Sunday, January 30, 2022

Genetic risk factors have a substantial impact on healthy life years (FinnGen)

This new preprint obtains very interesting results using data from the FinnGen cohort of 300k+ Finns (genotypes + medical records) and UK Biobank. 
Genetic risk factors have a substantial impact on healthy life years 
Sakari Jukarainen et al. 
The impact of genetic variation on overall disease burden has not been comprehensively evaluated. Here we introduce an approach to estimate the effect of different types of genetic risk factors on disease burden quantified through disability-adjusted life years (DALYs, “lost healthy life years”). We use genetic information from 735,748 individuals with registry-based follow-up of up to 48 years. At the individual level, rare variants had higher effects on DALYs than common variants, while common variants were more relevant for population-level disease burden. Among common variants, rs3798220 (LPA) had the strongest effect, with 1.18 DALYs attributable to carrying 1 vs 0 copies of the minor allele. Belonging to top 10% vs bottom 90% of a polygenic score for multisite chronic pain had an effect of 3.63 DALYs. Carrying a deleterious rare variant in LDLR, MYBPC3, or BRCA1/2 had an effect of around 4.1-13.1 DALYs. The population-level disease burden attributable to some common variants is comparable to the burden from modifiable risk factors such as high sodium intake and low physical activity. Genetic risk factors can explain a sizeable number of healthy life years lost both at the individual and population level, highlighting the importance of incorporating genetic information into public health efforts.
The figure below shows DALYs attributable to being in the top 10% vs bottom 90% of each PGS. (So, roughly, top 10% vs average individuals.) 

The Shorter Lifespan Polygenic Score is a kind of index similar to the Embryo Health Score used by Genomic Prediction. Note that the difference between 90+ percentile and average individuals is roughly 4 DALYs!  





In our 2019 sibling validation paper we showed that disease risk polygenic scores have roughly as much predictive power to differentiate high and low risk sibs as when applied to pairs of unrelated individuals. 


Thus the results above are indicative of DALY gains from embryo selection.


In 2018 we had Arno Palotie, one of the founders of FinnGen, at MSU to give a talk about the project. 



Wednesday, January 26, 2022

ManifoldOne podcast

Please visit Manifold1.com

First episode is up!

James Lee on Polygenic Prediction and Embryo Selection

Monday, January 24, 2022

Supreme Court To Take Up Harvard, UNC Affirmative Action Case


By coincidence, I was just in contact over the weekend with several of the people involved in the effort to end discrimination against Asian Americans in elite college admissions. 

This has been a long road, but perhaps victory is near. 
Supreme Court To Take Up Harvard, UNC Affirmative Action Case (Harvard Crimson)
... SFFA founder Edward J. Blum, who has spearheaded more than two dozen lawsuits challenging affirmative action and voting rights laws around the U.S., heralded the court’s move. “Harvard and the University of North Carolina have racially gerrymandered their freshman classes in order to achieve prescribed racial quotas,” he wrote in a statement. “Every college applicant should be judged as a unique individual, not as some representative of a racial or ethnic group.”
See previous posts: 

... The facts are just so embarrassing to Harvard that with some modest adjustment in its admissions practices it might be able to absorb a judgment against it and get on with life more or less as usual. The vagueness of the category on which Harvard was relying to make sure that it kept its Asian undergraduates at the level that it wished, the so-called personality score, is such a floppy nothing of an empty basket — that’s not gonna do anymore. 
There is something profoundly disturbing about Harvard using these flaccid categories to achieve something like a quota. The court papers show how the system was invented to keep the number of Jews down in the late 1920s and early 1930s. It’s all pretty bad, and part of the badness is that colleges have been both compelled and allowed to do what they’re doing under the rubric of "diversity," which conceals from view the actual operation of the whole system, and what they are in fact aiming to achieve. It’s substituting one vocabulary for another in a way that creates a climate of dishonesty. What goes on in the admissions office is increasingly mysterious, and what happens once students are admitted — that is something to which little attention is paid by educators themselves. 


Harvard Office of Institutional Research models: explicit racial penalty required to reproduce actual admit rates for Asian-Americans





Sunday, January 16, 2022

Preimplantation Genetic Testing for Aneuploidy: New Methods and Higher Pregnancy Rates (re-post with video)

The post below appeared originally November 1 2021, just after the annual American Society for Reproductive Medicine meeting in October. The video of Dr. Wiemer's talk was embargoed so I could not include it. However, now that the embargo has passed you can view it at the link below.


 


Comparison of Outcomes from Concurrent Use of 3 Different PGT-A Laboratories, Main program of ASRM 2021, Presented on October 18th by Klaus Wiemer, PhD.

Let me stress again that improved success rates resulting from higher accuracy in aneuploidy screening of embryos will affect millions of families around the world, and over 60% of all IVF families in the US.




[ NOTE ADDED NOVEMBER 12 2021: Research seminar videos from ASRM are embargoed until 12/31. So this video will not be available until then. ]

This talk describes a study of PGT-A (Preimplantation Genetic Testing - Aneuploidy, i.e., testing for chromosomal normality) using 2 different methods: NGS vs the new SNP array platform (LifeView) developed by my startup Genomic Prediction. 

The SNP array platform allows very accurate genotyping of each embryo at ~1 million locations in the genome, and the subsequent bioinformatic analysis produces a much more accurate prediction of chromosomal normality than the older methods. 

Millions of embryos are screened each year using PGT-A, about 60% of all IVF embryos in the US. 

Klaus Wiemer is the laborator director for Poma Fertility near Seattle. He conducted this study independently, without informing Genomic Prediction. There are ~3000 embryos in the dataset, all biopsied at Poma and samples allocated to three testing labs A,B,C using the two different methods. The family demographics (e.g., maternal age) were similar in all three groups. Lab B is Genomic Prediction and A,C are two of the largest IVF testing labs in the world, using NGS.

The results imply lower false-positive rates, lower false-negative rates, and higher accuracy overall from our methods. These lead to a significantly higher pregnancy success rate.

The new technology has the potential to help millions of families all over the world.

Comparison of Outcomes from Concurrent Use of 3 Different PGT-A Laboratories 
Oct 18 2021 annual meeting of the American Society for Reproductive Medicine (ASRM) 
Klaus Wiemer, PhD

While Down Syndrome population incidence (i.e., in babies born) is only ~1 percent, the incidence of aneuploidy in embryos is much higher. Aneuploidy is more likely to result in a failed pregnancy than in the birth of a Downs baby -- e.g., because the embryo fails to implant, or does not develop properly during the pregnancy. 

False positives mean fewer healthy embryos available for transfer, while false negatives mean that problematic embryos are transferred. Both of these screening accuracies affect the overall pregnancy success rate.

Saturday, January 15, 2022

Manifold Returns!

I'm working on the return of Manifold. I just did the first interview yesterday, with James Lee. I'm not sure when it will be released, but soon I hope.

Thanks to everyone who enjoyed the original show. I received a lot of requests to bring it back over the last 18 months since we went on hiatus. Please make suggestions for guests and show topics!

I'll try to get Dominic Cummings for a long interview. For now, see this great piece in the Guardian about his substack: 

Intoxicating, insidery and infuriating: everything I learned about Dominic Cummings from his £10-a-month blog


Here are the 51 episodes from our first run, with transcripts: https://manifoldlearning.com/


Michigan State University owns the copyright to the old stuff, so we may have to create a new channel and web site. This new project will be entirely mine and I will probably allow comments on the YouTube channel -- we turned those off because the university was/is risk averse about free expression ¯\_(ツ)_/¯ 

I hope to have Corey as a frequent guest on the show, but I will be restarting in solo mode. Here's the first episode where Corey and I introduced ourselves. Very rough around the edges, but still fun for me to listen to again.


Monday, January 10, 2022

Recent Papers on Socio-Economic Status and Student Achievement: Marks and O'Connell

I received the message below from Michael O'Connell, University College Dublin, and Gary Marks, University of Melbourne. 

See also this recent post: Social and Educational Mobility: Denmark vs USA (James Heckman), and links therein. 
Dear Scholar, 
 
There is a widely-held perception that many of life’s key outcomes are fundamentally driven by people’s socio-economic status (SES). More specifically, there is a view that children’s educational attainment is largely a by-product of their familial SES. As a consequence of this pervasive paradigm, much of the energy in seeking to ameliorate or resolve poor educational attainment is based around trying to use SES as a social lever. 

However, in the six papers listed below, published between 2019-2022, evidence has been gathered demonstrating that SES is only very modestly correlated with educational attainment. Furthermore, once a child’s cognitive ability is taken into account, even the modest link between SES and attainment diminishes to slight influence. This is true of datasets drawn from international groups of young people, as well as those from the US, UK, or Ireland. Future attempts to aid and study young people experiencing difficulty with educational attainment should be built on an awareness of the limited role of SES. 

Gary N Marks   Michael O’Connell 


1. O’Connell, M. and Marks, G.N. (2022) 
Cognitive ability and conscientiousness are more important than SES for educational attainment: An analysis of the UK Millennium Cohort Study
Personality and Individual Differences, 188 
https://doi.org/10.1016/j.paid.2021.111471 
Highlights Antecedents of educational attainment of great interest Dominant paradigm focuses on SES of children. Cognitive ability and conscientiousness have stronger record in research findings. Using new UK MCS longitudinal survey data, GCSE state exam performance assessed Cognitive ability and conscientiousness explained far more than SES measures 

 

2. Marks, G. N. (2021) 
Is the relationship between socioeconomic status (SES) and student achievement causal? Considering student and parent abilities
Educational Research and Evaluation, 10.1080/13803611.2021.1968442: 1-24. 
Abstract Most studies on the relationship between students’ socioeconomic status (SES) and student achievement assume that its effects are sizable and causal. A large variety of theoretical explanations have been proposed. However, the SES–achievement association may reflect, to some extent, the inter-relationships of parents’ abilities, SES, children’s abilities, and student achievement. The purpose of this study is to quantify the role of SES vis-à-vis child and parents’ abilities, and prior achievement. Analyses of a covariance matrix that includes supplementary correlations for fathers and mothers’ abilities derived from the literature indicate that more than half of the SES–achievement association can be accounted for by parents’ abilities. SES coefficients decline further with the addition of child’s abilities. With the addition of prior achievement, the SES coefficients are trivial implying that SES has little or no contemporaneous effects. These findings are not compatible with standard theoretical explanations for SES inequalities in achievement. 

 

3. Marks, G. N. and O’Connell, M. (2021) 
No Evidence for Cumulating Socioeconomic Advantage. Ability Explains Increasing SES Effects with Age on Children’s Domain Test scores 
Intelligence, 88    
https://doi.org/10.1016/j.intell.2021.101582 
Highlights Data analysed for five domains for children of the NLSY79 mothers study. SES effects increase for only some domains and not substantially. No increase in SES effects when considering mother's or children's prior ability. Effects of child's prior ability on test scores increase substantially with age. SES effects are small net of mother's ability. 
 
4. Marks, G. N. and O'Connell, M. (2021) 
Inadequacies in the SES–Achievement model: Evidence from PISA and other studies 
Review of Education, 9(3): e3293. 
https://doi.org/10.1002/rev3.3293 
Abstract Students’ socioeconomic status (SES) is central to much research and policy deliberation on educational inequalities. However, the SES model is under severe stress for several reasons. SES is an ill-defined concept, unlike parental education or family income. SES measures are frequently based on proxy reports from students; these are generally unreliable, sometimes endogenous to student achievement, only low to moderately intercorrelated, and exhibit low comparability across countries and over time. There are many explanations for SES inequalities in education, none of which achieves consensus among research and policy communities. SES has only moderate effects on student achievement, and its effects are especially weak when considering prior achievement, an important and relevant predictor. SES effects are substantially reduced when considering parent ability, which is causally prior to family SES. The alternative cognitive ability/genetic transmission model has far greater explanatory power; it provides logical and compelling explanations for a wide range of empirical findings from student achievement studies. The inadequacies of the SES model are hindering knowledge accumulation about student performance and the development of successful policies. 
 
5. O'Connell, M. and Marks, G. N. (2021) 
Are the effects of intelligence on student achievement and well-being largely functions of family income and social class? Evidence from a longitudinal study of Irish adolescents
Intelligence, 84: 101511. 10.1016/j.intell.2020.101511 
Highlights Power of cognitive ability and social class contrasted. Large representative sample from longitudinal study, waves 1–3, of 6216 children Outcomes were attainments, difficulties and relationships. Cognitive ability explained large amounts of variance. Social background only minor effects 
 
6. O'Connell, M. (2019) 
Is the impact of SES on educational performance overestimated? Evidence from the PISA survey 
Intelligence, 75: 41-47 
https://doi.org/10.1016/j.intell.2019.04.005 
Highlights Policy-makers overly attribute differences in educational performance to SES. PISA survey used to assess roles of parental education and household income. Combining them concealed differences in outcomes between rich and poor countries. Household income important in poor countries, parental education in rich countries.

Twilight Struggles: Kazakhstan edition


If you are scratching your head about what happened in Kazakhstan (perhaps as you were last summer about Afghanistan, remember that?), it may be because you only look at mainstream Western sources of information. You might also be the kind of person who swallows whole US propaganda stories about Ukraine, Syria, Xinjiang, Huawei, January 6, RussiaGate, etc.
Dmitry Orlov on January 09, 2022 · at 5:22 pm EST/EDT [ Comment on this blog post by the Saker ] 
What happened in KZ was a paramilitary attack meticulously organized but launched in haste by Western intelligence that had the goal of destroying the statehood of KZ. It was not an attempt to take it over (no time for that) but simply to destroy. The entire state structure was sufficiently rotten that the defense/security agencies couldn’t even pick sides and became demoralized and inactive, but once the Russians were called in to help they immediately knew which side would win and fell back in line. The West’s goal was to set KZ ablaze prior to the talks in Geneva in order to have a better negotiating position vis-à-vis Russia: “You want to divide spheres of influence? Well, we already did that for you—in Kazakhstan!” Keep in mind, the RU-KZ border is open, undefendable and almost 8000km long, running from Volgograd to Tomsk in Siberia, making KZ, as a failed state, a major headache for Russia. Obviously, Russia knew that KZ, rife with Western NGOs and accompanying corruption, and with a weakening economy, could easily be tipped over, and prepared for just this case. Now that the attack on KZ statehood has failed and a mop-up operation is in progress, this has given Russia a huge trump card for the Geneva talks. The West has played its cards and lost. There will be no more color revolutions in the post-Soviet space. Its operatives in KZ are being hunted down and eliminated. Those in positions of authority in KZ have learned the same lesson as Lukashenko: they cannot trust the West; they have to trust Moscow.
On any particular issue Orlov might be right or he might be wrong, but guaranteed on certain topics he knows a lot more than the "experts" found on television or in the NYTimes. In recent years I have spent significant time with Western foreign policy and defense "experts" in think tank settings and I have to say that they are often poorly informed or miscalibrated in the confidence levels assigned to their predictions. Sadly, elites in the West have largely been fooled by their own propaganda, and often have entirely unrealistic views of what is really happening in the world. Alternative sources of information, especially individuals with good local knowledge, are always useful.
Wikipedia: Dmitry Orlov (Russian: Дми́трий Орло́в; born 1962) is a Russian-American engineer and writer on subjects related to "potential economic, ecological and political decline and collapse in the United States", something he has called "permanent crisis".[1] Orlov believes collapse will be the result of huge military budgets, government deficits, an unresponsive political system ... 
Orlov was born in Leningrad (now Saint Petersburg) and moved to the United States at the age of 12. He has a BS in Computer Engineering and an MA in Applied Linguistics. He was an eyewitness to the collapse of the Soviet Union over several extended visits to his Russian homeland between the late 1980s and mid-1990s. ... 
In 2006 Orlov published an online manifesto, "The New Age of Sail." In 2007 he and his wife sold their apartment in Boston and bought a sailboat, fitted with solar panels and six months supply of propane, and capable of storing a large quantity of food stuffs. He calls it a “survival capsule.” ...

Here is a different (internal coup and counter-coup) interpretation of events which is quite unlike Orlov's. I have not seen any evidence presented yet of foreign involvement, but perhaps I am not looking at the right sources...

Friday, January 07, 2022

Ryan Petersen (Flexport); the best Generalist talent is in tech startups...

Dominic Cummings has a great substack, which I highly recommend. 

I just read his recent post 


which led me to the podcast interview below with Ryan Peterson, CEO and Founder of Flexport. It's one of the best startup founder interviews I've heard in a long time.

 

See also this YC blogpost about the interview.

Is there any doubt that the best "generalist" talent among people aged 20-35 is in tech startups? Petersen understands more about the world (or, at least, how to get stuff done in the world) than any typical business school or social science prof, or any typical big company exec or government official.

Dan Wang of Gavekal is a great China tech and economics analyst -- well calibrated, in my opinion.
... China has strong entrepreneurs as well as a strong state, and these two sometimes reinforce each other. An interesting fact I noticed recently is that the party secretary of Zhejiang province, one of the country’s most important, used to be a director of China’s manned space program. A skim through the Wikipedia pages of provincial party secretaries would reveal a diverse range of technocratic experiences. 
An important factor in China’s reform program includes not only a willingness to reshape the strategic landscape—like promoting manufacturing over the internet—but also a discernment of which foreign trends to resist. These include excessive globalization and financialization. Beijing diagnosed the problems with financialization earlier than the US, where the problem is now endemic. The leadership is targeting a high level of manufacturing output, rejecting the notion of comparative advantage. That static model constructed by economists with the aim of seducing undergrads has leaked out of the lecture hall and morphed into a political justification for only watching as American communities of engineering practice dissolved. And Beijing today looks prescient for having kept out the US social media companies that continuously infuriate their home government. ...

Only a retard (autistic economist) can overlook the pitfalls of blind acceptance of Comparative Advantage (Ricardo, etc.): "Gee, those guys are great at making hypersonic missiles and targeting radar. We'll let them do it and just buy the stuff from them. Everybody wins!" See Charlie Munger, Ricardo and finance.

Thursday, January 06, 2022

BOLA2 Copy Number Variation: Phenotype Effects From A Human Accelerated Region

Human Accelerated Regions (HARs) are regions of DNA that were conserved throughout prior (e.g., vertebrate) evolution but are significantly different in the human genome.
Allen Institute: ... of the known 3,171 human accelerated regions, 99 percent of these human-specific mutations fall into "non-coding" regions of DNA, or regions of DNA that don't contain instructions for making a protein. Many of them are in stretches of our genome known as enhancers, regions which regulate nearby genes, and about half of those are nestled in enhancers that are active in the developing human brain.
Our analysis of DNA regions used in predictors for common diseases and complex human traits found that large portions of phenotype variance reside in non-coding regions. This has important consequences for pleiotropy and for our understanding of genetic architecture. 

Regarding HARs, in a 2013 post Neanderthals Dumb? I wrote:

This figure is from the Supplement (p.62) of a recent Nature paper describing a high quality genome sequence obtained from the toe of a female Neanderthal who lived in the Altai mountains in Siberia. Interestingly, copy number variation at 16p11.2 is one of the structural variants identified in a recent deCODE study as related to IQ depression; see earlier post Structural genomic variants (CNVs) affect cognition.

From the Supplement (p.62):
Of particular interest is the modern human-specific duplication on 16p11.2 which encompasses the BOLA2 gene. This locus is the breakpoint of the 16p11.2 micro-deletion, which results in developmental delay, intellectual disability, and autism5,6. We genotyped the BOLA2 gene in 675 diverse human individuals sequenced to low coverage as part of the 1000 Genome Project Phase I7 to assess the population distribution of copy numbers in homo-sapiens (Figure S8.3). While both the Altai Neandertal and Denisova individual exhibit the ancestral diploid copy number as seen in all the non-human great apes, only a single human individual exhibits this diploid copy number state.

Modern humans typically have many (e.g., 3-10) copies of BOLA2. In Neanderthals and apes, 2 copies. 
Variation in copy number presumably affects gene expression, even if the actual protein (coding base pairs) structure is not changed. There may be other mechanisms at work, of course.

Mutations in this 16p11.2 region are associated with schizophrenia, autism, brain size, reduced IQ, anemia, and other things. 

Since 2013 a number of papers have investigated the phenotype effects of BOLA2 copy number variation (CNV) and/or the 16p11.2 duplication/deletion. The latter is more complex as it affects multiple genes in addition to BOLA2. In the future, using whole exome or whole genome data in UKB, it should be possible to focus more specifically on effects of BOLA2 CNV.

For reference I note some of the results below.
Phenome-wide Burden of Copy-Number Variation in the UK Biobank (2019) 
16p11.2 C deletion: "We observe significant increases, on the order of one standard deviation, in weight, BMI, hip and waist circumference, reticulocyte count, and Cystatin C measures for these individuals. The larger 593 kb CNV associates with similar measures of body size and fat, as well as hypertension, diabetes/HbA1c, and abdominal hernia. These results are also indicative of effects due to developmental delay; namely, decreased measures of memory, higher Townsend deprivation (an index of material deprivation which considers employment, home/auto ownership, and household overcrowding in a person's neighborhood) ..."   
Note the effect sizes, e.g., on Townsend deprivation index, are extremely large, roughly 1 SD. The effect size for Prospective Memory score (related to ability to read, remember, and execute directions) is 2 SD!

 

 

Medical consequences of pathogenic CNVs in adults: analysis of the UK Biobank (2019)
Population percentage in parenthesis: 

See also:

The Human-Specific BOLA2 Duplication Modifies Iron Homeostasis and Anemia Predisposition in Chromosome 16p11.2 Autism Individuals (2019)
Quantifying the Effects of 16p11.2 Copy Number Variants on Brain Structure: A Multisite Genetic-First Study (2018)