Studies that analyze huge, complex data sets, known as big data, may provide a new way to identify environmental factors that increase the risk of Parkinson’s disease (PD) and thus inform prevention efforts, according to Brad Racette, M.D., during an Oct. 18 lecture at NIEHS.
“Big data is often characterized by the three Vs — the extreme volume of the data, the velocity at which the data [is generated and] must be processed, and the variety of data that can be incorporated,” said Racette, from the Washington University School of Medicine in St. Louis. These studies require such massive computing power that two terabyte servers are barely adequate for Racette’s data analysis.
By comparison, most PD studies are small enough to analyze on a single personal computer. Typically based in neurology clinics, these studies may underrepresent certain populations. Moreover, as Racette pointed out, such efforts have limited ability to examine environmental factors.
Racette wants to combine years of Medicare records with detailed environmental data from the U.S. Environmental Protection Agency (EPA). He would then use an analytical technique called machine learning to look for patterns between environmental exposures and PD. “Let it spit out 25 new environmental toxins that we don’t know about,” said Racette. “I think that actually would change the field more than picking at the same environmental exposures for the next decade.”
Big data reveals important patterns
Using the Medicare data, Racette’s lab has already found patterns missed by clinic-based studies. His team reported a geographic clustering of PD cases in the Midwest and Northeast regions of the U.S.
Suspecting environmental influences, Racette and his colleagues used data from the EPA Toxics Release Inventory to study PD and local releases of manganese, copper, and lead — pollutants associated with the region’s industries. The team found 1.78 times more PD cases in urban counties with high manganese releases compared with urban counties with low manganese releases. They also observed a relationship, although it was weaker, between copper releases and PD.
Racette is also looking for symptoms of prodromal PD, or the period before motor control is impaired. Referring to this as the holy grail of PD, he and his colleagues launched a study of 90,000 cases and 120,000 control subjects. Earlier diagnosis may increase opportunities for effective treatments, said Racette, which have been elusive. He underscored the need for disease prevention, which is also a priority for NIEHS.
Cindy Lawler, Ph.D., head of the NIEHS Genes, Environment, and Health branch, agreed. “Prevention may be where you have the biggest impact,” she said. “If we can remove key exposures, we may be able to prevent the process of Parkinson’s from starting.”
Special funding mechanism enabled innovation
Racette’s funding includes a grant, called a K24, tailored to clinician-researchers who want to develop new projects and mentor junior scientists. “You have to understand this disease in order to study it,” said Racette, who has spent 20 years caring for Parkinson’s patients.
“The K24 grant has really supported innovation and the ability to mentor trainees on [the] high risk projects that have led to the work I’m presenting here,” he added. High risk is a term researchers use for studies that go in a new direction, without a base of previous findings to point the way.
“We don’t have many cases where clinicians were looking at patients and making this kind of connection,” said Carol Shreffler, Ph.D., program officer in the NIEHS Exposure, Response, and Technology branch. “This grant really helped him focus in on the environmental exposures.”
Wright Willis A, Evanoff BA, Lian M, Criswell SR, Racette BA. 2010. Geographic and ethnic variation in Parkinson disease: a population-based study of U.S. Medicare beneficiaries. Neuroepidemiology 34(3):143−151.
Willis AW, Evanoff BA, Lian M, Galarza A, Wegrzyn A, Schootman M, Racette BA. 2010. Metal emissions and urban incident Parkinson disease: a community health study of Medicare beneficiaries by using geographic information systems. Am J Epidemiol 172(12):1357−1363.
Racette BA, Willis AW. 2015. Time to change the blind men and the elephant approach to Parkinson disease? Neurology 85(2):190−196.
(Virginia Guidry, Ph.D., is a technical writer and public information specialist in the NIEHS Office of Communications and Public Liaison, and a regular contributor to the Environmental Factor.)