It takes some imagination to grasp the raw computational challenge researchers face trying to understand the molecular causes of cancer. John Weinstein, M.D., Ph.D., professor and chair of M. D. Anderson's Department of Bioinformatics and Computational Biology, likes to start with this imagery."If you unpacked the DNA in every cell of a single person and stretched it end to end, it would circle the equator 917,000 times -- the equivalent of 120 round trips to the sun. One error in replicating the genome in one unlucky place -- over a length of 120 trips to the sun and back -- can lead to cancer. Our challenges are to understand how that happens, and to know what to do about it if we can't prevent it in the first place."
Now, consider that there are usually many more errors spread across that expanse and that they tend to interact with each other in a maze of complexity. Add current techniques to profile genetic expression and variation that capture a galaxy of data, and you have enough information to choke traditional methods of analysis.
The Cancer Genome Atlas, a federally funded project to make sense of the genetics that drive 20 different cancers, has enlisted Weinstein and colleagues at M. D. Anderson to help with this problem. A five-year, $8.3 million grant will allow the team to apply cutting-edge approaches to find the right buckets of information in an ocean of data.
They will blend agile software development, Bayesian statistics, and a flexible and efficient database infrastructure called semantic web with M. D. Anderson's expertise in clinical and translational research to compile the most meaningful data in a tumor's tangled molecular profile.
Success will mean better treatment choices, improved risk assessment, diagnosis and prognosis. "The bottom line is personalizing cancer medicine," Weinstein says.
Leave a comment