Cancer’s Signature—Written in Blood

When it comes to deciphering the health of the body, the blood carries a potential mother lode of protein clues. Given the ease of extracting blood, such proteins could serve as efficient health barometers. But it’s tough to distinguish between the multitude of proteins naturally found in blood and those that are secreted into the blood—including those secreted by diseased tissue such as cancer. Their signal may get swamped by the many other proteins present in blood, thwarting efforts to discover useful information. Now, scientists have developed an algorithm that sorts through the multitude, expediting the search for blood-based cancer biomarkers.


“Figuring out which proteins are secreted into the blood is like searching for a needle in a big, big haystack,” says Ying Xu, PhD, professor of bioinformatics and computational biology at the University of Georgia. “This [algorithm] sorts through all that hay.”


To develop their algorithm, Xu and his colleagues began by scouring the literature for all proteins known to be secreted into the blood, regardless of their origins. They then analyzed the amino-acid sequences of these proteins to identify common features, such as signal peptides, transmembrane domains, solubility, and secondary structure. They discovered 18 features that were powerful predictors of blood secretion, and used them to train a computerized classifier.

This microarray shows genes that differ in regulation between cancerous and non- cancerous lung tissue. Ying Xu’s classifier can predict which of the proteins made by these genes may be useful as blood-based bio- markers. Courtesy of Ying Xu.


When the researchers applied the classifier to other data sets, it could distinguish proteins secreted into the blood from all other proteins in the blood with more than 80 percent accuracy. The results appear in the October 2008 issue of Bioinformatics.


Xu and his colleagues are now using microarrays to identify differences in gene expression levels between cancerous and non-cancerous stomach tissue. Using their classifier, they can then sift through the data to zero in on genes that produce proteins that are most likely to be secreted into the blood, followed by validation with mass spectrometry.


“We’ve already identified proteins that are elevated during different stages of stomach cancer,” Xu says. “Typically, in order to find out what stage it’s in, you’d have to actually cut the patients open and do a biopsy. Our markers could be the first markers to provide information about cancer stage.”


By applying his biomarker discovery pipeline to a range of cancers, Xu ultimately hopes to identify general biomarkers that apply to any cancer. He envisions doctors detecting various cancers at early stages with a simple blood test.


Bo Huang, PhD, a post-doctoral fellow at Vanderbilt University, hopes to use Xu’s classifier to find biomarkers for breast cancer. “These results provide a powerful method to discover potential biomarkers, not only for cancers but also for many other diseases,” Huang says.

Post new comment

The content of this field is kept private and will not be shown publicly.
This question is for testing whether you are a human visitor and to prevent automated spam submissions.
Enter the characters shown in the image.