SOURCE: The Genomics Institute of the Novartis Research Foundation

July 14, 2005 17:00 ET

Scientists Adapt Facial Recognition Technology to Scan the Genome for Cancer Genes

LA JOLLA, CA -- (MARKET WIRE) -- July 14, 2005 -- A team of scientists from La Jolla, California, led by researchers at GNF, have for the first time successfully combined microscopic cellular analysis with the genome-wide interrogation of individual gene function to find biological causes of cancer.

A common feature of all cancers is the uncontrolled growth of diseased cells. Despite the completion of the sequencing of the genome, scientists still do not know the full set of genes which may contribute to cancer growth. Recent years have seen the development of technologies that can be used to determine the function of individual genes, such as promoting cell growth, across entire genomes. One of these technologies, pioneered at GNF, is high throughput genetic gain-of-function screening. In this process the genome is broken down into its individual component genes, which are then systematically inserted into cells using a unique automated system designed for this purpose.

Microscopic analysis has traditionally been used to assess the functional effects of individual genes in cells, and scientists have had to rely upon observations of a trained eye to gather their data. Experiments across a genome would require functional gene expression followed by acquiring and accurately analyzing tens of thousands of microscopic cellular images in a short space of time with unerring accuracy. "This type of experiment has so far been out of the reach of scientists, which has limited the scope of what we can decipher from our knowledge of the genome, and hence how we can translate that knowledge into cures," commented Dr. Sumit Chanda, Group Leader at GNF and the lead author of the study published in Genome Research.

Now, through the automated acquisition of microscopic cellular images and the application of sophisticated image analysis methods, scientists from the Genomics Institute of the Novartis Research Foundation (GNF), the Scripps Research Institute, Beckman Coulter, and Vala Sciences, have demonstrated that it is possible to scan approximately 1/3rd of the genome in a single experiment to identify genes which cause significant increases in the growth of human cells. Using a process known as high-content imaging, which employs pattern recognition and image analysis algorithms adapted from facial recognition technology, tens of thousands of microscopic images of cells were analyzed to pinpoint genes in the human genome which may play a role in cancer -- without any human intervention or bias. "Essentially, we let the software tell us which of the 7,000 genes contained in our library are likely candidates to be cancer promoting genes, "oncogenes," which we then experimentally validate. By bringing microscopy to the field of high throughput functional genomics, we now have an extraordinary opportunity to systematically understand the role of complete human genome in any given disease, at the level of the individual cell," adds Dr. Chanda.

The study, entitled "Identification of Novel Mammalian Growth Regulatory Factors by Genome Scale Quantitative Image Analysis," is authored by Josephine N. Harada, Kristen E. Bower, Anthony P. Orth, Scott Callaway, Christian G. Nelson, Casey Laris, John B. Hogenesch, Peter K. Vogt, and Sumit K. Chanda. It appears online ( as a "Genome Research in Advance" paper on Friday, July 15, 2005, and will also be published in the August 2nd print issue of Genome Research.

About GNF:

The Genomics Institute of the Novartis Research Foundation (GNF; has integrated state-of-the-art technologies, enabling new approaches to complex biomedical problems in life science research. These technologies include genomics and proteomics tools, combinatorial chemistry, structural genomics and computational biology. The mission of the Institute is to exploit these technologies to identify new biological processes and understand the underlying mechanisms involved in human disease.

Contact Information