Experts who utilize gene databases on the Internet face the same type of problems that greet many Web surfers, so some of them have designed an easier and more accurate way to search for potential disease-causing genes. (PNAS)
15th July, 2002
Ludwig Institute for Cancer Research
With the click of a mouse, researchers can quickly analyze information from thousands of genes stored in special databases over the Internet. But experts who utilize these databases face the same type of problems that greet many Web surfers, such as sites that are confusing and hard to navigate--or worse, information that is incomplete or non-existent.
To overcome these shortcomings, a group of researchers have designed a Web site that provides an easier and more accurate way to search for potential disease-causing genes. The new site, known as SAGE Genie, is available online (http//cgap.nci.nih.gov/SAGE) as part of the Cancer Genome Anatomy Project run by the National Cancer Institute (NCI).
"Now that most of the human genome has been sequenced, it is important to accurately determine exactly where these genes are expressed, but this information is often hard to use or not available," said Dr. Gregory Riggins of Duke University Medical Center, the lead author of a study on SAGE Genie that is to published in the Proceedings of the National Academy of Sciences. "We have simply improved a good way to study genes and made it available to everyone online."
The improvements are based on highly respected technique for studying genes called serial analysis of gene expression, or SAGE, developed at Johns Hopkins. Although widely considered to be the most in-depth human gene-sequence profiling technique, even the most experienced gene hunters can have trouble navigating existing SAGE databases.
Researchers made several improvements as part of the upgrade. At Duke, a team led by Dr. Riggins went through nearly 7 million SAGE transcription tags to see which ones were the most reliable. Investigators at the Ludwig Institute for Cancer Research then developed a scoring method to alert users whether the tags were of high-quality or if they contained potential flaws.
Both teams helped to offset further errors, such as accounting for incomplete gene transcripts that can occur as experimental artifacts. Finally, specialists at the NCI developed a new site that incorporated these additions, providing a simplified interface that matches a particular transcript to where it is found in the body.
"This is very straight forward and user friendly," said Dr. Sandro de
Souza of the Ludwig Institute for Cancer Research, Sao Paulo Branch. "It
helps a large variety of scientists who want to check gene expression patterns
to find a disease marker or target for treatment."
© 1995-2002 Newswise