Designer babies are coming soon to an IVF clinic near you.
Nucleus Genomics, founded by Kian Sadeghi in 2020, when he was just 20, got its start analyzing genomes to weigh a person’s risk of everything from cancer to ADHD.
Sadeghi won over investors like Alexis Ohanian and Balaji Srinivasan, raising $32 million. Now, he wants to bring this level of advanced genetic testing to embryos.
While most IVF clinics test embryos for the potential of severe disease, Nucleus goes further, offering results meant to judge a future child’s chance of being overweight, bald, short or more. For $25,000, customers can get Nucleus’s top-tier offering: Sadeghi’s team will help you find a genetically optimal egg or sperm donor, surrogate, and, ultimately, embryo. “It’s basically a patient-first IVF built for longevity — built for genetic optimization,” he said. Nucleus aims to launch its service in Bay Area clinics within a few months.
But some scientists have raised the alarm. Nucleus tests for polygenic traits and conditions, meaning those caused by genes working in concert. Some studies show that many polygenic results are a crapshoot, with an overreliance on Eurocentric datasets. But believers say it’s a rapidly developing field that will get more accurate as the datasets grow. I caught up with Sadeghi to dive into Nucleus’ expanded services and what they mean for the future of babies.
The interview had been condensed and edited for clarity and length.
Clinics vary in what they will allow patients to select for in embryos. How hard was it to find IVF clinics that would partner with you on your technology?
Certain clinics are pro the technology and will embrace it and do what’s best for the patient. Others are not. Clinics ended up reaching out to us. We found that patients wanted this. I think it’s often lost in the conversation that patients demand better technology from their clinics and their physicians. They want to pick their future embryo. They want to know more about their future challenges.
Controversially, you offer embryo testing for traits like height and intelligence. Can you explain that decision?
When a couple approaches our technology, the first question in their minds is about health. Usually, couples don’t want their child to necessarily have a severe chromosomal abnormality. From there, they might test for cystic fibrosis or PKU or Tay-Sachs. Then they might test for polygenic predictors for things like breast cancer and for diabetes.
After a parent checks all these boxes, if they want to pick based on which embryo will have a couple more inches, that’s their right. That’s their choice. Remember, it’s always a relative optimization. If you want a kid who’s tall, and you marry someone who’s 4-foot-5, no amount of genetic optimization is going to help you there. The most important decision is your partner.
Is this a new form of eugenics?
Ultimately, what we’re doing is helping parents give their child the best start in life in every sense. The beautiful thing about genetics is that it’s not destiny. It’s never going to be destiny. There’s always going to be that nurture component, because that’s how life is. You can’t reduce someone to just their DNA, right? Of course not. Everybody knows that.
But many are afraid of a future in which people are selecting kids by traits like IQ or hair or eye color. Do you think that’s a future we should be working to stop?
It’s up to the patient. For a while, IVF itself was something that was quite controversial. Over time, the technology became much more adopted. It’s kind of like AI, where everyone’s like “AGI! AGI! AGI!,” and it’s this scary thing. But then you use ChatGPT and you realize very quickly that, oh, it’s just another tool. Eventually, patients are going to see genetic optimization as just another tool.
Your competitors, like Orchid and Heliospect Genomics, won’t admit that they conduct IQ testing on embryos, although there are reports saying they do. You have the opposite strategy: You are upfront about IQ testing on embryos.
A problem with the genetic optimization industry is that people have really opposite approaches. They want to hide things. They don’t want to say things they do. They want to serve a very small constituent of technocrats. To me, it’s crazy, because once you start trying to hide science, people actually overextrapolate and make up all sorts of stories that just simply aren’t true about the science. The original ethos of technology companies was not to serve a small group of people at exceptionally high prices — it was to drive down the cost so that anybody, anywhere, can be served.
There’s a real question to be asked here: Why is polygenic testing for embryos not ubiquitous? It offers clear value, it’s well elucidated, and Nucleus is bringing down the cost for everybody. I actually think it’s because the companies have not been pushing aggressively enough to say: This is not a technology that belongs only in Silicon Valley boardrooms. This is a technology that belongs in the hands of every single couple undergoing IVF.
I think you know the counterargument that it’s difficult for the average person to evaluate the quality of individual polygenic scores. Which is a problem, because some of these scores are way more validated by research than others. How do you communicate that to clients with no scientific background?
You’re correct that explaining polygenic predictors can be difficult. One of the criticisms of polygenic predictors is that they’re hard to explain. But we shouldn’t be presumptuous here in assuming the patient can’t understand.
To me, the challenge you articulate is an opportunity to actually educate patients, make those materials, and provide them support during the entire process. We have counselors with the patient every step of the way, from adult testing to embryonic testing. After embryonic testing, patients have a genetic counselor always there to guide them through it.
When I used Nucleus this year for a story about the rise of superbabies in Silicon Valley, I was originally given a 53% risk of type 2 diabetes, based on a study from 2018. A few weeks later, the number was updated with a far more comprehensive 2022 study, and the risk plummeted to 10.9%. How can customers ensure they’re getting accurate information?
That’s been on my mind a lot, because I do think there needs to be more insight and clarity for the patients. The true answer is something that we’re going to be announcing soon, but I can’t actually talk about it yet.
One thing I will say is, think about AI models. People understand that AI models are going to get better and better. There’s a social contract of, I’m using this tool, and it may be wrong on something important, but I need to exercise my judgment and a healthy dose of skepticism. I don’t think it’s a perfect analogy when it comes to polygenic predictors, because people interact with them differently. But I do think there needs to be an expectation and clear communication from the industry that these models will get better, and they will change. That is not something we should hide.