Flexible Hybrid Cloud Infrastructure for Seamless Integration and Use of Human Biomolecular Data and Reference Maps [1 of 5]

The Human BioMolecular Atlas Program (HuBMAP) is redefining our understanding of the human body by recovering multi-scale tissue organization -- anatomical, histological, and molecular -- at unprecedented resolution, through computational integration of diverse experimental measurements. The HuBMAP Integration, Visualization & Engagement (HIVE) Collaboratory is

an effort among interdisciplinary components developing pipelines for data ingestion and processing, enabling visualization of datasets spanning dozens of biomolecular assays on the HuBMAP portal, leading the development of a human common coordinate framework (CCF), constructing molecularly and spatially resolved reference maps of human tissues, developing

mapping frameworks for the interpretation of new datasets, and coordinating extensive collaborative activities both within HuBMAP and with the broader community. In the production phase of HuBMAP, the HIVE will construct a Human Reference Atlas (HRA), establishing the

HuBMAP Portal as the “go-to” resource for human tissue reference maps and multimodal singlecell data. The next iteration of the HIVE will coalesce the HuBMAP Consortium around a joint vision, develop cutting-edge and scalable tools to achieve it, and ensure its open dissemination to partners and users across the wider international community.

As the HIVE Infrastructure Component (IC), the Pittsburgh Supercomputing Center (PSC), the University of Pittsburgh (Pitt), and Stanford University will provide infrastructure, based on our flexible hybrid cloud microservices architecture, along with community engagement, that will support delivery of this vision in the production phase. To accomplish this, we will focus our efforts

in the following key areas: 1) Curation and Ingestion: Increased automation of data ingestion from HuBMAP data providers, community partners, and the general research community to maximize efficiency and usefulness for building the HRA; 2) Integration: Automated integration and mapping of ingested data to the HRA based on data standards; 3) Findability and Accessibility:

Manifestation of backend resources in the modular architecture of APIs and containers, services, and documentation that minimize user friction in integrated searching, querying, analyzing/aligning and viewing of tissue maps at multiple spatial scales and among multiple layers of information; 4) Interoperability: Extension of the HuBMAP Knowledge Graph to translate

HuBMAP data, HRA assets, and community data among one another via ontologies; 5) Analysis: Infrastructure support to maximally enable users with scalable analyses and workflows among both HuBMAP and user-contributed data and tools, including integration and mapping against the HRA; and 6) Sustainability: Sustainment of open tools, data, and infrastructure for reuse beyond

the production phase. We will grow and harden our model for collaboration, coordination, and engagement led by the IC, with substantial leadership from all HIVE members and participation from all HuBMAP Members.