The Arcus Archives at the Children’s Hospital of Philadelphia (CHOP) aims to collect and describe research data from across the Research Institute. To accomplish this, the Arcus Library Science team has established processes, structures, and descriptive metadata schemas informed by all phases of the research data preservation and reuse lifecycle at CHOP including archiving, cataloging, display, and usability. This paper introduces the Arcus Archives metadata schema and the key structures it relies upon. Specifically, this paper explains how a small set of dataTypes subfields within a hierarchical, archivally informed metadata structure utilize a shared file directory organization structure to thoroughly and accurately process, catalog, and surface for discovery metadata about petabytes of archived pediatric research data from multiple and growing data modalities. The metadata schema itself is flexible but consistent enough to apply to a myriad of data types produced during the conduct of pediatric research. Each metadata record reflects the unique archival arrangement done for each collection. It encompasses the framework of a shared recommended “project template” file directory structure which includes manifest and protocol files that allow for meaningful capture and organization of numerous complex data files and their relationships to one another.