PerspectiveCell Biology

Whole cell maps chart a course for 21st-century cell biology

See allHide authors and affiliations

Science  26 May 2017:
Vol. 356, Issue 6340, pp. 806-807
DOI: 10.1126/science.aan5955

Navigating a course for understanding human cell biology will require collaboration across disciplines, funding agencies, and research efforts, from individual investigators to large consortia.


Cells are complex machines constructed from genetic blueprints generated by mutation and evolutionary forces, whose information is expressed under the influence of the cellular environment. Although each cell in any human has the same genes, complex regulatory networks determine which genes are expressed, creating the large variety of specialized cell types that constitute our bodies. To gain a deeper understanding of how cells function, we need to methodically study gene expression and cell structure in the context of the activities that drive cell behaviors. This will provide a framework with which to untangle and model the plethora of dynamic, interacting components that enable life. Recent initiatives, including one reported on page 820 of this issue by Thul et al. (1), and new methodologies suggest that now is the time to undertake an ambitious challenge: conjoin genomic, epigenetic, and structural studies to create a whole cell atlas representing the full variety of cell types and states in the human body.

The major types of human cells were originally identified and classified by the functions of the tissues in which they reside and by morphological criteria, including cell shape and organization of the defining cytoplasmic organelles. The profiles of expressed genes and proteins support these classifications and are revealing new cellular subtypes. Multiscale structural studies complement these gene-centric insights, uncovering detailed views of the molecular machinery through studies of individual components, their interactions with other molecules, and the posttranslational modifications that regulate them.

These approaches present a formidable, multiscale unification challenge. Until recently, most genomic and proteomic data were generated from large heterogeneous samples, rather than from single cells, often obscuring cell-to-cell differences and frustrating attempts to map cell behaviors onto specific patterns of gene and protein expression (2). This suggests that the current catalog of cell types may be incomplete or even misleading. Furthermore, adequate criteria may not have been developed for either defining and categorizing cell types and the states in which they reside or identifying the rules for their interconversions, to reflect the cell as a complex, multistate system (3).

Structural data, by contrast, are usually derived from either high-resolution studies of purified components or microscopic analyses of small cellular regions. Therefore, there is no accurate view of the interior of a whole cell—where the organelles, molecular machines, and complexes reside and interact, how they change during cellular activities, and how they vary from cell to cell within a tissue. It is also not known what the relationships are between the patterns of gene expression, epigenetic modification, and cell organization during various cellular activities, such as the cell division cycle, cellular differentiation, or in response to environmental effects.

The demonstrated potential of genomic studies to identify expressed genes, regulatory circuits, and normal and pathologic cell types and states has catalyzed the pursuit of developing a human cell atlas that reveals the gene expression profiles for every cell in a healthy human body. The Allen Brain Atlas (4) showed the feasibility and impact of such a genomic atlas, using the mouse and human brains as a single organ system (5). The U.S. National Institutes of Health is using genomics to define cell types through its Brain Research through Advancing Innovative Neurotechnologies (BRAIN) Initiative. The emerging Human Cell Atlas (6) project aspires to bring together an international community, creating comprehensive genetic reference maps of all human cells to understand human health and to diagnose, monitor, and treat disease.

The impressive insights emerging from microscopic analyses of cells suggests addressing a parallel challenge: cellular cartography. For example, the Allen Institute for Cell Science (7) is creating dynamic three-dimensional (3D) multiscale integrated cells that reveal the organization and interactions of cells in tissues; organization of organelles and molecular complexes within and surrounding cells; and molecular-scale structures, posttranslational modifications, and interactions of major cellular components over time. Such a dynamic 3D map could have enormous potential in defining, understanding, and predicting cell types and states, both normal and pathologic, assigning potential function to newly identified proteins, and identifying signatures of distinct cell states and the interconversions among them—through analyses of the morphologies, organization, and dynamics of cells and intracellular structures (8). The project currently analyzes living human induced pluripotent stem cells and genome-edited cell lines, but will also include single-cell genomics studies. The Oregon Health and Science University's Center for Spatial Systems Biomedicine (9) uses a range of microscopy, from electron to light, to study the structural and mechanical features of cancer cells, conjoining them with genomic analyses to determine how altered genes and structure lead to disease, thereby enhancing diagnosis and treatment.

The Human Protein Atlas (10) is a long-standing project that developed a large panel of highly specific and well characterized antibodies to determine where proteins reside in tissues throughout the human body. Thul et al. applied this impressive resource to localize ∼12,000 human proteins to 30 subcellular structures in a set of 22 established human cell lines. This big step forward in cellular cartography creates an enormous and highly valuable community resource—assigning a major fraction of the cellular proteome to specific cellular locations, showing both cell structure and where proteins localize in various organelles.

The recent genomic initiatives and the emerging efforts in cellular cartography suggest that the time is ripe for the community to address a long-term, grand challenge: conjoin genomic, epigenetic, and structural approaches to create a single, comprehensive, dynamic, multiscale, and high-dimensional human cell atlas, representing the major cell types and states, including their interconversions, in the human body. This atlas would also incorporate differentiation and responses to environmental changes and perturbations. It would frame and unify decades of structural, genomic, and cell biologic studies with analyses, theories, and models, giving us the ability to zoom visually and mathematically from body to molecular scales. It would provide a long-term research focus with tangible goals, a resource for hypothesis generation, and enormous insight and practical value along the way.

This effort will require coordination among the current (and future) independent atlas initiatives, the development of standards, and the engagement and integration of individual investigators, studying cellular processes in depth, outside these initiatives. It will also benefit from new technologies and broad multidisciplinary collaborations among bioscientists, physicians, engineers, physical scientists, mathematicians and statisticians, computational scientists, data and information scientists, and scientific illustrators. A challenge of this scale will require a large, collective effort to build resources to characterize cell types and cell states and the controls systems that regulate them. It will further require industry, philanthropic, and government support, as well as leaders who can engage, guide, and empower the broader research community. The recent announcement by the Chan Zuckerberg initiative (11) is a step in this direction, providing support to develop technologies for a whole-body human atlas project that incorporates genomic and image-based components. In the wake of the Human Genome Project, building the human cell atlas envisioned here is an essential and timely goal for 21st-century cell biology.

References and Notes

  1. Acknowledgments: We thank P. G. Allen for his vision, encouragement, and support. We also thank our colleagues for suggestions and insights.
View Abstract

Stay Connected to Science

Navigate This Article