Leaders in Pharmaceutical Business Intelligence (LPBI) Group

The Need for an Informatics Solution in Translational Medicine

Curator: Larry H. Bernstein, MD, FCAP

White paper

http://www.perkinelmer.com/PDFs/downloads/PerkinElmer-White-Paper-The-Need-for-an-Informatics-Solution-in-Translational-Medicine.pdf

Informatics Designed for the Translational Scientist Developing treatments that take individual variability into account (“personalized medicine”) has given rise to a new discipline in science: translational research or translational medicine. Scientists in this field work to translate biological phenomena into targeted, evidence-based medicines that improve health and treat disease by more optimally matching drugs and individuals. Currently, at least 95 percent of pharmaceutical companies are performing translational research and the translational efforts are driving many of the new therapies entering the clinic today. But those advances don’t come for free. According to the National Center for Advancing Translational Science, translational medicine has “increased research costs and complexity,” and is on par with more traditional clinical challenges of recruiting, study design, and regulatory burdens in driving clinical study costs.

1 It enables translational researchers to easily search, access, and integrate complex, multivariate data, leading to proof or refutation of hypotheses and new questions and discoveries.

2 It’s designed and built from the ground up to serve translational scientists; an out-of-the-box solution, not a generic solution topped off for translational purposes.

3 The universe of supported data types is flexible and ever-expanding as new data types are identified as useful for translational research.

4 It leverages the cloud to improve productivity and collaboration while lowering total costs.

Current tools do not enable the translational researchers to engage directly and intuitively with the available data to affirm or refute a hypothesis. There is no easy means for scientists to search for and access integrated data so they can better identify and characterize biomarkers and develop the most efficient drug to treat a specific disease. Even the types of questions they can ask of the data are limited.

To gain the computational and bioinformatics power to analyze all the data, translational scientists most often call on IT counterparts or biostatisticians and data scientists to create custom applications. This creates its own problems. First, it can restrict the type of inquiry researchers can pose, inadequately focusing on the aftermath of an instrument run, for example. Secondly, it can take several iterations (not to mention days or weeks) before IT is able to serve up what the researcher needs – even if they deliver exactly what the researcher asked for.

New science needs new information solutions – self-service solutions that enable any scientist to engage directly with data more quickly and at a lower cost. These new solutions must address a different type of workflow, one that starts with a scientific question rather than the outcome of an experiment.

“Unless you can start harnessing data and making sense of it, in an automated way, with systems that are engineered to solve big data problems, you’ll be overwhelmed by the data very quickly,” says Nicolas Encina, vice president of the Innovation Lab at PerkinElmer. “You can no longer effectively manage this data manually and you certainly can’t analyze or process it manually either.”

“Too often, people think about data oriented from the informaticist’s or technologist’s point of view,” says Daniel Weaver, senior product manager for translational medicine informatics. “PerkinElmer Signals™ for Translational presents the data in a way a regular scientist will be able to understand. It’s organized around concepts a scientist gets, around the subjects of clinical trials, patient visits, samples collected, etc.”

Before PerkinElmer Signals™ for Translational, most scientists would query data, for example, based on results from a certain day or sample run. To glean more knowledge required manual analysis of multiple data sets layered in Excel spreadsheets. With the growth of data from R&D and clinical research, this task became even more challenging. The new self-service PerkinElmer Signals™ for Translational platform, however, automatically gathers disparate data to answer more open-ended questions, such as, “Do elderly female patients with KRAS mutant breast cancer have increased localization of protein ‘X’ to the nucleus?”