The Allen Institute for Immunology this month publicly introduced a new data-sharing portal called the Human Immune System Explorer (HISE) that the organization hopes will unleash a new wave of longitudinal immunology studies.
Initially developed as an informatics framework to support the Seattle-based institute and its partners, HISE is now being released to the general bioinformatics research community with the goal of creating the “largest interactive single-cell database of its kind on the human immune system,” according to Paul Meijer, director of software development, database, and pipelines at the Allen Institute for Immunology.
The HISE portal is built on the Google Cloud Platform to facilitate data sharing as well as cloud-based analysis and interpretation. It has been integrated with Google Cloud’s Vertex AI Workbench artificial intelligence platform, simplifying the process of adding Jupyter Notebook to manage experiments. The Allen Institute also built its own user interface on top of Vertex AI.
“We’ve taken Vertex … and customized it for an immunology environment,” Meijer explained.
The portal is meant to manage ever-growing datasets and support long-term studies on immune system diseases.
The initial release of HISE features protocols describing the Allen Institute’s research pipeline, data and a related visualization of the research team’s findings that delays in processing blood samples can change the molecular profile of immune cells, and two new analytic methods and applications called TEA-seq and PALMO. The software also features a collection of computational methods called BarWare that tracks the original source of cells.
PALMO stands for Platform for Analyzing Longitudinal Multi-Omics Data. TEA-seq is an assay that measures transcriptomics, epitopes, and chromatin accessibility through single-cell RNA sequencing. Allen Institute of Immunology scientists described the latter in an article in eLife last year.
Meijer called TEA-seq a “very complex” wet-lab pipeline combined with an equally complex dry-lab pipeline.
“We built these pipelines so that the first analysis of a batch of samples can be done in a standardized and automatic way,” Meijer explained. This will help reduce personnel costs for multiomics analysis, he said.
The immunology division of the Allen Institute debuted in 2018. It now has a team of about 60, including immunologists, computational and molecular biologists, and engineers.
In 2019, the Allen Institute for Immunology, in partnership with the Fred Hutchinson Cancer Research Center, the Benaroya Research Institute, the University of Pennsylvania, the University of California, San Diego, and Colorado University, launched a longitudinal study of health and disease in hundreds of subjects by collecting and analyzing blood and tissue samples through genetic sequencing and other nonmolecular laboratory methods. They also collected demographic, phenotypic, and lifestyle data for what was supposed to be a two-year research project.
“We wanted to do something that was longitudinal and could do multiomics and could really focus down the road on integration of these many different assay types,” Meijer said. “It’s a full integration of human metadata as well as pure assay data,” Meijer said.
HISE also features a virtual collaboration space that Meijer expects will germinate many scientific papers. “If you have just an analysis framework where all the data is, that’s great, but you also need a place to say, ‘Well, why did we do this research?'” he said.
The flexibility of HISE was on display after COVID-19 prompted the World Health Organization to declare a pandemic in March 2000.
“We were just basically getting underway and then the pandemic happened,” Meijer recalled. “We realized that what we had built, both in terms of a dry-lab analysis framework and as well as our wet-lab discipline, could really be used in COVID research because it’s a longitudinal study” of what came to be known as long COVID.
The institute is still following its original research plan, but now includes the study of long COVID.
The Allen Institute for Immunology has funding for five to 10 years to develop HISE, which is somewhat unusual in the world of bioinformatics. The money is coming from internal sources rather than an external grant, according to a spokesperson.
“It allows you to … solve the immediate problem, but let’s also solve it in a structural way so we don’t have to throw it away and then start over again on the next problem,” Meijer said. “That’s what we’re trying to do, and that allows us to expand the exploration portal in the future to have more tools available.”
The guaranteed funding and the strategy it enabled allowed the institute to start the COVID-19 research rather quickly. “Right from the start, we built a framework where we can have multiple research studies going around with different sets of collaborators,” Meijer said.
HISE differs from the US National Institutes of Health’s Human BioMolecular Atlas Program (HuBMAP), which aims to map the molecular level of healthy cells and tissues in the human body, because the Allen Institute project is not an atlas.
While a sister organization, the Allen Institute for Brain Science, has developed an atlas called BrainMap.org, the Allen Institute for Immunology is more interested in building a research framework than an atlas.
“One of our goals is to be translational. We are looking for mechanisms for cure or diagnostics or prevention,” Meijer said.
Meijer called the general release a first step, but said that the Allen Institute is already expanding the portal to make it more interactive. “Our hope is sometime next year, we can offer people [the opportunity] to do actual analysis using some of the tools we’ve built internally,” he said.
The Allen Institute spokesperson said that the organization will release more data and applications for HISE sometime in 2023.