New reports give more personalized and accurate results by incorporating genetic ancestry as a continuous spectrum instead of grouping customers into ancestry categories.
The question looks so simple, just four words: “What is your race?” We see it on forms seemingly everywhere we go — at the doctor’s office, work, and school. But the question’s simplicity belies the deep complexity of racial, ethnic, and ancestral origins. In the 2020 Census, over 10 percent of the United States population identified as multiracial.
Similarly, at 23andMe, we found that many consented 23andMe research participants answered “I’m not sure” rather than choosing one of the options when asked which population they “most closely” identify with. As the nation becomes increasingly diverse, the assumption that a person’s identity can be captured meaningfully by a single racial or ethnic category becomes ever more flawed.
Reflecting Genetic Diversity
At 23andMe, we know that every customer has a unique Ancestry Composition, a genetic mosaic formed by countless ancestors over generations. To reflect this genetic diversity better, we are updating the algorithm that powers new reports available to 23andMe+ Premium members to reflect genetic ancestry as a continuous spectrum instead of as distinct groups.
The new reports are powered by 23andMe research and based on polygenic scores, which estimate someone’s overall genetic likelihood of having a trait or disease based on the combined impact of thousands of genetic variants. Any given health condition may be more or less common from one global population to another.
Likewise, any given genetic variant may be more frequent in some populations than others. To tease apart all these interrelated factors and give accurate results to customers of varying ancestral backgrounds, customer results based on polygenic scores must account for ancestry in some way.
Polygenic Scores
For several years, 23andMe developed reports using a polygenic score algorithm that incorporated data from consented research participants with predominantly East Asian, European, Hispanic/Latino, Black/African American, South Asian, or West Asian/North African ancestry.
This algorithm, released in 2019, leveraged the fact that consented 23andMe research participants form the world’s largest, most ancestrally diverse recontactable genetic database. However, it was also limited because it relied on grouping customers with other genetically similar individuals to produce their report results.
The shift to representing genetic ancestry as a continuum is in line with emerging best practices in the scientific community; as we were developing this updated algorithm in 2023, the National Academies of Science, Engineering, and Medicine released guidelines recommending that scientific analyses represent genetic diversity as a continuous spectrum, rather than genetic groups.
Tailored Results
With that recommendation in mind, the new reports released this month take the inclusivity of the 2019 algorithm a step further by including consented research participants with mixed ancestry who do not fit into any of the genetic groups from the previous update. With this improvement, a customer’s results are calculated based on their personal, unique pattern of genetic ancestry, without comparison to groups of genetically similar people.
This allows each customer to receive results tailored to their unique genetic makeup, whether or not there is a large group of consented research participants with similar ancestry. The new algorithm was presented at the American Society of Human Genetics Meeting in November 2023. For more scientific details, see our white paper.
New reports released to 23andMe+ Premium members will be powered by these more personalized and accurate results starting in March 2024, beginning with three new reports on common forms of cancer.
