Eric Topol challenged claims that genetic scores can accurately predict IQ, a measure from intelligence testing, and height when selecting embryos

Eric Topol challenged claims that genetic scores can accurately predict IQ, a measure from intelligence testing, and height when selecting embryos

On August 16, cardiologist Eric Topol wrote about companies that charge up to 50 thousand dollars for embryo selection and claim they can accurately predict IQ, height, and other traits. In a post on X, he questioned these claims because a polygenic score combines many DNA variants into a single genetic score.

During in vitro fertilization (IVF), one embryo is selected from several for transfer to the uterus. A polygenic score combines many DNA variants, each of which has a weak association with a trait. In a large population, such a score may capture a statistical difference. In IVF, however, it is applied to a small number of genetically similar embryos. The future child’s height and cognitive test results will also depend on the environment in which the child grows up.

Topol wrote:

“As if their IQ, height, and other traits could be accurately predicted with polygenic risk scores. They cannot.”

A 2019 study examined this specific question within families: what benefit comes from selecting the embryo with the highest score? The authors combined a mathematical model, simulations, and data from large families whose children had already reached adulthood. With the models available in 2019 and the number of embryos typically available during IVF, selecting an embryo rather than choosing one at random produced an average gain of about 2,5 centimeters in height or 2,5 IQ points. Among 28 families with as many as 20 adult children, the child with the highest height score was the tallest in 7 cases. An average difference in a large sample and a prediction for one future child answer different questions.

Selecting embryos to reduce the risk of a specific disease requires a different calculation. In a 2021 model, the result depended on the selection strategy, the disease, and the number of embryos. Excluding embryos with very high risk scores produced only a small reduction in risk. Selecting the embryo with the lowest risk score produced a larger relative reduction when enough embryos were available. Future height and IQ require different calculations and produce different results from those used to estimate disease risk.

In its 2026 position statement, the American Society for Reproductive Medicine placed the use of PGT-P, or preimplantation genetic testing for polygenic disorders, to select for height, intelligence, and eye color outside the scope of reproductive medicine. The document describes PGT-P as an emerging, unproven technology and recommends against offering it as a clinical service.

A polygenic score converts many weak genetic associations into a probabilistic estimate. Choosing among several related embryos requires a prediction for one person, not an average difference across a large population. Topol’s criticism concerned this shift from a statistical estimate to a promise about an individual.