Skip to main content
Premium Trial:

Request an Annual Quote

Method Developed to Glean Genetic Interactions from Cellular Morphology and Genetic Screen Data

By a GenomeWeb staff reporter

NEW YORK (GenomeWeb News) – Researchers from the Massachusetts Institute of Technology and Harvard University have developed a computational method for discerning genetic interactions by imaging the morphology of individual cells that are being tested with a genetic screen.

The researchers came up with a computational method for translating morphological data from cell imaging experiments into information on genetic interactions in individual Drosophila cells being screened using RNA interference. They then used this approach to investigate the interactions affecting a signaling pathway involving the small GTPase enzyme Rho. The research appears online today in Genome Research.

"These images are an enormous source of data that is only beginning to be tapped," senior author Bonnie Berger, an applied mathematics professor and head of computation and biology at MIT, said in a statement. "We realized we had enough data to go beyond classification and start to understand the mechanism behind the differences in shape."

Enzymes participating in cellular signaling networks often act as so-called molecular switches, the team explained, activating or inhibiting other proteins within the pathway. While targeted biochemical, genetic, and cell biology-based approaches have uncovered the basic arrangements of these networks, they added, a refined view of many signaling pathways remains elusive.

For the current study, the team came up with a high-throughput approach for getting hints about signaling interactions based on morphological traits, such as cell shape, DNA characteristics, or the localization of proteins and/or organelles in the cell.

These image-based approaches offer an indirect picture of what's happening when specific genes are mutated in — or missing from — the cell, the team explained, and offer an opportunity to observe the cellular consequences of these changes.

"[N]o successful method, to our knowledge, has been developed for systematically identifying genetic interactions or predicting signaling relationships using image-based data from high-throughput screens," Berger and her co-authors wrote.

Berger and her team used a combination of RNAi targeting 13 RhoGAPs and single cell imaging to investigate the relationships between different genes and proteins in the RhoGAP/GTPase pathway in a Drosophila cell line called BG-2. Past studies suggest the pathway is involved in everything from adhesion between cells to cell motility.

The team first developed a classification model to categorize cellular features reflecting up- or downstream changes to the Rho pathway from cell imaging data collected under a variety of treatment conditions, measuring 145 geometric and nine status features for each cell. They then developed a computational method for integrating single-cell morphological data with RNAi screening information.

In the process, the researchers were able to pinpoint genetic interactions involving RhoGAP/GTPase signaling pathway components. And by combining data generated by knocking down individual genes and pairs of genes, the team was able to refine their view of so-called within- and between-pathway interactions.

Those involved noted that more research is needed to validate the findings of their high-throughput findings — particularly to rule out false positive interactions between genes.

Still, although their prior knowledge of some features within the Rho network aided the high-throughput approach, they say a similar method may also be useful for discerning relationships between proteins in other pathways as well.

"This work demonstrates the fundamental fact that high-throughput morphological data can be used in a systematic, successful fashion to identify genetic interactions and, using additional elementary knowledge of network structure, to infer signaling relations," the researchers wrote. "[T]he computational framework presented here represents an initial approach to the problem that will serve as a basis for future enhancements."