Building Comprehensive Single-Cell Atlases

Humans have been mapping their world for as long as history has been recorded. In biology, scientists are acting on this instinct by creating cell atlases. Cell atlasing is the systematic effort to create detailed reference maps of all cell types within an organism. Using technologies such as single-cell genomics, transcriptomics, and spatial analysis, researchers are identifying the type, location, and functional state of each cell. This information is building comprehensive resources for understanding complex biological systems in health and disease.

Advances driving cell atlases

Technological developments are driving cell atlas construction in ways that were not feasible just a few years ago. Among the most consequential is the ability to scale single-cell experiments, which has made it possible to build much larger and more detailed cell atlases. “We can now do orders of magnitude more cells and samples than we could just a year or two ago,” stated Charlie Roco, Ph.D., CTO and Co-Founder of Parse Biosciences. This was recently demonstrated in the Tahoe-100M, a collaborative effort between Parse, Ultima Genomics, and Tahoe Therapeutics, which led to the world’s largest individual single-cell dataset at 100M cells across 60,000 samples. Roco noted that with an estimated 37 trillion cells in the human body, efforts like these are still only scratching the surface of what's possible in cellular analysis.

David Peoples, Chief Financial and Business Officer at Ultima Genomics, explained that recent advances in single-cell and sequencing platforms, as well as CRISPR technologies, have made it possible to create more detailed and comprehensive cell atlases. Improvements in these methods now allow researchers to profile millions of individual cells, identify and characterize rare populations, and explore their roles across tissues, developmental stages, and disease contexts. Peoples believes that the continued decline in sequencing costs has been a major driver of this progress, making large-scale experiments more feasible and affordable. Lower per-sample and per-read costs mean researchers can process exponentially more cells and replicate experiments across diverse biological conditions. These developments have enabled the creation of comprehensive cell atlases and led to a deeper exploration of biological systems.

Notable cell mapping efforts

One of the most well-known examples of cell atlasing is the Human Cell Atlas (HCA) Consortium, an initiative to build a global, open-access 3D reference map of all human cell types.¹ The project combines transcriptomics, spatial genomics, machine learning, and other advanced tools to uncover patterns in cell variation and assemble detailed reference maps. As of late 2024, the HCA contains data from over 9,200 donors and 62.7 million cells across 478 projects.² Draft atlases from 18 biological networks are already available through its data portal. In a major release last year, the consortium published over 40 studies that highlighted new tissue maps, AI tools, and biological discoveries, including skeletal and placenta development, gut inflammation, and diverse population data. These advances marked a major step toward the first draft atlas.

While many atlasing projects are focused on whole organisms or healthy tissues, other efforts target specific systems or diseases. Examples include the BRAIN Initiative Cell Atlas Network (BICAN), which maps the diverse cell types of the human brain; the Cancer Genome Atlas, which catalogs genomic changes across dozens of cancer types; and the Malaria Cell Atlas, which charts the parasite’s life cycle at single-cell resolution. Similarly, disease-focused atlases have been used to study immune responses in conditions like COVID-19. In one such study, researchers atlased peripheral blood from COVID-19 patients to identify immune pathways linked to long COVID.³ They found shared and sex-specific signals, including altered monocyte activation, suggesting that treatment strategies may need to be personalized.

These atlasing efforts are only at the beginning, with many new studies now underway. For instance, Roco shared that Parse’s technology is powering large-scale cell atlasing projects at Mount Sinai and Vanderbilt. At Mount Sinai, researchers are profiling alternative splicing to uncover molecular drivers of neurodegeneration, while at Vanderbilt, the focus is on mapping neutralizing antibody-producing plasmablasts to measles, mumps, and rubella to advance therapeutic discovery. Peoples also highlighted Ultima Genomics’ partnerships with the Chan Zuckerberg Initiative, Arc Institute, Tahoe Therapeutics, and Myllia Biotechnology. These collaborations span cell atlasing for AI-driven drug discovery, CRISPR-based drug target discovery, and the creation of a virtual cell atlas.

Designing an atlas project

With initiatives like these gaining momentum, researchers are now considering how to design atlases of their own. Like all experiments, successful atlas projects require careful planning to ensure meaningful results. Roco explained that the first step is to define the project’s goals and determine how many cells to profile to meet them. In addition, tissue selection, sample preparation, and the computational pipeline are all critical components. He noted that having high-quality samples and sample preparation at the outset is essential for generating high-quality single-cell data. Once the data are generated, decisions about how to perform cell type annotation become important. Aligning that process with community standards, Roco added, helps ensure the dataset contributes to broader atlas-building efforts.

In addition to thoughtful design, Peoples emphasized the importance of scale, diversity, and data quality. He explained that cost reductions achieved by Ultima have made it possible to expand experimental designs to include more samples, time points, and population diversity, strengthening both statistical power and the generalizability of results. Achieving consistent quality also requires minimizing batch effects and technical variability through uniform sample collection, robust library preparation methods, and standardized data processing pipelines.

Batch effects can be especially challenging when datasets are assembled from many small runs. Roco noted that Parse’s fixation-based workflow allows researchers to preserve and store samples, separating collection from downstream barcoding and processing. “Fixation, combined with Parse’s ability to multiplex hundreds to thousands of samples at a time, means that you can pool samples together and effectively put them into the same batch to remove a large amount of risk to potential batch effects,” he said.

The future of cell atlasing

Looking ahead, experts anticipate that continued advances will expand both the scope and the impact of atlas projects. “We see the integration of affordable, massively scalable sequencing with cutting-edge single-cell library prep and computational methods as enabling a new era of bioAI-driven discovery,” stated Peoples. As large-scale atlases become increasingly comprehensive and representative, he said, they will provide foundational reference maps of normal and disease biology. These resources will allow researchers to more precisely define cellular dysregulation, discover biomarkers, and stratify patient populations. The ability to generate atlas-scale datasets cost-effectively with platforms like Ultima’s, Peoples added, will help accelerate both basic research and precision medicine.

“The ability to create a reference map of all cell types that exist in the human body would have profound implications,” Roco stated. Drawing parallels to the Human Genome Project, he explained that mapping genes allowed scientists to place them in context within the genome and understand how variations or mutations might affect health. In the same way, a complete cell-type map could reveal how different cells interact, what occurs when certain cells lose function, and offer valuable tools for drug developers to design more effective treatments. Acknowledging the progress made so far, Roco emphasized that the scientific community still has substantial work ahead, with his team contributing as part of the broader effort.

References

1. Regev A, Teichmann SA, Lander ES, et al. The Human Cell Atlas. Elife. 2017;6:e27041. Published 2017 Dec 5. doi:10.7554/eLife.27041

2. Parums DV. Editorial: The Human Cell Atlas. What Is It and Where Could It Take Us?. Med Sci Monit. 2025;30:e947707. Published 2025 Jan 1. doi:10.12659/MSM.947707

3. Hamlin RE, Pienkos SM, Chan L, et al. Sex differences and immune correlates of Long Covid development, symptom persistence, and resolution. Sci Transl Med. 2024;16(773):eadr1032. doi:10.1126/scitranslmed.adr1032