UCLA genetics professor Alex Young on polygenic embryo selection: how IVF can halve disease risk and why academia punishes researchers who get involved

Polygenic embryo selection: halving disease risk from routine IVF data

Alex Young, a statistical geneticist at UCLA and advisor to Heracyte, is working on one of the more politically charged corners of human genetics: using polygenic scores to predict disease risk and traits in IVF embryos before implantation.

The technology is not new in principle. Genetic testing in IVF dates to the 1990s, when clinics began screening embryos to avoid sex-linked diseases. What has changed is the computational side. The collapse in sequencing costs over the past two decades generated datasets large enough to build meaningful predictors. The UK Biobank — 500,000 people with whole-genome data and linked medical records — became the flagship resource, and Young credits its relatively open access policy for much of the subsequent research progress. He notes the US has been "quite behind" in making comparable datasets available, particularly for commercial use.

What the technology can actually do

For diseases with a single gene exerting an outsized effect, the results are striking. Young says polygenic embryo selection can "pretty much eliminate" the risk of passing on type one diabetes in at-risk couples, and substantially reduce Alzheimer's transmission. For more complex diseases like type two diabetes, the effect is smaller but still material: a couple with a family history carrying a 20% offspring risk could, Young argues, bring that down to roughly 10%.

Heracyte has also built what Young describes as a more powerful predictor of IQ than researchers previously thought achievable, using deep learning to create better-curated psychometric data before applying a Bayesian statistical model.

“You can roughly say half the risk your offspring gets some disease... I also had some collaborators pull out of a paper I had in review at Nature and I also had a job offer rescinded from a top US university... Heracyte's been working on rare damaging mutations not typically included in these scores — including AI models like AlphaFold, AlphaMissense to predict what risks these rare mutations might confer.”

The algorithm Young built

Young's own contribution to Heracyte was not a prediction model but a data access layer. He developed a hidden Markov model that takes data from the routine aneuploidy test — already performed in approximately 60% of IVF cycles in the US — and converts it into a comprehensive genome profile of each embryo. The practical effect is that couples don't need a separate, purpose-built genetic test. Under HIPAA in the US and GDPR in Europe, patients have a legal right to request that routine test data from their clinic, send it to Heracyte, and receive disease risk and trait predictions from it. The algorithm lowers both the cost and the regulatory barrier to access.

Academic blowback

Young paid a price for getting involved. When his Heracyte advisory role became known, collaborators withdrew from a paper he had under review at Nature, and a job offer from a top US university was rescinded. He frames the backlash as ideological: the reproductive genetics space, he argues, has been "neglected" partly because of political opposition, even as pharmaceutical companies like Regeneron have invested heavily in human genetic data for drug target discovery — a use case that attracts far less controversy.

Regulation and market dynamics

The US IVF market is largely unregulated, which Young sees as a double-edged situation. Light-touch oversight could keep out bad actors — he flags misleading test results from at least one competitor as a concern — but he worries any formal regulatory framework would quickly kill the innovation that has allowed US-based companies to develop products that simply cannot be offered in Europe.

Heracyte is targeting a fundraise this year. Revenue is described as growing, and Young anticipates a preference cascade: early adopters have skewed toward wealthy, tech-adjacent customers, and he expects broader demand to follow as the technology becomes more visible.

What comes next

Young sees the current polygenic score approach as capturing only a fraction of the genome's total signal. Rare damaging mutations — not typically included in existing scores — represent the next frontier, and Heracyte is already using AI tools including AlphaFold and AlphaMissense to predict what risks those mutations might carry. The company has an early product using rare variants to predict neurodevelopmental disorder risk.

Further out, Young envisions a stack combining in vitro gametogenesis (creating eggs from adult cells, enabling the generation of thousands of embryos), polygenic selection across that pool, and CRISPR-based editing of the most promising candidates — including, potentially, trait enhancement rather than purely disease removal. He is candid that this is speculative and far from clinical reality, but it is the direction he finds genuinely exciting.

Young is also a cancer patient who underwent fertility preservation before chemotherapy, which he says made him "a lot more pro-IVF" and gave him a firsthand view of how poorly the standard oncology playbook serves patients who want to be aggressive about treatment options.