BERKELEY, Calif. – Researchers showed progress accelerating the search for gene-based cures for cancer and expanding the field of computer theory at an annual event sponsored by the University of California at Berkeley. Computer scientist David A. Patterson called for a million genome warehouse to advance work on a cure for cancer. Today separate repositories hold less than 10,000 pieces of genetic information, many of them only partial representations of genes. "There’s a chance for computer science to help build fast and accurate genetic pipelines and accelerate the move to personalized therapies--I want this in time to help me and my family," he said, noting researchers today often delete genetic data after completing experiments. Patterson helped develop a tool called SNAP that provides significantly faster and more accurate genetic analysis that tools typically used by cancer researchers today. Benchmarking tools are still needed to improve what are still highly subjective methods used in the field, he said.
Patterson aims to apply big data tools under development at Berkeley to the challenge of analyzing reams of genomic data on cancer.
The tools include Spark, a programming language for computer clusters that provides functions similar to Map Reduce used by Google’s search engine. The Berkeley Data Analytics System is an open source engine in development based on Spark.