(RxWiki News) The Google algorithm for searching and prioritizing pages on the Internet is constantly changing. Researchers have developed their own version of this coveted equation and put it to work to find new cancer answers.
Scientists from Dresden University of Technology in Germany have used a customized take on Google's PageRank algorithm to rank the importance of some 20,000 proteins for their genetic involvement in pancreatic cancer.
The massive sifting has led researchers to identify seven proteins they think can help determine the course of the disease and assist clinicians in making treatment decisions.
"Learn about all your treatment options."
Biomarkers are molecules that cancer (and other diseased) cells make.
Biomarkers can indicate the presence of cancer very early on when they show up in the blood and other body fluids.
Zeroing in on these proteins has been extremely difficult and time intensive.
The investigators adopted the Google strategy of identifying both a web page's content and how it's linked to other pages. The more links in - and out - the more important the protein.
Next, the researchers created what they call a "protein Facebook" to map out how the proteins are connected to each other through various physical and functional interactions.
"Once we added the network information in our analysis, our biomarkers became more reproducible," said the paper's first author, Christof Winter.
They used the network information and then the algorithm to see how the data overlapped with an earlier University of North Carolina study.
In this manner, researchers were able to connect with proteins that can assess the aggressiveness of pancreatic cancer.
Knowing how likely the disease is to grow and spread can help doctors figure out whether or not the person will benefit from chemotherapy.
While this marks an improvement over existing tools, the new biomarkers aren't perfect by a long shot. They still need to be validated in larger studies before they can be used on patients.
Once that happens, the next step would be to develop drugs that could target these proteins. The research team is currently working with a German biotech firm, RESprotect which is conducting a clinical trial on a potential pancreatic cancer drug.
This study was published May 16, 2012 in PLoS Computational Biology.
The research was supported by Roland Ernst Stiftung für Gesundheitswesen, MeDDrive TU Dresden, the EU project Ponte and the national projects GoOn, Format, and GeneCloud.
The authors have declared no competing financial interests exist.