Community Data Hub for Integrative Visualization

Project Summary The rich, diverse, and complementary multimodal datasets provided by the Common Fund Data Ecosystem's (CFDE) Data Coordinating Centers (DCCs) provide exciting opportunities for integrative analysis and exploration. Here, we propose a systematic data visualization-driven approach to investigate the model-based

predictions of the target genes of candidate cis regulatory elements (cCREs). Our project will leverage biomolecular data from key repositories in the CFDE, including the Genotype-Tissue Expression (GTEx) project, the Human BioMolecular Atlas Program (HuBMAP), the Cellular Senescence Network (SenNet), and

the 4D Nucleome (4DN) consortium, and will expand to include genomic data from external large projects such as the Encyclopedia of DNA Elements (ENCODE) project. Our motivating use case is the integrative exploration and validation of predictions for target genes of cCREs in the context of relevant linkage

disequilibrium (LD), molecular quantitative trait loci (QTLs), epigenomic, transcriptomic, (spatial and non-spatial) single-cell, and Chromosome Conformation Capture (3C/Hi-C) data. Central to this objective are (1) a flexible data management system tailormade for visualization that addresses the challenges of organizing

private and federated public datasets, (2) tools for the creation of custom, interactive exploratory visualizations, and (3) the ability to save and share visualizations in a FAIR manner. To achieve our goals, we will develop the Common Fund Visualization Hub, a cloud-based platform enabling unified access to data

resources in the CFDE alongside user-contributed, private data in projects. It will provide a fluid metadata tagging system that supports structured and unstructured tags. It will support multiscale, genomic visualization using HiGlass along with a user interface for on-the-fly visualization construction based on the

Gosling visualization grammar. It will also support multimodal data visualization using Vitessce and enhance reusability through sharing of visualizations and storytelling features to communicate discovery processes. The Common Fund Visualization Hub will thus facilitate comprehensive data interpretation and

communication through integrative, interactive analysis of diverse genomic and bioimaging datasets. Using this platform, we will evaluate and compare regulatory element-to-gene predictions from leading machine learning models based on different evidence-based methodologies. By streamlining integrative visualization of

existing CFDE data, our platform aims to help maximize the utility of CFDE investments to accelerate biomedical discovery. Our platform will empower community members to independently browse, record, and present their own data combined with data sourced from across the ecosystem, enhancing collaboration and

communication both within and across research groups, as well as with the broader community and the general public.