BIG DATA

The National Center for Genome Analysis Support (NCGAS) at Indiana University has received a joint five-year, $4 million resource improvement grant from the National Cancer Institute. The grant is a partnership with the Broad Institute of MIT and Harvard, the lead institution, and the Center for Information Services and High Performance Computing (ZIH) at Technische Universität in Dresden. The partnership’s goal is to advance cancer research.

NCGAS, along with ZIH, will continue to optimize key software and make it available to the cancer research community on IU cluster supercomputers, NCGAS Director Bill Barnett said. Researchers at the Broad Institute and Hebrew University developed the software, known as Trinity. Used by researchers studying transcriptomes (sets of all ribonucleic acid, or RNA, molecules) of model and non-model organisms, Trinity produces high-quality transcript reconstructions. These assemblies allow scientists to identify and study genes that are active within any living creature, with or without having a sequenced genome.

Cancer transcriptomes often have rearranged genomes and harbor transcripts that are not encoded in a reference genome, like those derived from tumor viruses or microbes.  This funding will ensure cancer researchers have more comprehensive assemblies of cancer transcriptomes, as well as access to bioinformatics resources to rapidly obtain those transcriptome assemblies and related findings. The goal is for these expanded research capabilities to lead to improved understandings of cancer biology and ultimately to more effective cancer treatments.

So far, Trinity has quadrupled in speed since the collaboration began in 2011. “By optimizing Trinity, researchers can now do four times as much science,” Barnett said. “Broad continues to improve their software, and Indiana will continue to optimize it to make it run better.”

The grant has annual benchmarks over the course of the five years to ensure continual refinement. “It will be a constant cycle of code improvement, optimization, then delivering new software for researchers to use on systems we support for them,” notes Barnett.

Code optimization will include improvements to runtime performance, memory usage, CPU usage, and storage usage, particularly on High Performance Computing (HPC) machines, said Robert Henschel, manager of the NCGAS Scientific Applications and Performance Tuning group. “We keep the quality of the software high while being more efficient using computational resources, like supercomputers,” Henschel said.

Cancer research is new territory for NCGAS, Barnett added. “This grant is important for us because it allows us to serve a community that we haven’t been able to before,” Barnett said. “And it allows us to continue working with the premier research institutions in the country. IU and the Broad Institute are leaders who can collaborate and leverage each other’s expertise. And it’s important that the result will be in advancing therapies for treating cancer.”