Friday, May 18, 2012

Scientists use search algorithm for cancer studies

Google uses their Pagerank technology to rank the websites that turn up in its search engine. A higher page rank means a website is more relevant, and it should therefore end up higher in the search results. This technology is used by cancer researchers to uncover proteins that could play an important role in detecting the presence and progression of tumours. Such tools should help clinicians by deciding what kind of treatment their patients need. Previously, scientists already showed that computer simulation can help create new cancer drugs.

Using their algorithm, the scientists scanned around 20.000 genes, all coding for individual proteins, and tried to correlate them with clinical outcome in pancreatic cancer. That makes these proteins function as biomarkers: measuring them signifies whether a patient is likely to have cancer, or whether it progresses aggressively. Scientists have already found large numbers of biomarkers for a variety of diseases, but the Google-like search algorithm is set to make their predictive value higher. It not only uses information about the protein itself, but also about its interactions. Basically, that means it does not only look at its content, but also at the links it is making, to put it in web terms.
The algorithm provided the scientists with a likelihood that a certain protein holds a predictive value for the aggressiveness of pancreatic cancer. After looking at the most promising ones, they found a match with a previous study that also found biomarkers for predicting how aggressive the disease is. It confirms that their tool is working, although more data is required to actually prove it is of use in the clinic. Nevertheless, the data does show that using a Google-like algorithm that combines protein information and its network can improve biomarker detection.

Computers keep on getting better, and medical biology also benefits from that. It allows us to investigate complex interactions, such as the one in the present study: it is impossible to analyze 20.000 proteins and their interactions by hand. Using computer algorithms, such as the one Google has developed for their search business, can help make things easier. Improving these tools means we can determine the faith of a patient with more accuracy, although human biology will remain full of surprises. 

No comments:

Post a Comment