IU team makes significant contributions to constructing a Human Reference Atlas

Katy Börner’s team from the Luddy School of Informatics, Computing, and Engineering’s Cyberinfrastructure for Network Science Center has made significant contributions to constructing a Human Reference Atlas and has led or co-authored six research articles in a just-released HuBMAP package in Nature.

Börner, Victor H. Yngve distinguished professor of engineering and information science, and CNS director, leads one of the two mapping components within the NIH-funded Human BioMolecular Atlas Program. Her team includes experts from Indiana University, Stanford University, California Institute of Technology, Harvard Medical School, and the EMBL’s European Bioinformatics Institute in the UK.

The perspective, “Advances and Perspectives for the Human BioMolecular Atlas Program (HuBMAP),” provides an overview of recent accomplishments and detailed maps at single-cell resolution.

Ellen Quardokus, research scientist at IU, and Andrea Radtke, senior scientist at NIH, both members of the MC-IU team, lead the effort that resulted in “Organ Mapping Antibody Panels (OMAPs): A community resource for standardized multiplexed tissue imaging.”

Organ Mapping Antibody Panels (OMAPs) and accompanying Antibody Validation Reports (AVRs) reflect the hard work and dedication of a team of domain experts in academia, industry, and government. Together, we have worked to overcome the cost, time, and expertise required to map human tissues using multiplexed antibody-based imaging. In addition to providing data on hundreds of community-validated antibodies for diverse imaging methods such as CODEX, IBEX, and Cell DIVE, OMAPs are aligned to data in the Human Reference Atlas, providing evidence for cell types in specific anatomical structures. We are excited to share these resources with the growing spatial biology community and look forward to accelerating discovery together.”

Andrea Radtke, senior scientist at NIH

Yingnan Ju, Luddy School PhD candidate, and Börner collaborated with the GE research team on “3D reconstruction of skin and spatial mapping of immune cell density, vascular distance and effects of sun exposure and aging”.

“This paper seamlessly bridges the gap between practical industrial applications and my academic pursuits” Ju said. “We successfully integrated data visualization and machine learning, synthesizing the invaluable insights from my GE internship with my research at Indiana University. This collaboration has not only laid a solid foundation for future exploration but also added substantial value to my Ph.D. experience. I am excited and look forward to further collaboration.”

CNS Research Scientist Andreas Bueckle was a contact author for “Tissue Registration and Exploration User Interfaces in support of a Human Reference Atlas.” The paper details the registration and exploration user interfaces that empower users to spatially register their data to one of the 66 3D reference organs and semantically and spatially explore the thousands of tissue blocks now mapped to the HRA. Users can do that in a 3D environment, deployed to a web browser and reachable by anyone with an internet connection and with hooks into the HuBMAP Data Portal.

“Building a Human Reference Atlas is a gigantic endeavor that requires skills and abilities from a diverse set of experts,” Bueckle said. “This necessitates tools with state-of-the-art user interfaces.”

Yashvardhan Jain, CNS Research Software Engineer is the lead author of “Robust and generalizable segmentation of human functional tissue units.”

“Construction of the Human Reference Atlas (HRA) requires a combination of human intelligence and machine intelligence,” Jain said. “Detecting and segmenting structures such as functional tissue units in human tissue images is an important step and needs efficient machine learning algorithms that can automate such tasks, enabling processing and analysis of datasets at scale.”

Jain said global data science competition platforms such as Kaggle can be leveraged to engage data scientists, machine learning engineers and researchers across the globe to accelerate the development of efficient, generalizable and scalable algorithms that can aid the development of such automated tools.

Griffin Weber, Harvard Medical School lead the team that published “Anatomical structures, cell types, and biomarkers of the healthy human blood vasculature”.

“The Human Reference Atlas-Vasculature Common Coordinate Framework (HRA-VCCF) dataset is the first open, computer-readable, and comprehensive database of the adult human blood vasculature,” he said. “It contains a list of blood vessels and their branching relationships, as well as associated cell types and biomarkers, the vessel type, anastomoses, portal systems, microvasculature, functional tissue units, links to 3D reference objects, vessel geometries (length, diameter), and mappings to anatomical structures the vessels supply or drain. Because the vasculature extends to all parts of the body, this database forms a core part of the Human Reference Atlas (HRA), linking together organ-specific datasets.”

You can explore a video of the envisioned VCCF.

The MC-IU team authored, “Specimen, Biological Structure, and Spatial Ontologies in Support of a Human Reference Atlas.”

The Human Reference Atlas is a multiscale, pan-tissue, three-dimensional digital atlas of the healthy human body. It provides standard terminologies and data structures for describing specimens, biological structures, and spatial positions linked to existing ontologies. As of July 2023, the atlas comprises 30 organs with 4,340 named anatomical structures, 1,474 cell types, and 3,842 characteristic biomarkers plus more than 1,200 anatomically correct reference objects. All data is available for free at the HRA Portal.

“We are excited to see the atlas grow in coverage and quality during the HuBMAP Production phase,” Börner said.

Their work was among a set of nine scientific papers released in the Nature family of journals describing breakthroughs in high resolution, multiscale and multimodal mapping of the human body in two and three dimensions.

HuBMAP researchers have published more than 270 scientific papers, presenting advances in software and experimental technologies, creating maps of healthy human tissues at the cellular level.