Skip to main content
Premium Trial:

Request an Annual Quote

HMS Team Develops Data-Analysis Approach For High-Content Genome-Wide RNAi Screens

Premium
Investigators at Harvard Medical School, including Norbert Perrimon and scientists in his laboratory, and Columbia University recently said they have developed a framework that enables them to automatically recognize cellular phenotypes in the context of the Rho family of small GTPases.
 
The method, which has components to segment and analyze microscopic images, uses RNA interference to perturb the function of genes involved in Rac signaling.
 
The investigators generated data in the form of high-content, three-channel, fluorescence microscopy images of Drosophila Kc167 cultured cells stained with markers that enabled them to visualize DNA, polymerized actin filaments, and the constitutively activated Rho protein RacV12.
 
They validated the approach using a database that contained more than 1,000 samples of three predefined cellular phenotypes and relied on a cross-validation technique to obtain a generalization error rate.    
 
The researchers also used this approach to analyze whole high-content fluorescence images of Drosophila cells for further HCS-based gene-function analysis.
 
Their work was published last month in the Journal of Biomolecular Screening.
 
Stephen T.C. Wong, who is currently director of the Center for Biotechnology and Informatics at The Methodist Hospital Research Institute in Houston, Texas, and corresponding author on the paper, spoke with CBA News this week about how this approach can be applied to drug discovery and human cells.
 

 
Can you give me a little background on this work?
 
This is one of the first papers ever to discuss how you can automatically segment, classify, and even discover new phenotypes using genome-wide RNAi screening.
 
This is really exciting, because genome-wide screening of RNAi assays is now more feasible. In the past, these assays generated so many images, researchers did not know how to analyze them.
 
This paper discusses the automation of the segmentation and classification of these cellular images for the phenotyping of the cell-based assays.
 
How is this technology integrated into drug discovery?
 
Once you can do genome-wide RNAi screening, you can do a whole-genome screen to understand the particular mechanisms of the disease in which you are interested. You do not have to guess.
 
In this particular case, I am interested in a specific gene family, the Rho. This relates to the migration of cells. It is cancer related.
 
This assay is very important to making the next generation of cancer drugs. And we will be able to discover a lot of things that people could not discover in the past.
 
You did your work on Drosophila cells. Would this technology also work with other kinds of cells, such as human cells?
 
Yes. This is really a general pipeline; it can be applied to many different types of cells, including human cells. The design is really modular.
 
What do you see as the next step in this research?
 
I am interested in cancer. The next step is to create a pathway network. I can make a systems-biology model for these mechanisms in this signaling pathway.
 
The other thing is to try different kinds of cancer cells, such as prostate cancer or breast cancer, because they really have the same mechanism.
 
We can try mammalian cells as well. We have a project ongoing now where we are screening some cancer stem cells, and we will be applying this technology to genome-wide RNAi screening of cancer stem cells.    
 
Is this technology something you would be interested in commercializing?
 
I was not thinking about that because I am an academic. If people are interested in commercializing it, I am happy to license the technology to them. I am not interested in making money at the moment. I am more interested in the science.
 
How would you characterize the significance of this work?
 
This work is opening up new opportunities in genome-wide RNAi screening or related drug assays, and the powerful integration of imaging tools with more genome-wide "-omic" approaches to analyze the biological circuits of organisms and biological mechanisms of disease, in this case, cancer metastasis.