Why only 12% of reproductive clinicians approve of polygenic embryo screening

If you're a reproductive endocrinologist or genetic counselor, you've probably settled on some version of this: polygenic embryo screening is interesting, but not ready. Professional societies urge caution. The evidence base is still developing. Better to wait.

That's a reasonable position. But there's a number that should change how you think about it.

A recent survey of U.S. REIs found that only 12% generally approve of PGT-P. Yet when the same clinicians were asked whether PGT-P should be allowed, 44% said yes. Another 42% were uncertain, not opposed.

Read that again. Almost half of the clinicians who don't "approve" of PGT-P would still vote to allow it. The vast majority aren't saying "this doesn't work." They're saying "I don't have a way to evaluate whether this works."

That's not a scientific objection. That's a missing framework.

The label is hiding a consensus

Here's the part most people miss. The same survey broke out support by what PGT-P screens for. When it's physical and psychiatric health conditions, 55 to 59% of REIs support it. When it's behavioral or physical traits like height, support drops to 6 to 7%.

That's a massive gap. And it means the label "PGT-P" is doing most of the damage. Clinicians hear the term and think of a single technology that includes everything from disease screening to trait selection. But those are completely different things, and the majority already support the disease screening side.

Herasight's screening focuses on conditions like type 2 diabetes, breast cancer, schizophrenia, Alzheimer's, and coronary artery disease. Not height. Not eye color. The technology that clinicians already support, separated from the umbrella they don't.

The concerns are real. We addressed them.

Professional societies raised a specific methodological concern about PGT-P, and it's legitimate: embryos from the same IVF cycle are biological siblings. The genetic variance between them is roughly half of what you'd see between unrelated people. A polygenic score validated on population data can look impressive and still fail to differentiate between siblings.

We validated on sibling pairs specifically because that's the test that actually matters. Moore et al. 2025 tested 17 disease scores on within-family data. 16 of 17 showed no decrease in predictive performance compared to population-level associations. The scores capture genuine genetic signal, not population structure artifacts that vanish when you control for shared environment.

But that's only part of the story.

Standard polygenic risk scores trained on European-ancestry data lose accuracy for other populations. The ISPG's own analysis found that risk reductions for embryos of African ancestry were about half those for European ancestry using standard approaches. A score that works within families but only for one ancestry group isn't useful for the patient population most clinics actually serve.

We calibrate and publish performance across 8+ ancestry groups. Moore et al. showed similar relative risk reductions for type 2 diabetes across European, African, and East Asian ancestries.

And there's one more thing clinicians don't always think about. Without family context, a risk score is just a number against a population average. That's like reporting a cholesterol reading without knowing whether the patient's parents both had heart attacks at 50. We set each family's baseline from ancestry-specific prevalence and parental polygenic scores, then adjust for reported disease in relatives. The report is interpretable for that patient, not their population's average.

Your clinic probably doesn't need to change anything

Most IVF clinics already run PGT-A for chromosomal screening. ImputePGTA reconstructs polygenic risk scores from that existing data. No new biopsy. No new sample logistics. Patients who already have PGT-A results can add polygenic screening retroactively.

For new patients, our standard workflow uses 2x low-pass whole-genome sequencing on embryos with 30x parental sequencing. We include carrier screening. Turnaround is about 8 weeks, with faster options for urgent frozen-embryo cases.

The survey also found that 85% of REIs cited confusion over results as a top concern. That's fair. Most genetic counselors don't have training in polygenic screening. Ours do. We walk through results with each family: what the numbers mean given their specific conditions and family history, and what the tradeoffs look like across their actual embryos. That conversation is different from anything your patients are getting right now, and we handle it.

We showed up

We presented three papers at ACMG 2026, covering within-family validation and embryo genome imputation alongside type 1 diabetes risk modeling. No other PGT-P company was present.

We did that because we think the evaluation standard for this technology should be public and the science should be open for scrutiny. If the concern is "we don't have enough evidence to evaluate this," the answer is to put the evidence in front of you and let you evaluate it.

If you want to see the validation data or talk about what integration looks like for your clinic, reach out to our team.

‍