Source: Kagenmi/Getty Images
Source: Kagenmi/Getty Images

Direct-to-consumer genetic testing company 23andMe has announced it will offer a new report that provides an assessment of an individual’s risk of developing type 2 diabetes. Development of the report drew upon the data of more than 2.5 million 23andMe customers who consented for their data to be used for research purposes.

“Diabetes is a significant health issue in the United States that is expected to impact nearly half of the population. When customers learn about their genetic likelihood of developing type 2 diabetes, we believe there is an opportunity to motivate them to change their lifestyle and ultimately to help them prevent the disease,” said Anne Wojcicki, CEO and co-founder of 23andMe, in a press release.

The company developed the polygenic risk score (PRS) for type 2 diabetes by conducting a genome-wide association study (GWAS) that examined more than 1000 genetic variants. The report, which is made available to 23andMe Health+Ancestry Service customers, also provides information about how their age, weight, and lifestyle can influence the development of type 2 diabetes and what actions may prevent, or delay onset of, the disease.

The 23andMe Health+Ancestry report is the latest offering from a company transitioning from a simple reporter of genetic findings to a company capable of leveraging its vast library of genetic data for DTC and research purposes alike. 23andMe currently has data from more than 5 million patients in its research platform.

It’s the kind of data that has significant implications for drug development and the creation of precision medicine care models. In October 2018, for example, pharma giant GlaxoSmithKline (GSK)made a $300 million equity investment in the company, which intends to leverage the data of 23andMe’s research platform across a range of its therapeutic development programs including Parkinson’s disease.

The availability of the new PRS for type 2 diabetes also helps support a partnership the company announced earlier this year with Lark Health, a company that leverages artificial intelligence and patient monitoring for the management of chronic health conditions. Under the collaboration, 23andMe customers can use their genetic data in conjunction with Lark’s platform for its Diabetes Prevention Program.

23andMe also released a white paper detailing the methods it used in the GWAS to develop the PRS for the new report. The report provides the following information (and more):

  • A qualitative summary of the customer's genetic predisposition to type 2 diabetes (“increased” or “typical” likelihood of developing type 2 diabetes);
  • an estimate of the customer's remaining lifetime risk of type 2 diabetes, based on genetics, age, and self-reported ancestry;
  • estimates of 10-year risk for people with similar PRS profiles by age (range: 20–70years), illustrating peak likelihood of developing type 2 diabetes during midlife;
  • a “prevalence explorer” tool that illustrates how age, BMI, diet, and exercise habits impact type 2 diabetes prevalence in 23andMe research participants with similar type 2 diabetes PRS;
  • Information about lifestyle choices previously shown to lower the likelihood of type 2 diabetes, including participating in Diabetes Prevention Programs (Knowler, et al. 2002);
  • general information about the causes, symptoms, complications, and risk factors for type 2 diabetes;
  • limitations of the report (e.g., that it does not cover every possible genetic variant that could impact the likelihood of developing type 2 diabetes);
  • a “Scientific Details” page that summarizes the methodology used to generate the report, screening guidelines recommended by the American Diabetes Association, additional information about non-genetic factors that impact the likelihood of developing type 2 diabetes, and reference.

In the discussion section of the white paper, 23andMe notes this first method of reporting type 2 diabetes risk has a number of limitations and areas for improvement. One area would be to incorporate additional modeling methods to the one used, which fit a logistic regression to a selected set of variants and covariates. It also noted the common problem of a lack of diversity in the data and that the model was trained using a GWAS of European ancestry—a gap the company hopes to address in the future across all ethnicities as diverse databases of genetic databases continue to grow rapidly.

The company also hopes that over time the PRS can be used for more than just risk assessment.

“While we see potential for the inclusion of PRS into many aspects of both diagnosis and treatment of [type 2 diabetes] someday, the PRS powering this report is intended to only inform customers of their genetic likelihood for developing [type 2 diabetes],” the white paper noted. “Customers concerned about their [type 2 diabetes] risk will be informed about healthy diet and exercise choices, will be informed about diabetes prevention programs, and will be advised to talk to a healthcare professional. More research into optimal risk thresholds, model design, and overall PGS validation is needed before PRS such as this are used for more specific clinical decision making.”

This site uses Akismet to reduce spam. Learn how your comment data is processed.