Oregon Health and Science University, Lawrence Berkley National Laboratory, and University of North Carolina
The primary goal of the proposed SRI is to enable the ARAs and ARS to leverage the KPs to answer translational questions in a reproducible manner, using a sustainable and collaborative infrastructure. This vision requires a consensus-driven approach to maturing and establishing Translator standards; it also requires normalization services that implement these standards and render the results from one component useful across any other. The feasibility phase of Translator (hereafter “Phase 1”) saw a shifting line between federated vs. centralized database approaches, with a trend towards increasing centralization of core knowledge resources. We see this trend continuing toward more core knowledge resources, unified within a wider data lake and supported by federated architecture of analytic tools, reference datasets, and specialized high-volume analytic services. Our proposal is organized around the following themes:
Community governance coordination
Architecture and API specifications
Biolink model
Integrated reference ontologies
Knowledge graph and data lakes
Next-generation Shared Translator Services
A registry of Translator KPs, ARAs, and shared services