cellxgene: Using Cell Visualization to Help Scientists Better Understand Complex Datasets in Health and Disease

Apr 28, 2020

Brushes of rainbow color in shades of blue, yellow and purple radiate outward from a white background, representing tissues colored by category across cells.
A visualization in cellxgene with coloring by tissue-type category across cells.

The information gathered from biomedical research can include terabytes of data that are difficult and time-consuming to analyze. Cellxgene is a tool that helps scientists to explore and visualize high dimensional single-cell datasets in an interactive way, allowing them to surface important information that could lead to discoveries in treating disease.

This easy-to-use, open source cell visualization tool integrates with existing tools like SciPy and Jupyter. Scientists can visualize gene expression at the level of an entire dataset or a particular subset of cells to help identify cell types, which can in turn help identify diseases. cellxgene allows researchers to keep track of and annotate cell types as they identify them in their dataset, and the tool also enables plotting one gene’s expression level against another gene to compare how these different genes are expressed across a dataset.

Researchers can view metadata, such as which cells came from a certain tissue type, and then color-code these attributes to display more context about the cells they’re viewing. Computational biologists and experimental scientists use cellxgene to explore information from the global Human Cell Atlas project, like a single-cell atlas of the human lung that is helping researchers understand COVID-19 transmission. Many institutions, such as the Chan Zuckerberg Biohub, Cincinnati Children’s Hospital, Novartis, and Johns Hopkins University, have set up their own public sites featuring their data within cellxgene.

Visit

chanzuckerberg.github.io/cellxgene starfish Building a Data Coordination Platform