Researchers at Columbia University published a paper today in Nature Genetics describing a machine-learning algorithm, developed under DARPA’s SIMPLEX program, that could rapidly scan massive genomic data sets and infer the ancestral makeup of an individual. This knowledge could help identify disease-carrying genetic mutations in an individual to provide an optimum personalized diagnosis, treatment or prevention program on the basis of an individual’s unique genetic profile.

“We’re very pleased with how SIMPLEX researchers are creating novel mathematical frameworks and machine learning tools to analyze scientific data across diverse complex science and engineering disciplines to enable generation of big hypotheses,” said Reza Ghanadan, DARPA program manager. “The paper published today is another example of how SIMPLEX-developed tools are unifying vast amounts of scientific data and knowledge into frameworks that can advance our understanding in diverse areas such as personalized medicine, neuroscience, materials science, and other domains.”