Jonathan Anomaly

239 posts

Jonathan Anomaly banner
Jonathan Anomaly

Jonathan Anomaly

@JonathanAnomaly

https://t.co/kd5ubQYRxj

Katılım Temmuz 2022
181 Takip Edilen1.6K Takipçiler
Sabitlenmiş Tweet
Jonathan Anomaly
Jonathan Anomaly@JonathanAnomaly·
1/ Today we launch an ambitious paper on the ethics of embryo screening. While the technology is new, our hopes and fears about our future children are as old as the Greek myths, including stories about Hera, goddess of fertility and the namesake of our company @herasight
Jonathan Anomaly tweet media
English
9
50
259
165.3K
Adam Butler
Adam Butler@GestaltU·
Listening to @hsu_steve and @AlexTISYoung discuss the ethics of embryo selection is disorienting. Seems disingenuous to laugh off the difference between screening out Huntington’s and optimizing for height or IQ, as if the distinction is some kind of philosophical naivety. A child with Huntington’s will lose motor control, lose cognition, and die. Sparing someone that is mercy. Now take height. Taller men earn more, get promoted faster, do better in dating markets. The advantage is real and well documented. So yes, selecting for height would genuinely benefit your child. But notice what kind of benefit it is. Positional. Your child gains because other people are shorter. You aren’t preventing suffering, you’re purchasing competitive advantage through your child’s biology. And height is the simple case. The moment you accept the logic of optimization, you enter dozens of arms races simultaneously, across traits whose interactions we barely understand. Polygenic traits share genetic architecture. They trade off in ways population genetics is only beginning to map. Now multiply that by millions of parents all optimizing at once, each against their own biased objective function, and you’re running a massively parallel experiment on the human gene pool with no coordination, no agreed-upon objective, and no rollback. We aren’t just bad at specifying the Pareto frontier across these traits. We don’t even know what we’re trading off against what, at what scale, with what population-level dynamics. Again, I’m 100% for funding genetic research, and I’m not even aspiring to a moratorium on positive screening - mostly because it would just drive it into more lenient jurisdictions. But we should have a society level conversation about what we’re doing and attempt some forms of coordinated regulations like we enforce in global organ trading, child trafficking etc. Technology is enabling arms races of increasing complexity and consequence at the exact time in our history that we appear least able to coordinate to bind negative outcomes. It’s important for top researchers and leaders in the field to take actual stands on these potentially existential issues, rather than pretending they don’t exist, or that they’re someone else’s problem.
English
5
4
24
8.6K
Jonathan Anomaly retweetledi
SynBioBeta
SynBioBeta@SynBioBeta·
@JonathanAnomaly said embryo genetics is moving beyond monogenic screening toward polygenic prediction validated within families, with IVG, gene editing, and synthetic chromosomes now back in the conversation. #SynBioBeta #Biotech
SynBioBeta tweet mediaSynBioBeta tweet mediaSynBioBeta tweet mediaSynBioBeta tweet media
English
0
11
23
1.9K
Jonathan Anomaly
Jonathan Anomaly@JonathanAnomaly·
@octal But it’s worth mentioning very few people have 20 embryos
English
1
0
1
117
Ryan Lackey
Ryan Lackey@octal·
Herasight can do 5-10 iq point selection over 20 embryos?
English
5
0
7
1.8K
Jonathan Anomaly retweetledi
steve hsu
steve hsu@hsu_steve·
Embryo Selection and Frontier Genomics with Dr. Alex Young – Manifold #111 Dr. Alex Young, a statistical geneticist and assistant professor in the Human Genetics department at UCLA, joins Steve Hsu to discuss the cutting edge of genomic prediction. They cover his research on polygenic embryo screening in IVF (including the ImputePGTA method), family-based DNA analysis, missing heritability, and the implications of polygenic scores for traits like education and disease. Alex also discusses his recent battles with cancer. @AlexTISYoung Chapter Markers: (00:00) - Alex Young Bio (06:36) - Biobank Era Genetics (10:49) - Missing Heritability Debate (27:18) - Embryo Selection Controversy (50:32) - Embryo Selection Backlash (53:42) - Mexico City Admixture Study (01:00:13) - Censorship Via Data Access Control (01:05:02) - Battle With Cancer and Circulating Tumor DNA (ctDNA)
English
3
22
105
19.5K
Jonathan Anomaly retweetledi
Foresight Institute
Foresight Institute@foresightinst·
45 million deaths a year are caused by biological aging. We treat most of them as inevitable. Tomorrow, @RaianyRomanni joins our seminar series to challenge that assumption. Drawing on interviews with 102 scientists and original economic modeling, she makes the case that advancing the science of aging may offer the highest social return of any investment in biomedical research. Request to join: luma.com/uxt39mla
Foresight Institute tweet media
English
0
5
7
532
Madhava Jay
Madhava Jay@madhavajay·
@JonathanAnomaly And i'm down to have that conversation because its fun to explore these thought experiments. 😊 What are your views on where this is all heading and what are the steel-man and straw-man arguments in your view?
English
1
0
0
29
Jonathan Anomaly
Jonathan Anomaly@JonathanAnomaly·
@madhavajay Skepticism is the essence of science, and is merited. But check out our validation studies at Herasight.com. No question our predictors work. The question is only how powerful they are, which depends in part on family history and ancestry
English
1
0
2
50
Madhava Jay
Madhava Jay@madhavajay·
As someone who did pgt-m for a serious disease and has an abundance of healthy embryos there is the question of how to stack rank them. In theory I would choose to use that data if it was accurate but im skeptical of the state of the art and would be concerned one bias might rule out other traits that might be more desirable.
English
1
0
0
219
Jonathan Anomaly
Jonathan Anomaly@JonathanAnomaly·
@SashaGusevPosts @GarettJones Please stop tweeting at me. I’ve been enjoying my weekend. You should do the same. I generously replied to your accusations, and you followed up with more accusations. You’re at Harvard, not middle school. Time to grow up
English
2
0
2
139
Sasha Gusev
Sasha Gusev@SashaGusevPosts·
@GarettJones @JonathanAnomaly I get the sense you guys don't spend a lot of time outside your bubble, so basic questions like "what are the details of the study you're running?" or "why are you unwilling to bet on this?" produce a great deal of confusion and frustration.
English
1
0
3
169
Sasha Gusev
Sasha Gusev@SashaGusevPosts·
It's very telling that the embryo selection companies are not interested in rigorous prospective validation -- the milestone all clinical companies strive for and all sketchy wellness cures avoid like the plague.
Sasha Gusev tweet media
New York Magazine@NYMag

Many of Herasight’s first customers were Bay Area rationalists. The company’s promise to use data gleaned from embryo genomes to predict the future of a child-to-be appealed to IQ-obsessed quants who were already thinking about their “reproductive health stack” of supplements and peptides. “It’s hard not to love this technology,” called polygenic screening, writes Scott Alexander, the author of the rationalist blog Astral Codex Ten. Peter Thiel, Alexis Ohanian, and Coinbase’s Brian Armstrong have all invested in companies that provide polygenic embryo screening. Now this innovation seems poised to breach Silicon Valley containment. In the fall, CBS ran two separate specials on polygenic screening, one on Herasight and one on Nucleus Genomics. And in November, Nucleus ran a high-profile series of ads in the New York subway system. “IQ IS 50% GENETIC” one read. In the Broadway–Lafayette Street station, commuters were confronted by banners with the tagline “HAVE YOUR BEST BABY.” The sales pitch is appealingly straightforward: Choose one embryo and gain four inches in height; choose another and add nine IQ points. In reality, behind each number is a series of nesting uncertainties inherent in prenatal testing, in how genetic risk is calculated, and in the IVF process itself, which, taken together, makes this technology deeply unreliable. “How many lawsuits are going to happen,” one clinician wondered, “because you’ve supposedly chosen an embryo that’s going to be tall, beautiful, and smart, and they’re short, squat, thick, and a little dull?” Christopher Cox reports on the shaky science behind polygenic embryo screening and the prospective clients who don’t seem to mind: nymag.visitlink.me/Hn3i66

English
8
23
128
13.3K
Jonathan Anomaly
Jonathan Anomaly@JonathanAnomaly·
It’s interesting that Sasha always assumes the worst motives in people when he is emotionally invested in a topic. He can’t imagine the someone might disagree with him on a consequential topic in good faith. It’s textbook motivated reasoning. Our science team is about as strong as anyone can imagine in this field. The real question is why he would assume bad motivations and poor scientific methods about many of his colleagues. And why he’d need to express this on twitter when we’ve repeatedly offered to have discussions with him, as we do with others in the field. We’ve been nothing but cordial, despite his constant attacks. I’ve had enough engaging with someone who always assumes the worst in people. It is a vice masquerading as virtue
English
1
0
2
271
Sasha Gusev
Sasha Gusev@SashaGusevPosts·
@GarettJones @JonathanAnomaly Sure, $100 that Herasight never replicates their published IQ PGS correlation of 0.45 in a prospective within-family analysis of actual customers (e.g comparing selected to non-selected siblings). $50 that they don't publish such a study at all for any polygenic score.
English
2
0
16
405
Jonathan Anomaly
Jonathan Anomaly@JonathanAnomaly·
@SashaGusevPosts By the way, this is why people don’t engage with journalists. Please speak to Alex Young or me personally and we’re happy to tell you more. There’s no point outlining every way in which journalists and editors mislead the public
English
0
0
3
67
Jonathan Anomaly
Jonathan Anomaly@JonathanAnomaly·
That quote was (unsurprisingly) taken out of context btw. I explained that we can know BRCA mutations are associated with breast cancer and know that testing embryos for BRCA works without needing to wait half a century after babies are born, and that a similar point applies to some polygenic scores (eg we already know some of the causal variants associated with T1D)
English
1
0
4
601
Sasha Gusev
Sasha Gusev@SashaGusevPosts·
@JonathanAnomaly Was the highlighted quote taken out of context? Can you provide a fuller statement?
English
2
0
10
541
Anshul Kundaje
Anshul Kundaje@anshulkundaje·
@herasight would love to hear your take here. If you are so confident about the technology, it would make sense to bet on a serious follow up study covering different traits & diseases to everyone that you are right?
Sasha Gusev@SashaGusevPosts

It's very telling that the embryo selection companies are not interested in rigorous prospective validation -- the milestone all clinical companies strive for and all sketchy wellness cures avoid like the plague.

English
1
0
12
1.7K
Jonathan Anomaly
Jonathan Anomaly@JonathanAnomaly·
@unboxpolitics @shae_mcl We're going to continue to improve both the r2 of our predictors and the power in our pleiotropy studies as we access more data. And, despite what journalists like to write about, our main focus is on disease, not IQ. That's also the focus of the vast majority of our customers.
English
0
0
1
38
Vinay Tummarakota
Vinay Tummarakota@unboxpolitics·
@shae_mcl @JonathanAnomaly Yeah, a within-family study of Herasight’s IQ PGS would rule out environmental confounding. Unfortunately, it seems like they didn’t have enough power to do that in their analysis of pleiotropy.
Vinay Tummarakota tweet media
English
2
0
3
79
Shae McLaughlin
Shae McLaughlin@shae_mcl·
Great article that does a good job of explaining the significant limitations of embryo selection based on polygenic scores. If someone tells me they plan on IQ-selecting their embryos, it’s a negative IQ signal for me.
New York Magazine@NYMag

Many of Herasight’s first customers were Bay Area rationalists. The company’s promise to use data gleaned from embryo genomes to predict the future of a child-to-be appealed to IQ-obsessed quants who were already thinking about their “reproductive health stack” of supplements and peptides. “It’s hard not to love this technology,” called polygenic screening, writes Scott Alexander, the author of the rationalist blog Astral Codex Ten. Peter Thiel, Alexis Ohanian, and Coinbase’s Brian Armstrong have all invested in companies that provide polygenic embryo screening. Now this innovation seems poised to breach Silicon Valley containment. In the fall, CBS ran two separate specials on polygenic screening, one on Herasight and one on Nucleus Genomics. And in November, Nucleus ran a high-profile series of ads in the New York subway system. “IQ IS 50% GENETIC” one read. In the Broadway–Lafayette Street station, commuters were confronted by banners with the tagline “HAVE YOUR BEST BABY.” The sales pitch is appealingly straightforward: Choose one embryo and gain four inches in height; choose another and add nine IQ points. In reality, behind each number is a series of nesting uncertainties inherent in prenatal testing, in how genetic risk is calculated, and in the IVF process itself, which, taken together, makes this technology deeply unreliable. “How many lawsuits are going to happen,” one clinician wondered, “because you’ve supposedly chosen an embryo that’s going to be tall, beautiful, and smart, and they’re short, squat, thick, and a little dull?” Christopher Cox reports on the shaky science behind polygenic embryo screening and the prospective clients who don’t seem to mind: nymag.visitlink.me/Hn3i66

English
3
7
34
3.8K
Shae McLaughlin
Shae McLaughlin@shae_mcl·
I read the technical paper when it came out. I think that at ~0.45 (I'd rather use the observed within-family estimate of 0.35, but I'll be generous), the predictor is effectively useless for individual prediction. The 95% PI full width for an individual prediction is ~52 IQ points (3.49 SD).
English
2
0
3
190