Processing Genomic Sequences in Minutes

A new computational tool can divide the work of analyzing biologic data into manageable fragments, so that large data sets can run on many processors simultaneously, solving large-scale problems in minutes, rather than weeks.

Called ScalaBLAST, the tool was developed at the U.S. Department of Energy's Pacific Northwest National Laboratory (PNNL, Richland, WA, USA; www.pnl.gov). PNNL's supercomputer has 1,960 processors, lots of memory, and the ability to tackle large problems. However, without special modifications, software does not run any faster on this computer than it would on a personal computer. In order to speed it up and get answers to complicated biologic questions more quickly, PNNL researchers "parallelized” the software, using Global Arrays, a powerful programming toolkit, by creating algorithms to divide up the work.

PNNL researchers state that ScalaBLAST may be used to process complex genomic sequences, essential to understanding the building blocks of the genome. Now they can manage the large influx of data resulting from new questions that arise during human genome research. Prior to this tool, it took researchers 10 days to analyze one organism. Now, they can analyze 13 organisms within nine hours, making the time-to-solution hundreds of times faster.

"Access to and understanding the pieces of genome sequences will allow researchers to understand the body's cellular machinery and discover clues to some types of cancer,” observed T.P. Straatsma, a PNNL senior research scientist. "And it will help in developing drugs or detection methods to be used for particular diseases.”






Related Links:
PNNL