Big Data analysis outfit GNS teams with Inova on predicting preterm birth

Could crunching data from genomes and electronic health records prevent some of the thousands of U.S. infant deaths from preterm births? GNS Healthcare, a Big Data analytics firm, and Inova Translational Medicine Institute (ITMI) have embarked on a new project to create computer models that predict risk of preterm birth. So maybe.

Cambridge, MA-based GNS Healthcare says that the models would be shared with academics, healthcare groups and biopharma organizations to improve diagnostics and treatments of preterm birth. This is one of the latest in a series of deals GNS has landed to create computer models for researchers, including its work funded by Johnson & Johnson ($JNJ) to create digital maps to uncover the inner workings of disease in multiple sclerosis. The initiative, called Orion Bionetworks, was launched earlier this year.

Preterm births present a new conundrum for GNS CEO Colin Hill and his digital data sleuths to solve. A 2006 paper in the journal Pediatrics shed light on the grim problem in the United States, where authors said 12% of babies arrive too early, resulting in 10,000 deaths annually. Researchers have suspected nonspecific genetic culprits in preterm births, but they lack knowledge about specific genetic triggers. And there are factors outside of genetics at play in the problem.

GNS plans to build computer models with data from ITMI at Inova Fairfax Hospital in Falls Church, VA. Inova has accumulated large amounts of information on the genetics and clinical outcomes of families that have experienced preterm birth. The software company uses advanced machine learning algorithms and other methods to automatically connect health problems like preterm birth with molecular information from next-generation genomic sequencing and other biological analyses.

There are no guarantees that such Big Data analysis holds the key to cracking the mysteries of a complex health problem like preterm birth. Yet Hill and his colleagues are confident that humans need lots of help from machines to make use of huge amounts of health data to advance personalized treatments.

"These models will create new ways for clinicians and scientists to understand these interactions and will accelerate the discovery of new diagnostic tools and treatments for this condition, as well as other complex conditions," Hill said in a statement.

- here's the release