Home >> ALL ISSUES >> 2014 Issues >> NGS informatics catching up to clinical demands

NGS informatics catching up to clinical demands

image_pdfCreate PDF

William Check, PhD

November 2014—When Birgit H. Funke, PhD, gave a talk earlier this year on incorporating bioinformatic tools and pipelines into medical NGS, at Molecular Medicine Tri-Con 2014, one of her slides showed the main bioinformatics activities needed to support sequencing. Among them were designing and building pipelines to manage genetic data, writing scripts for data analysis pipelines, and building custom applications.

But the point she emphasized most was “Clinical add-on: documentation + validation.” Dr. Funke, who is assistant professor of pathology at Massachusetts General Hospital/Harvard Medical School and director of clinical research and development for the Laboratory for Molecular Medicine at Partners HealthCare, said that working with an awareness of clinical application is new for bioinformaticians. In an interview with CAP TODAY, she illustrates this with a brief dialogue she had experienced:

Lab director: “Where did you store that script?”

Bioinformatician: “I don’t know, but I can rewrite it for you.”

Dr. Funke

Dr. Funke

While this attitude is all right for research work, it just won’t cut it in the world of CAP-accredited, CLIA-certified laboratory testing. “We have to validate everything, and that includes scripts,” Dr. Funke says. “Building a team of clinical bioinformaticians is a painful process. It is almost a new discipline.” It takes about a year, she estimates, for a bioinformatician to get on the same page with the rest of the laboratory.

“We need individuals who know how to code but who also understand genetics and the rigor of clinical lab work,” Dr. Funke adds. One big reason is that there is not much software designed for clinical next-generation sequencing. “I have yet to come across a lab that did not have to integrate those programs into their infrastructure,” she says. Bioinformaticians are equipped to do this. “It is critical to have at least one bioinformatician dedicated to your clinical analysis. There are more such people than before,” she says, “but they are still hard to come by.”

Why such an emphasis on trained bioinformaticians? Certainly in her Tri-Con talk, and in a webinar she gave in July in the “NGS 101 for the Clinic” series for the Association for Molecular Pathology, Dr. Funke described in detail the technical side of NGS and its associated informatics pipeline. But she said in the webinar: “The wet lab steps are the least of your problems. The more tricky issues surround bioinformatics analysis.” And it is for the latter that bioinformaticians are needed.

Dr. Nagarajan

Dr. Nagarajan

At Washington University School of Medicine, Rakesh Nagarajan, MD, PhD, and his colleagues also recognize that skilled, clinically aware bioinformaticians are essential for clinical NGS. “All of our bioinformaticians started as master’s-level personnel at the university’s Center for Biomedical Informatics,” which he directs, Dr. Nagarajan tells CAP TODAY. After a classical training in bioinformatics, they received training in NGS. They were then further trained in clinical concepts in the clinical lab and clinical IT environment. Traditional bioinformaticians are aware of pipelines and details and can tweak parameters in the research arena to see different outcomes, says Dr. Nagarajan, associate professor of pathology and immunology and of genetics. “In contrast, in the clinical arena we have one shot with patient data on a clinically validated pipeline, and that can’t be changed in an ad hoc fashion. We had to drill those concepts into them.”

There is a need, too, for pathologists with an understanding of bioinformatics pipelines. “There is a need for folks like myself who hacked away and trained ourselves,” Dr. Nagarajan says. “We kind of paved the road when there wasn’t one. In these times, we need interdisciplinary teams, including liaisons, who can speak across disciplines and can direct and train bioinformaticians and train clinicians and genomicists in the caveats of bioinformatics tools.” While he and his contemporaries essentially trained themselves, he says, “In the future there need to be formal training programs.”

CAP TODAY
X