Software developers receive $4 million for cancer research

IU’s National Center for Genome Analysis Support received a $4 million RNA research software grant.

The software, called Trinity, processes the measurements of different types of gene levels.

The grant is specifically intended to improve the ability of Trinity to assemble graphs that show the presence of prospective cancer cells.

“By optimizing Trinity, researchers can now do four times as much science,” NCGAS Director Bill Barnett said in a press release. 

Trinity is currently used by IU, the Broad Institute at the Massachusetts Institute of Technology and Harvard University — who helped develop it — and the Center for Information Services and High Performance Computing at Technische University in Dresden, Germany.

The software processes high-quality RNA sequence assemblies, a code scientists use to determine how active certain genes are in different tissues, which can hint at how likely cancer is in the area being studied. Scientists can use this data to discover which cells to target and watch out for.

Given by the National Cancer Institute, the grant will pay to optimize aspects of the software like coding, memory and data storage.

Ultimately, the improved research could lead to more effective cancer treatments.

But Robert Henschel, manager of Scientific Applications and Performance Tuning at NCGAS, said the memory needed to run Trinity demands a lot, and the cost to provide it is high.

A transcriptome is the small portion of the genetic code that is transcribed into RNA molecules, and it consists of less than five percent of a human’s DNA code.

The amount of memory it takes to put together a transcriptome is too big for a normal computer, Henschel said.

Some of the grant money will buy two new servers to run Trinity and to help take on some of the burden.

“Trinity requires very large amounts of main memory to assemble a transcriptome, thus requiring special servers,” Henschel said. “If we can reduce the amount of memory needed during computation, we can allow more people to run Trinity, broadening the user base of the software.”

With the new servers, more researchers will be able to measure more
information.

But despite the large figure the grant boasts, Henschel said the portion of the money IU will receive pays for little more than one person to work on making Trinity more
efficient.

It’s not a small workload, he said.

“Our optimization work involves understanding the computational performance of Trinity across the runtime of the various components that make up a complete Trinity execution,” he said. “We will be looking at all the different components, understanding their runtime behavior and devising ways to make them run faster or, for example, consume less memory.”

The rest of the software optimization will be done by the Broad Institute and the Center for Information Services and High Performance Computing.

The grant comes with a five-year benchmark plan, where refinement of the computer program will be checked each year.

The software has already increased four-fold in speed since 2011, according to the release.

But even after the increase in speed, Henschel said it can still take Trinity longer than a week to produce a transcriptome.

More timely results and a smoother run will happen over the next five years as developers work to make Trinity more efficient.

“In this process, we mostly rely on best practices in the field of software engineering and development of scientific applications,” Henschel said.

Researchers are hopeful cancer care will see a boost in efficiency after the software is improved.

“It’s important that the result will be in advancing therapies for treating cancer,” Barnett said.

Follow reporter Ashley Jenkins on Twitter @ashley_morga.