Turning basic science into precision medicine offering with millions of people

One of the promises of precision medicine is early risk prediction for common conditions. Such predictions enable focusing prevention efforts on high risk individuals, promote early intervention, and empower patients to better plan their lives. The advent of polygenic risk scores ushered in the machine learning framework to implement early prediction based on genetic information. However, there are multiple challenges in turning raw predictions into clinical information, ranging from statistical issues, via regulatory aspects, to the participants ability to comprehend those risks.

Here, we describe our approach in implementing large-scale genetic risk predictions within global direct-to-consumer genetic tests at MyHeritage. Our approach leverages data and experience from >3 million people that took our genetic test and thousands who answered our questionnaires, rendering us one of the largest databases for genetic information. We used these datasets to develop and validate a clinical test that can predict a range of conditions, from risk to heart disease to breast cancer. We used multiple statistical tools to understand the limitations of our predictions and ensure that only eligible participants receive those tests. Moreover, we set a unique regulatory framework that leverages scalable telemedicine services to provide physician oversight and offer genetic counseling to contextualize the results. Finally, we combine various expertise, including cognitive psychology and design principles, to deliver results to participants.

This talk will show the process it takes to transform findings in fundamental science into a complete offering that empowers the lives of participants in this new world of affordable genetic tests.