Software or Web Services to Re-Represent Existing Scientific Data and Knowledge into a Knowledge Graph Format

Fast Track Proposals will be accepted. Direct-to-Phase II will be accepted. Number of anticipated awards: 1-3 Budget (total costs): Phase I: $ 300,000 for up to 1 year. Phase II: $ 1.5 million for up to 3 years. Background Scientific research produces more data and knowledge than ever before and it is becoming increasingly challenging for researchers to digest and analyze the data to derive new knowledge and insights. Increasingly, researchers will depend on computer algorithms to assist them with data analysis and with the discovery of new information and knowledge. For example, during the COVID-19 pandemic, the discrepancy between the speed of data generation and the ability to digest and analyze the data was acutely illustrated with almost 25 thousand new papers published in the first half of 20201 . The public health response against emerging pandemic threats and the treatment and control of existing infectious- and immune-mediated diseases requires the continuous analysis of newly emerging data, information, and knowledge; a task that is increasingly complicated by the large volume of data that is being generated. Knowledge graphs (KG's) have emerged as a prominent technology for data representation and integration and are widely used in the industry. KG's represent information about entities and relationships between entities in a semantically rich way that enables efficient data retrieval and analysis by computational algorithms, such as automated reasoners and AI, and have shown great promise for data management and knowledge discovery. A familiar use case for KG's in the scientific domain is for the analysis of collaborations between authors based on published papers, for the exploration of pathways in molecular biology, and for drug repurposing. A major bottleneck for leveraging KG's for biomedical research and discovery is the difficulty to represent data and knowledge into a KG-compatible format. Many biomedical data are stored in formats that are not compatible with KG's, such as spreadsheets and databases, and cannot be used by automated reasoners and other computer-assisted knowledge discovery methods, unless these data are represented in a KG format. Given the complexity of accurately representing semantic information on scientific data and related methods, automated methods and software will be essential to scale up the amount of data on infectious- and immune-mediated diseases available in KG's for accelerating scientific discovery