and clinical concepts relevant to pharmacogenetics. It uses classes and instances to represent shared names, meanings, and relationships between concepts.

The authors employ a "bottom-up" approach for modeling experimental data, adapting the ontology to fit the data provided by initial research groups. Conversely, a "top-down" approach is used for domain knowledge modeling to ensure a broader conceptual framework.

Datasets

The paper describes the initial data being integrated into PharmGKB from five research groups funded by the NIH. These groups are studying various aspects of pharmacogenetics, including:

  • Enzymes involved in drug metabolism (e.g., TPMT, HNMT, COMT).
  • Transporter genes (e.g., SERT, VMAT2) and their impact on antidepressant efficacy and toxicity.
  • Pharmacogenetics of anticancer agents (e.g., irinotecan, etoposide).
  • Tamoxifen metabolism by cytochrome P450 enzymes.
  • Genetic factors influencing patient response to asthma medications.

The paper also references existing public databases like dbSNP and GenBank, indicating that PharmGKB aims to be compatible with and leverage these resources.

Key Insights and Ideas

  • The critical role of ontologies: The paper strongly emphasizes that a well-defined ontology is crucial for managing the complexity of pharmacogenetic data. It provides a shared nomenclature, facilitates consistent data interpretation, and enables effective querying.
  • Integrating experimental data and domain knowledge: PharmGKB's strength lies in its ability to integrate raw experimental data with conceptual knowledge about the domain. This allows for richer queries and validation, such as confirming that entered metabolites are indeed associated with a given drug.
  • Bottom-up for data, top-down for knowledge: The dual approach to ontology development (bottom-up for experimental data, top-down for domain knowledge) allows for iterative development and early consideration of the broader conceptual landscape.
  • The need for standardization: The paper highlights the significant challenge posed by the lack of standard terminologies and representations in pharmacogenetics, underscoring the value of PharmGKB in establishing such standards.

Key Results

The paper focuses on the methodology and design of PharmGKB rather than presenting quantitative results from extensive data analysis. However, it reports on preliminary experience with gene-sequence data submissions:

  • The developed ontology is compatible with the dbSNP database, supporting automatic dbSNP submissions.
  • A test phase with collaborators from the Mayo Clinic successfully incorporated experimental data into the XML format required for direct XML submissions, which corresponds to the PharmGKB ontology.
  • The ontology development process revealed subtleties in definition that required clarification for shared understanding between developers and data submitters.

Limitations and Future Work

The authors acknowledge several limitations and outline future directions:

  • Complexity of data integration: The paper notes the inherent complexity of integrating data from diverse sources and the ongoing challenge of ensuring consistency.
  • Evolving nature of knowledge: Pharmacogenetics is a rapidly advancing field, requiring continuous updates and management of the knowledge base.
  • Future work includes:
    • Expanding the content to cover a broader scope of pharmacogenetic research.
  • Enhancing constraint representation for improved data validation.
  • Developing change-documentation features for better management of evolving knowledge.
  • Implementing merging techniques for updating the knowledge base.
  • Improving user query methods for more intuitive genotype-phenotype association retrieval.

Summary updated

Abstract

Research directed toward discovering how genetic factors influence a patient's response to drugs requires coordination of data produced from laboratory experiments, computational methods, and clinical studies. A public repository of pharmacogenetic data should accelerate progress in the field of pharmacogenetics by organizing and disseminating public datasets. We are developing a pharmacogenetics knowledge base (PharmGKB) to support the storage and retrieval of both experimental data and conceptual knowledge. PharmGKB is an Internet-based resource that integrates complex biological, pharmacological, and clinical data in such a way that researchers can submit their data and users can retrieve information to investigate genotype-phenotype correlations. Successful management of the names, meaning, and organization of concepts used within the system is crucial. We have selected a frame-based knowledge-representation system for development of an ontology of concepts and relationships that represent the domain and that permit storage of experimental data. Preliminary experience shows that the ontology we have developed for gene-sequence data allows us to accept, store, and query data submissions.
BibTeX
@misc{oliver2001ontology,
  author = {Diane E. Oliver and Daniel L. Rubin and Joshua M. Stuart and Micheal Hewett and Teri E. Klein and Russ B. Altman},
  title = {{ONTOLOGY DEVELOPMENT FOR A PHARMACOGENETICS KNOWLEDGE BASE}},
  year = {2001},
  doi = {10.1142/9789812799623\_0007},
}
Figure