A new genetic technique developed by a research team from Michigan State University uses a machine-learning algorithm to build predictors for human traits such as height, bone density, and even the level of education a person might achieve, purely based on one's genome.
"While we have validated this tool for these three outcomes, we can now apply this method to predict other complex traits related to health risks such as heart disease, diabetes, and breast cancer," said lead investigator Stephen Hsu of MSU. "This is only the beginning."
The research, featured in the October issue of Genetics, analyzed the complete genetic makeup of nearly 500,000 adults in the United Kingdom using machine learning—the process through which computers learn from data. In validation tests, the computer accurately predicted every subject’s height within roughly an inch. While bone density and educational attainment predictors were not as precise, they were accurate enough to identify outlying individuals who were at risk of osteoporosis or of struggling in school.
Traditional genetic testing typically looks for a specific change in a person's genes or chromosomes that can indicate a higher risk for diseases. Hsu's model considers numerous genomic differences and builds a predictor based on the tens of thousands of variations.
Using data from the UK Biobank, an international resource for health information, Hsu and his team put the algorithm to work, evaluating each participant's DNA and teaching the computer to pull out these distinct differences.
"The algorithm looks at the genetic makeup and height of each person," Hsu said. "The computer learns from each person and ultimately produces a predictor that can determine how tall they are from their genome alone."
Hsu's team will continue to improve the algorithms while tapping into larger, more diverse data sets. Doing this would further validate the techniques and continue to help map out the genetic architecture of these important traits and disease risks.
"Our team believes this is the future of medicine," he said. "For the patient, a genomic test can be as simple as a cheek swab, with a cost of about $50. Once we calculate the predictors for genetically based diseases, early intervention can save billions of dollars in treatment costs, and more importantly, save lives."