Genetic dispositions play a major role in individual disease risk and treatment response. Genomic medicine, in which medical decisions are refined by genetic information of particular patients, is becoming increasingly important. Here we describe our work and future visions around the creation of a distributed infrastructure for pharmacogenetic data and medical decision support, based on industry standards such as the Web Ontology Language (OWL) and the Arden Syntax.
There is growing consensus in the medical and pharmaceutical community that further progress in the development of new therapies will necessitate a fundamental change in medical practice: away from broadly defined disease concepts and therapeutic regimes, and towards a fine-tuned evidence-based, personalized medicine. Genomic medicine is an important component of personalized medicine, and refers to a system in which medical decisions are refined by combining medical history with current physiological indicators against a genetic background for a particular patient [1]. Since genetics plays a major role in determining the response to a broad range of therapeutic treatments, the appropriate use of this pharmacogenetic information for guiding treatment decisions has the potential to improve the efficacy of treatments and reduce the incidence of adverse drug events.
While nearly one fourth of all outpatients in the US received one or more drugs for which pharmacogenetic knowledge is available [2], it is still not common that pharmacogenetic findings are used in medical practice. Doctors are usually not specifically trained in genomic medicine, the cost-benefit trade-off of genetic testing is often unclear, and there is not enough time to incorporate potentially complex pharmacogenetic reasoning in routine medical decision making.
Therefore, the development of decision support systems capable of handling pharmacogenetic data is clearly essential to the realization of personalized medicine. These systems need to provide accurate and timely reminders and decision support tailored to each individual patient, drug and therapeutic regime. However, creators of decision support systems for genomic medicine face the challenge of working with highly heterogeneous information concerning the relationship between genetics and drug responses based on limited trials. They need to deal with distributed, incomplete and possibly contradictory information.
Here, we describe our ongoing work and future visions of employing information technologies to address this problem and towards 1) seamless integration of relevant pharmacogenetic data in a distributed setting, 2) the exploitation of clinically relevant pharmacogenetic knowledge in clinical decision support and 3) the design and dissemination of clinical decision support systems that improve the quality of health care delivery.
Several relevant data sources have already become available in an open, interlinked format, or will be made available soon. We, together with other participants of the Health Care and Life Science Interest Group [3] of the World Wide Web Consortium (W3C, [4]), worked on making several relevant datasets accessible in RDF/OWL [5] format. The extraction and conversion of additional relevant datasets such as the Pharmacogenomics Knowledge Base (PharmGKB [6]), Drugbank [7], Online Mendelian Inheritance in Man (OMIM [8]), dbSNP [9] or SNPedia [10] is currently ongoing.
In addition to manually curated data, natural language processing has been successfully used to identify pharmacogenomic information, such as gene-drug-disease relationships [11] or descriptions of new molecular diagnostics [12].
Organisations dedicated to reviewing current evidence and publishing recommendations about pharmacogenetics have emerged. For example, the Clinical Pharmacogenetics Implementation Consortium (CPIC) was recently initiated in the context of the PharmGKB. The CPIC members create, curate, review, and update written summaries and recommendations for implementing specific pharmacogenetic practices. Levels of evidence and strength of recommendations are documented. Another example of such an organisation is the Evaluation of Genomic Applications in Practice and Prevention initiative (EGAPP [13]). The text-based recommendations provided by such initiatives can be easily formalized as rules for clinical decision support.
Ontologies help improve interoperability and data consistency. Several ontologies relevant to pharmacogenetics have become available in recent years. The Translational Medicine Ontology (TMO, [14]) provides a foundation upon which chemical, genomic and proteomic data can be harmonized and linked to disease, treatments and electronic health records. The Suggested Ontology for Pharmacogenomics (SO-PHARM, [15]) was the first to demonstrate how pharmacogenomic knowledge can be captured based on the Open Biomedical Ontologies (OBO) resources. The Sequence Ontology aims to describe the features and attributes of biological sequences [16]. It holds terms and relations of value for describing genetic variation including single nucleotide polymorphisms (SNPs) at the sequence level.
Our work is guided by international standardisation efforts, and we also participate in standardisation activities. The most important standardisation organizations in this context are Health Level 7 (HL7 [17]); and the World Wide Web Consortium (W3C), which develops standards for lar
This content is AI-processed based on open access ArXiv data.
