Streamlining a Solution

When it comes to research involving computational biology algorithms, researchers can face the same set of obstacles as those conducting biological, chemical, and physical experiments: complex and inconsistent parameters, including protocols and outcome interpretations. Raja Mazumder, Ph.D., associate professor of biochemistry and molecular medicine at the GW School of Medicine and Health Sciences, however, is working to streamline that system to ensure that the regulatory sciences industry has unified standards to simplify the replication of experiments and confirmation of the results.

Mazumder’s solution is based on “biocompute objects,” which he describes as “human-readable or machine-readable.” Biocompute objects are records that include various data – software arguments and version information, references to all outputs, associated metadata of executable programs, a usability domain, validation kits – that allow for consistent and accurate computation evaluations, particularly for research submitted to the FDA.

“When somebody says ‘I have this amazing diagnostic for breast cancer, and I’m using some software to do this after the sequencing is done,’ right now there is no mechanism to communicate all of the computational steps to the FDA in a clear and consistent manner,” Mazumder explained. “We proposed to the FDA to develop a standardized biocompute object, which will allow this communication. It has far-reaching potential.”

This March, Mazumder will look to the scientific community for feedback during a public workshop held at the National Institutes of Health. “[T]his is not a new problem, but we are the first to talk about this from a regulatory science viewpoint. Everyone does research but when you’re trying to get something to the bedside, that’s a whole new ballgame,” Mazumder said. “We have some ideas, but we don’t want to just come up with this and push it down on everybody. If there’s no agreement, it’s not going to fly.”

At the workshop, participants will discuss aspects of HTS and relevant workflows, with a goal of creating a set of recommendations for the FDA on how to report computations.

For more information on the HTS-CSRS workshop, which will take place on March 16 and 17 at the NIH Porter Conference Center in Bethesda, Maryland, visit https://hive.biochemistry.gwu.edu/htscsrs/workshop_2017.