Human Cell Morphology Atlas Links Genes and Traits
Using a novel technique for genome-wide imaging screens scientists are creating perturbation atlases of cell morphology.
Complete the form below to unlock access to ALL audio articles.
Visualizing a cell after editing a specific gene can reveal insights into the gene's function. This, in turn, allows scientists to link genes to known disease-associated pathways, which can reveal novel therapeutic targets. One of the key challenges of the modern genomics era is studying the thousands of genes in a single human cell at scale.
To overcome this, Dr. J.T. Neal, institute scientist and principal investigator at the Broad Institute of MIT & Harvard, and colleagues have developed PERISCOPE (perturbation effect readout in situ via single-cell optical phenotyping), which combines the power of microscopy imaging with genome-scale CRISPR screens.
The approach utilizes the Cell Painting assay, in which small-molecule dyes are used to label cellular compartments that are then imaged in fluorescence channels. Machine-learning models then detect subtle changes in the images, such as changes to the size of the nucleus.
Cell Painting is combined with optical pooled screening in PERISCOPE, which “barcodes” cells and uses CRISPR to systematically turn off genes in those cells.
Using this approach, Neal et al. were able to produce the first open-source genome-wide perturbation atlas of cell morphology in human cells. The atlas contains 3 genome-wide genotype-phenotype maps comprising CRISPR-Cas9-based knockouts of >20,000 genes in >30 million cells.
What is cell morphology?
Cell morphology describes the shape, size, structure and appearance of a cell, including the arrangement of organelles within the cell. The morphology of a cell is a rich source of information about cell function in health and disease.
“Large-scale DNA sequencing has revolutionized our ability to catalog human genetic diversity, but it has also led to a new challenge: understanding how these genetic differences affect human biology,” Neal told Technology Networks. “Systematically linking genes and DNA variants to diseases and traits is still a significant challenge in biomedical research. Our aim was to build an approach capable of tackling both the complexity and the scale of this problem.”
The atlas and screening platform represent a rich resource for connecting genes to cellular functions, which holds potential in drug discovery and could be used to assess how genetic variants drive responses to drugs.
“Perturbing genes and studying the associated changes in morphology allows us to link genes to known disease-associated pathways and processes. PERISCOPE allows us to do this at genome-scale routinely,” explained Neal. “Because PERISCOPE allows us to study the effects of many genes or gene variants simultaneously, we can also use it to carry out very large experiments in which we assess how genetic variants drive responses to drugs. This holds immense potential for drug discovery efforts.”
The researchers believe the accessibility and proposed cost-effectiveness of PERISCOPE-style screens position it as a democratizing platform for linking genotypes to cellular phenotypes. “Notably, the cost is remarkably low per-cell profile: ~US$0.001 per cell for the described HeLa datasets (including labor, materials and analysis, but not equipment),” the researchers stated.
Interrogating poorly characterized genes
In PERISCOPE, a library of guide RNAs are introduced into the cells. Next, expression of the Cas9 enzyme is induced to disable the guide-targeted genes. The guide RNAs are then converted to complementary DNA, creating “barcodes” of the gene knocked out in each cell. This barcoding also enables the study of many genes in a single batch of cells. Finally, a standard widefield microscope is used to record images of the Cell Painting stains and the four-color barcodes in the cells, followed by automated image analysis to extract cell features and link them to guide RNAs.
Using PERISCOPE to create atlases of the effects of knocking out genes in human lung cancer cells and HeLa cells cultured either in traditional cell culture medium or physiologic medium illustrated known biology but also revealed new information about poorly characterized genes.
“Using our morphological atlas, we were able to discover the function of TMEM251, a previously uncharacterized gene, demonstrating that it is essential for directing enzymes to lysosomes. This is an exciting proof-of-principle and shows the immense discovery potential of this type of screen,” said Neal.
The TMEM251 gene has previously been linked to distinct lysosome storage diseases, however its function has remained elusive. The researcher's findings now validate a role for TMEM251 in lysosomal protein trafficking through the mannose-6-phosphate system.
The team is now working on expanding the range of traits that PERISCOPE can capture as well as working with collaborators to create perturbation atlases in other cell types.
Neal concluded, “The genetic perturbation information in the atlas can be used by anyone for assessing the function of their favorite genes. We have also shared all of the images and their associated metadata and these can be used to train new machine learning algorithms. This is just the beginning, though, and we are continuing to advance the PERISCOPE platform by increasing its sensitivity and scale, and by expanding it into a much larger range of disease-associated cell types.”
Reference: Ramezani M, Weisbart E, Bauman J, et al. A genome-wide atlas of human cell morphology. Nat Methods. 2025;22(3):621-633. doi: 10.1038/s41592-024-02537-7
