POI-KB: Design and Development of a Knowledgebase for Accelerating Perioperative Organ Injury Research and Translation

Acute perioperative organ injury (POI) causing single- or multiple-organ failure is the leading cause of mortality and morbidity in surgical patients. Strategies to prevent or treat POI can significantly improve perioperative outcomes. Despite the intense investigational efforts in POI research, a large gap still exists in respective domain knowledge crosstalk and research efforts synchronization, which hinders the advancement of POI research. Therefore, there is an urgent need to develop a knowledge base platform that will bridge the multidisciplinary and translational gap among health science researchers who identify the important healthcare issues, basic science researchers who reveal the disease mechanisms underlying POI, and physician scientists who carry out important clinical trials to bring these preventive/therapeutic interventions to clinical practice. Interrogating large-scale public data resources offers unprecedented opportunities for advancing organ injury research. The ability to interrogate large-scale data resources has become a central focus of many research efforts. While successful data integration efforts have resulted in novel diagnostics, therapies, and prevention strategies, many data resources remain underutilized. First, researchers often lack even the most basic tools for navigating this complex data landscape which limits the full potential of leveraging the data for research. Secondly, researchers may also lack knowledge on what kinds of research questions for a specific topic can be addressed by those existing data resources. As POI research tends to be multidisciplinary and the lack of a knowledgebase for POI hinders the effort on leveraging existing public data resources for POI research. Here, we propose to develop Perioperative Organ Injury Knowledge Base (POI-KB), an innovative knowledgebase to systematically catalogue and curate POI information from public data resources through a multidisciplinary team science effort. POI-KB will consist of four components: POI-Ontology – an ontology capturing biomedical concepts (e.g., diseases, genes, drugs) discovered in relevant POI literature and other public knowledgebases through formal ontological development process; POI-Graph – a knowledge graph capturing relations in POI-Ontology through text mining, database fusion, ontologies and network mapping methods, followed by human verification; POI-Catalyst – a discovery engine for POI to accelerate POI research by indexing existing resources and datasets using POI-Ontology to facilitate resource discovery and facilitating hypothesis generation using semantic web technologies; and POI-Portal – a web portal to host tools and resources developed and foster community engagement for POI-KB. We aim to engage and evaluate POI-KB in supporting use cases in POI from various research communities across the translational spectrum. We will work with stakeholders to gather feedback and ensure the POI-KB effectively addresses the needs of both researchers and clinicians. We will demonstrate the potential of the POI-KB through impactful use cases, generating new scientific knowledge and mechanistic hypotheses for further testing.