Commentary on Roseman (2018)

Note: A large part of the paper and my subsequent commentary discusses the interpretation of the arguments that certain individuals in the scientific literature are making and the implications thereof. I am not any of the individuals in question and thus cannot speak directly for them, but only to my interpretation of their work and what I expect they would say. If anyone is still interested in what these people think, I would recommend contacting them.

A recent article by Charles Roseman (2018) aims to come at the race-IQ debate from a metascientific perspective as to what methodologies and techniques produce valid inferences and forms of evidence. Specifically, Roseman, from an anti-hereditarian perspective, charges many anti-hereditarians with (what amounts to me to be) obscurantism and erecting unnecessary burdens of proof for mechanistic explanations.

There are essentially two arguments that Roseman takes issue with. The first is an a priori argument against genetic group differences from the complexity of genetic outcomes. He primarily cites authors in the public sphere, such as Nathaniel Comfort and Gavin Evans, rather than academic criticisms.

Roseman describes Comfort’s criticism of Plomin:

Similarly, in a review of recent book with a hereditarian outlook (Plomin, 2018), Comfort (2018) describes the volume as “yet another expression of the discredited, simplistic idea that genes alone control human nature seems particularly insidious.”

and Evans critique of the revival of scientific hereditarianism:

Intelligence — even the rather specific version measured by IQ — involves a network of potentially thousands of genes, which probably takes at least 100 millennia to evolve appreciably. Given that so many genes, operating in different parts of the brain, contribute in some way to intelligence, it is hardly surprising that there is scant evidence of cognitive advance, at least over the last 100,000 years

At the conceptual level, too warm an embrace of complexity is a denial of the possibility of theory. All understanding relies on abstraction and simplification. If life is so very complex and we ask that every last detail be accounted for before an explanation is accepted, any attempt at a science of human behavior or anything else that includes genes or environment is an exercise in futility

When the stakes are as high as they are in a question like genetic explanations of group differences (Kitcher 1985), some would argue that a much higher burden of proof is necessary to adjudicate the scientific claims. As such, the complexity of the relationship between genes and behaviors – specifically the numerous confounds (Taylor 2007), context-dependency (Moore 2018) and inability to make high-level inferences (MacKenzie 1980) – goes against hereditarianism by undercutting their core claims. The issue is not, per se, that these critics are arguing that the complexity of phenotypic development needs to be completely solved before we can say anything useful about phenotypic development, but more that the complexity of this development precludes us from making the inferences hereditarians need for their scientific claims (Turkheimer 2017).

It may be that environmental effects swamp out genetic effects
or otherwise stymie random genetic drift and natural selection, but there is nothing inherent to complex traits that will slow their evolution over short time spans.

In my simulations, as polygenicity increases (and behavioral traits are the most polygenic traits there are), the variance of the expected mean differences on drift decreases. If correct, this means it is unlikely that drift will be a sufficient explanans for group differences in behavioral traits. Moreover, given the fact that IQ is almost universally posited to be fitness increasing, the increased efficacy of selection in Africa would mean that deleterious variants would be purged more effectively even with negligible fitness differentials (Aris-Brosou 2019). This means that we expect a leftward shift of the expected mean gap on drift towards Africans having a higher genotypic IQ. That is the opposite of the claim posited by hereditarians.

From a more rhetorical standpoint, critiquing and rejecting hereditarian arguments about the relationships among genome, organism, and environment by pointing to a past set of failures behavior genetics and accusing them of oversimplifying a complex situation is a self-defeating strategy

I do not have access to the most recent version of this paper, but I do not understand either the initial claim made here or the evidence marshaled in support of it. The fact that psychiatric genomics has attempted to estimate the influence of genetic effects on psychiatric traits does not obviate the criticisms that critics of hereditarianism have made in the past with respect to the relationship between environments, genes, organisms and development.

If genetic accounts of behavior begin to mimic their morphological and physiological cousins and as more complex studies of genes and behavior are lavishly funded, ground erodes from beneath the palisades of the complexity argument and it is not clear where antihereditarians of the complexity stripe can further retreat.

It seems here that Roseman is arguing that the complexity argument is only contingently successful on the historical and current failure of behavioral genetics to succeed at what the critics see as its goals, but that this could very well change in the future, destroying the basis of the complexity argument. While this may be true for some complexity theorists, it is definitely untrue for Turkheimer. He has long stated that he predicted that GWAS would identify SNPs associated with behavioral traits (Turkheimer 2000, 2012) and has even argued that they migh recapitulate the twin heritability estimates – the hypothetical success of behavioral genetic studies would be less of a condemnation of his position than a confirmation. As for other critics, it depends exactly upon what sort of “more complex studies of genes and behaviors” that Roseman is referring to – I cannot fathom new studies that would eradicate the issue with the complexity of behavioral development, but that may be the poverty of my imagination.

Do we need to have a molecule by molecule account of how a gene works to be confident that there are genetic effects on phenotypic variation?

Here Roseman conflates two related but importantly distinct questions: whether there are genetic effects on phenotypic variation and whether a particular set of SNPs or genes are causally associated with phenotypes. The former has been answered, at least for the vast majority of critics of hereditarianism (as noted by Roseman, Turkheimer’s weak genetic explanation satisfies the ‘genetic effects on phenotypic variation’ requirement). The question these critics are focused on is rather the latter, as the particular set of SNPs that are associated with a phenotype are often used for social and epistemological purposes (Aaronovith 2018; Piffer 2019), ones in contrast to the purposes of anti-hereditarians. As such, it is prudent that the correct SNPs be identified and the manner in which they become associated with phenotypes be ascertained.

It is indisputable that nutritional deprivation stunts growth in children and adolescents and we knew that well before we knew anything about the molecular machinery of growth plates

The difference between this example and purported results from GWAS is one of causality. One can use experimental, quasi-experimental and propensity-score matching designs to estimate the causal impact of nutritional deprivation of growth in youth, but there is no such analogous methodology for genome-wide association studies. Moreover, the mechanism by which two things are associated is against important for the social and epistemological questions that plague the hereditarianism debate.

Similarly, the links between smoking and lung cancer were well established before the details of the physiological operations of any cancer were understood

Typical to those wishing to defend causal interpretations of certain associations without sufficient evidence, the author here appeals to the history of the relationship between smoking and lung cancer. Here, he slips between the question the critics are asking – causality – and what GWAS has actually achieved: “links”. During the beginning of the 20th century, there was certainly strong evidence that associations between smoking and lung cancer existed. One could not yet conclude causality in any definitive form, but only try to make a best guess as to the true effect size. Luckily, a large body of research has confirmed the association studies from the beginning of the century using model organisms and experiments (Khang 2015; Lucas & Harris 2018). What to infer from current research results depends on your statistical approach to inference (frequentist vs Bayesian) and a number of other interpretative and adjudicative factors. Anti-hereditarians who base their critiques on the necessity of mechanistic explanations

We can tell similar stories for phenomena as disparate as the realization that sewers and clean water supplies were important for health (Chadwick, 1842) and the efficacy of vaccination (Lombard et al., 2007). In all of these cases, the big picture of how causal influences were flowing through the world were clear before the molecule- or microbe-eye view of the mechanisms of causation were worked out.

The discussion of epidemiology is again particularly revealing. The question of causality has long haunted epidemiology journals, who have struggled with the philosophy and statistics of demonstrating causation (Glymour 1998; Mackenzie & Pearl 2018; Parascandola & Weed 2001). The existence of epidemiology’s Hill criterion, and several modifications for different subfields and disciplines have established clear criterion for how to create a case for the causal impact of one variable on another. Yet again, GWAS do not fit these criterion, at least according to the critics.

A quick inspection of the diversity of life makes the difficulty with this position clear. A comparison of Ankylosaurus clubs (Arbour and Currie, 2015), Stegosaurus spikes (Cobb, 2009), and the tailless Manx (Todd, 1961) shows that tails can evolve in elaborate ways. The genetic basis of these tail characteristics (outside of the Manx) are as obscure as that of cognitive characteristics as are the vast majority of traits that have ever evolved. Evolution takes place nonetheless, showing that there are mechanisms that generate genetic variation amenable to sustaining evolutionary responses even if our understanding of relationships between genotype and phenotype in any one generation is less than concrete. This is a simple extension of the Darwinian observation that there is no difference in kind between variation within and among species.

Here, I fear Roseman has misunderstood Turkheimer’s distinction between weak and strong genetic explanations. It is not about whether we currently have access to a mechanistic account of the development of the trait and how it becomes associated with an SNP, but whether there exists such an account at all or whether it is coherent to ask for such an account. Turkheimer would say that most physiological traits fall into the category of traits that do in fact have “strong genetic explanations”, but the question is which behavioral traits do and which do not have strong genetic explanations: the real nature-nurture debate (Turkheimer 2016). Moreover, he is confusing several critiques with one another. I doubt Turkheimer has any issue with the evolutionary claims Roseman lists or the ability to identify selection in other species barring a mechanistic explanation. His claims are about the implications of weak vs strong genetic explanations for sociological questions (like approaches to crime).

It is not clear what the demand for strong genetic explanation is supposed to deliver for anti-hereditarians. A mechanistic explanation is no less context dependent than statements about variation at the population level.

If all that exists for behavioral traits ends up being weak genetic explanation (a la Turkheimer 1998, 2016), then it doesn’t seem like hereditarianism is a scientifically decidable hypothesis like hereditarians argue. Moreover, the demand for a mechanism comes first for a demand for the demonstration of causality. Anti-hereditarians are often skeptics of GWAS and argue that the associations identified are confounded by population structure, associative mating, and other forms of gene-environment correlation (Baverstock 2019; Richardson and Jones 2019). A mechanism goes a long way to demonstrating causality, and in some cases actually ends up proving it. The context-dependency of a mechanism only demonstrates the futility of hereditarianism rather than vindicating variational approaches to understanding phenotypic development. Moreover, what anti-hereditarians who view mechanistic explanation as a requisite factor for adjudicating the hereditarian question argue is that certain mechanisms by which SNPs can become associated with phenotypes are not amenable to hereditarianism. For instance, if it ends up being that height is highly heritable (say 80%) and that height causally impacts IQ  (via some socially mediated pathway), then of course IQ will be statistically heritable. But it isn’t “genetic” in the typical colloquial sense that hereditarians need – if simply changing our social norms will decrease the additive genetic variance of the phenotype, that contradicts a great number of claims in the hereditarian literature (e.g. Jensen 1969; Rushton 2012). Indeed, if it ends up being that (in our hypothetical) all of the additive genetic variation for IQ is mechanistically explained in this manner, then I suspect the ‘so-called’ environmentalists would consider themselves vindicated on the matter, as the proximate developmental mechanism is environmental. So understanding the mechanisms helps us understand the inferences that can be made from the association of a SNP to a phenotype – different types of associations have different biological, epidemiological and social implications.

My fear is that choosing to only accept extreme and well-characterized clinical phenotypes or simple Mendelian characteristics with good molecular level mechanistic explanations as genuinely genetic casts genetics in biased ways.

This is a misrepresentation of Turkheimer’s view on how genetic associations should be understood. His point is that a coherent molecular genetics of say, divorce, are not forthcoming. The proper level of analysis for understanding the origin is not “bottom up from the genes”, but at the phenotypic level itself (Turkheimer et. al 2014). These are not any less legitimate genetic associations, but they do not have simple mechanistic explanations. To be clear, Turkheimer doesn’t think there is a fine line between a “true” genetic association (‘true positive’) and a confounded genetic association (‘false positive’) – the question of causality may not end up having a fact of the matter for GWAS. But that does not dismiss the fact that Turkheimer has never claimed that these associations are not “genuinely genetic” or that the heritability is fictitious.

methods

Method 1

Consider N currently identified alleles associated with the trait. Assume there are P alleles associated with the trait across the entire genome.

If the PGS difference computed from those N alleles is PGS_N, then the extrapolated difference is PGS_N \cdot \frac{P}{N}

Benefits:

  • Simple to do
  • Finding numbers of associated alleles across the entire genome is not difficult

Downsides:

  • Highly dependent on your choice of original alleles.
  • Depends significantly on the distribution of effect sizes in your original set of N alleles and P \setminus N (e.g. assumes that that the mean effect size for the alleles in P \setminus N is the same as the mean effect size for the alleles in N)

Method 2

Consider N currently identified alleles associated with the trait. Assume these N alleles explain X% of the variance, but it has been estimated that Y% of the variance in the trait is explained by alleles across the entire genome.

If the PGS difference computed from these N alleles is PGS_N, then we should extrapolate the total PGS difference from genes across the entire genome to be PGS_N \cdot \frac{Y}{X}.

Benefits:

  • Doesn’t require mean effect sizes to be equal between discovered and undiscovered alleles

Downsides:

  • Requires accurate estimation of total heritability and current explained heritability.

Method 3

Consider K currently identified alleles with effect sizes following \beta_K \sim \mathcal{N}(\mu, \sigma^2). Call the frequency difference between two populations (consistently ordered, e.g. always African frequency minus European frequency) at each allele i, \digamma_i. Then the contribution of each allele i to the mean genotypic difference is \digamma_i\beta_i.

In R, fit a GAM model for \digamma_i\beta_i \sim \beta_i. This should give you a good nonlinear fit of the expected contribution to the genotypic gap for an allele of a given effect size \beta. Call this predicted contribution G(\beta).

Then, compare your current distribution of effect sizes \beta_K to the distribution of effect sizes you actually expected when the whole-genome is sequenced, \beta_M. Sum over the whole-genome distribution as such:

\sum_{\text{genome}} G(\beta)

Benefits:

  • Doesn’t require assumptions about the distribution of effect sizes in current GWAS or in the actual set of associated alleles

Downsides:

  • If there are a large number of large effect sizes alleles in your expected actual distribution, then the out of sample prediction can create wonky predictions, especially because the frequencies are going to be way off.
  • If you have a non-random subset of alleles with respect to any one given effect size, it can bias the estimates for the rest of the genome.

jews, genes, and diseases

Jewish Genetic History

Origin

The origin of the Ashkenazi Jewish population is one of the most contentious debate going on in modern population genetic history, with two major camps. One posits that Ashkenazi Jews have substantial Middle Eastern heritage, owing primarily to the Levant, who then migrated out of the region at some point in time into Italy, and then Eastern Europe (Ostrer & Skorecki 2013). The other group posits that Ashkenazi Jews were primarily a group of converts from the Caucus mountain region who donned Jewish culture and customs before moving to Eastern Europe (Das et. al 2017). Whichever group ends up being right, it is quite clear that Jewish population history is very complicated, with many migrations, substantial amounts of converts, and population admixture.

Bottlenecks, Founders Effects and Ne

Despite Cochran et. al (2006)‘s claims that there is no evidence for a bottleneck in Jewish history, there is both testimonial & recorded history and genetic evidence for severe genetic bottlenecks from multiple lines of evidence. Behar et. al (2004b) finds contrasting signals, but show the most parsimonious explanation is a bottleneck with other unique demographic characteristics. The same groups analysis of mtDNA data from Ashkenazi Jewish genomes reaches a similar conclusion, demonstrating a genetic bottleneck (Behar et. al 2004a). Recent work on the Ashkenazi Jewish genomes from the gnomAD data set indicate similar results, showing a bottleneck smaller than most posited in the literature (see e.g. Carmelli & Cavalli-Sforza 1979) of about 350 individuals (Carmi et. al 2014), which has been recently been replicated for Eastern and Western European Jews (Gladstein & Hammer 2018).

Recent research by Bray et. al (2010) demonstrates clear evidence for founder’s effects, showing that the high rates of IBD must be the result of founder’s effects rather than selection. [1] Older research has long-demonstrated that founders effects are necessary to explain the patterns in the data (Morton et. al 1982).

Jewish IQ

One of the key claims in Cochran et. al’s article is that Jewish IQ is so high that it needs an exceptional explanation. Despite their claims, there are still many outstanding questions on the IQ of Ashkenazi Jews. Particularly, the question of the quantity of the alleged Ashkenazi Jewish advantage in IQ is to be resolved, as well as whether this advantaged is limited to Ashkenazi Jewish immigrants, or is general to both Israeli Ashkenazi Jews and diaspora Jews.

The first aspect of their claims about Ashkenazi IQ that should be questioned is the difference between Ashkenazi Jewish IQ and ‘Oriental’ Jewish IQ:

The first is that the Ashkenazi Jews have the highest average IQ of any ethnic group, combined with an unusual cognitive profile, while no similar elevation of intelligence was observed among Jews in classical times nor is one seen in Sephardic and Oriental Jews today.

Cochran et. al (2006)

It has been in fact observed that “Oriental Jews”, when placed in kibbutzum, reach the same IQs as Ashkenazi Jewish individuals in kibbutzum (Hirsch 1972, p. 29, in Hunt 1972, citing Benjamin Bloom).

The first citation Cochran et. al (2006) give in support of the idea that Jewish IQ is high and has been high for quite some time is Russel and Lewis (1900). One can easily verify that the relevant portion of the text that Cochran et. al are citing exists:

In 1900 in London Jews took a disproportionate number of academic prizes and scholarships in spite of their poverty

Cochran et. al (2006)

As the 1900 book states:

The foreign children at the East End board schools are universally allowed to be sharper and more intelligent than the English, and they carry off a large proportion of prizes and scholarships.

Russel and Lewis (1900)

But the issue is that this is not the only discussion that the book has about the intelligence of Jewish children. First, the authors attribute the observed intelligence to three factors. The first factor is:

There can be little doubt that this is due, in part at any rate, to the extra teaching which they receive in Hebrew

Russel and Lewis (1900)

While the other two are mentioned in a footnote:

Two other considerations, however, must be allowed some weight : (i) the greater precocity of Jewish children ; (2) the greater stimulus and encouragement afforded in Jewish homes.

Russel and Lewis (1900)

Clearly Russel and Lewis are not rampant hereditarians attributing alleged Jewish success to biological endowment.

The second consideration in evaluating Cochran et. al (2006)‘s citation of Russel and Lewis (1900) is whether the Russel and Lewis (1900) had sufficient data to support their initial claim that:

The foreign children at the East End board schools are universally allowed to be sharper and more intelligent than the English, and they carry off a large proportion of prizes and scholarships.

Russel and Lewis (1900)

I have been unable to locate any actual statistics in the book or any clear citation(s) as to where they are getting their alleged figures from, making it unclear how to interpret the claim that the Jews of London were more intelligent than gentiles at the time, especially in the absence of a reported mean difference.

Finally, and most problematically, is the claim by Cochran that Jewish migrants to London were concentrated in poverty. While the class position of Jewish migrants in London during that time period is a topic that could be subject to an entire book, it is clear that the positioning of Jewish immigrants as overwhelmed by poverty is not a fair representation. Vaughn (2005)‘s analysis of data on the occupations of Jewish migrants in London shows that they have broadly similar levels of occupational attainments to gentiles, but that this differs by their spatial location and the actual composition. It is worth noting, however, that the methodology used to conduct the analysis is inherently limited by the data of the time, with the occupational strata being crude representations of the actual socioeconomic position of the individuals being measured (see also Lederhendler n.d.). Vaughn (1994) reports that by 1881, ~15% of the Jewish population was in the upper class, while one can note that the listed occupations are quite similar to the ones that comprised the allegedly ‘cognitively demanding’ occupations Cochran cites that allegedly caused selective pressures in the Ashkenazi Jewish population [2]. Moreover, when we actually look at the boom Cochran et. al are citing, we note that they would not describe the Jewish life in London the same way. Russel and Lewis (1900) instead claim that:

It should be clearly recognised that they form a community which is not stagnant in poverty, but everywhere bubbling up with life and enterprise

Russel and Lewis (1900)

They also note that:

but we may note in passing that the almost invariable presence of intellectual interests amongst Jews causes degraded types of poverty to be of rare occurrence amongst them

Russel and Lewis (1900)

A book dedicated to British Jewry documents a transition in the socioeconomic position of British Jews from the early 19th century to the late 19th century, meaning Cochran et. al (2006) may have gotten their dates confused. Endelman (2002)‘s third chapter is titled “Poverty to Prosperity (1800–1870)” and details the path British Jews took from being “impoverished, poorly educated, dependent on low-status street trades and other forms of petty commerce, popularly identified with crime, violence, and chicanery, widely viewed as disreputable and alien” at the beginning of the 19th century (Endelman 2002, p. 79) to being “native English speakers, bourgeois in their domestic habits and public enthusiasms, full citizens of the British state” at the end (ibid, p. 79).

However, it remains uncertain that the groups of Jews referred to when discussing academic accomplishment and the Jews referred to when discussing socioeconomic position are the same, and any extrapolation as to whether the observed scores are concordant or discordant with environmentalist theory cannot be made. Regardless, to the extent that we trust Russel and Lewis (1900)‘s unevidenced claims of Jewish academic success, we should similarly accept their observation of low poverty rather than the high poverty Cochran et. al (2006) claims.

In their paper, Cochran et. al (2006) also cite Hughes (1928) [3], a fellow progenitor of the theory of the innate superiority of Jews over gentiles. As such, it deserves much more careful examination, especially given that it claims to provide conclusive proof free from confounds.

The questions employed in the evaluation of intelligence seem to be particularly culturally limited:

General Intelligence.
If it takes longer to boil an egg than it does to boil potatoes, write “egg”; if not, write XXIII in ordinary figures.

If MARMALADE contains more M’s than MINIMUM, write M; if not, say how many capital M’s there are in this question.

In the following the right word has to be underlined:
Ancient. Antique. Same, opposite, unknown

money people much- not have poor . True, false, not known.

(kind, tall, short) man is (often, always, never) cruel to animals

May is darker than Lucy, but fairer than Kate. Who is fairer-Lucy or Kate? .. Kate, Lucy, I cannot tell.

Hughes (1928)

Moreover, it seems that this investigation was conducted by the very same Cyril Burt who has long faced accusations of fraud against him (Tucker 1994; Tucker 1997):

In I926 a grant from the Jewish Health Organisation of Great Britain made possible the carrying out of an extensive investigation into the subject in London elementary schools. It was conducted by the present writer, with the assistance of Miss Mary Davies and under the direction of Professor Cyril Burt.

Hughes (1928)

Making its veracity questionable, at best.

Finally, we should examine the actual composition of the sample. Given that the authors do not list the schools actually in their sample, it makes it problematic for assessing the representativeness of the results (as later noted by Brill 1936)

Three schools were chosen, each containing about an equal number of
Jews and non-Jews in one and the same school. The three schools were representative of three levels of social and economic status: thus School A, situated in North London, contained both Jews and non-Jews from better class homes; Schools B and C, situated in East London, contained, respectively, poor and very poor children.

Hughes (1928)

Despite their claims, there is contrary evidence both from the beginning of the century (from the 1900s to the 1940s) and the end (1950s-1990s) as to the existence, nature and magnitude of the Jewish IQ advantage.

For instance, Murdoch (1920) [cited in Young (1922)] found that Jews and white Americans scored the same, above Italians and “negroes [sic]”. Jordan (1921) found no consistent differences in intellectual functioning between Romanian Jews, Russian Jews and native whites. Pintner & Keller (1922) reported that Jewish children have an IQ of 95, slightly below the norm of 100. Feingold (1924) reported that Jewish children were slightly below that of American and foreign-born children in junior year of high school.

Wolberg (1927) (cited in Garth 1930) administered a test on the recognition of geometric figures and found that Jews performed worse than gentiles. Pintner & Arsenian (1937) used the Pintner intelligence test (Pintner 1927) [a typical early 20th century IQ test wherein IQ was calculated by the division of ‘mental age’ by ‘biological age’] and found that Jews averaged around 102 points on both the language and non-language sections, barely above the norms. Brown (1940)‘s systematic comparison of the intelligence of Jewish and Scandanavian children found that there were no differences within the same occupational groupings. The study was representative, systematic, used proper metrics of socioeconomic status and controlled for other errors in the field. Held (1941) was also unable to find consistent differences on the American Council Psychological Examination. Shuey (1942)‘s (re)analysis of the performance of Jewish college students found that they are below Protestants, but that the gaps decrease when relevant factors are taken into account.

More recently, Boris Levinson has performed a number of studies evaluating the intelligence of Jewish students, particularly examining the origin of their unique factorial structure (Levinson 19571958, 1960). A more detailed assessment of the merits and implications of Levinson’s research must be relegated to a later date.

As noted by Ferguson (2008), even more recent literature gives equivocal answers. For instance, Dershowitz & Frankel (1975) have the average of Ashkenazi Jewish IQ below that of White Anglo-Saxon Protestants, while the data from Project Talent indicates a Jewish IQ of only 102 to 103 (Seligman 1992; calculation from here).

IQism

Another important part of Cochran’s claims are the typical parts of the IQist mythology: IQ predicts everything, IQ is super/highly heritable. Both of these claims, however, have a lot more dust on them than one might think.

First, we should consider the predictive aspect:

IQ tests predict a host of characteristics of individuals including educational attainment, job performance, income, health, and other non-obvious characteristics like susceptibility to Alzheimer’s disease

Cochran et. al (2006)

The authors here confuses “prediction” and “understanding”/”explanation”, “prediction” and “correlation”, and “prediction” and “causation”. These are three important mistakes. The first mistake is confusing the ability to predict something with the ability to explain or understand something. We can predict all sorts of things using highly complex mathematical algorithms, but this has little to do with our ability to understand the core questions at the heart of science (Shmueli 2010). Moreover, the claims that IQ “predicts” something are built on a small set of correlations. If correlations were prediction, then surely all social sciences are in good luck! Social scientists have accumulated correlations for decades, which has been the cause of their downfall. They have confused correlation and prediction: being able to identify a correlation tells you less about what you will be able to predict will happen (Ferber 1956; Taleb 2019). Indeed, when it comes to IQ, the correlations with other variables are small at best, and decrease when you stop the usage of poor statistical practices (Taleb 2019). The final, and most important mistake, is the one that claims IQ actually causes these outcome variables, a much higher bar that not a single IQ study has met thus far.

In general the search for social and nutritional causes of IQ differences has not led to any convincing results and most workers now regard IQ as a biological rather than a social variable. It is highly heritable—correlations between identical twins reared apart are 0.7–0.8

Cochran et. al (2006)

This statement is not only false, but it is self-evidently false, as even hereditarians like Lynn (1990) and Rindermann (2018) concede that nutrition and education (the latter) causally influence IQ. It is absurd to claim that having ones head bashed in or being starved for the first half of someone’s life will not causally impact their observed test scores, the question is to what extent do environmental influences on IQ persist, and there is substantial evidence that it goes quite a ways (Ritchie et. al 2018).

IQ test scores are highly heritable, almost always greater than 0.5 when adult scores are studied.

Cochran et. al (2006)

IQ test scores are certainly heritable (Turkheimer 1998, 2016), but the actual magnitude and interpretation of reported h^2 figures are of great controversy. For instance, several groups in the 1970s, 80s and 90s have reported heritability coefficients below Cochran’s lower bound of 0.5 and far below his preferred figure of 0.8 (Feldman & Ramachandran 2018).

Lower heritability estimates are found for children’s IQ: the IQ of children does seem to reflect in part environmental influences like the social class of the home in which the child is reared, but these influences disappear as the child matures and are essentially gone in adulthood.

Cochran et. al (2006)

It is unclear whether environmental influences “disappear”, become increasingly confounded by behavioral genetic modeling assumption violations, or if environmental influences change from those influenced by the home to those of peers and such (Dickens & Flynn 2001).

In the same way enrichment programs like Head Start cause a transient elevation in IQ scores of children but these effects disappear as the child matures.

Cochran et. al (2006)

A more parsimonious explanation of Head Start is that stopping environmental supplementation stops the increase of IQ.

The heritability of IQ is probably lower than 0.80 in most human populations, and it may be as low as 0.50, so there are apparently some environmental effects on IQ. Since siblings and twins raised apart are as similar as those raised together, it has become commonplace to speak of “non-shared environment”, which means that siblings are exposed to different environments even when raised together. It is important to realize that so-called environmental effects include non-additive gene interactions like dominance and epistasis as well as testing error. The correlation between one IQ test score and another taken later may be as low of 0.8 or so

Cochran et. al (2006)

While it is true that “non-shared environment” is an unclear concept (Turkheimer & Waldron 2000), it is not true that dominance and epistasis are captured in the non-shared environment component. Because the effects of dominance scale with relatedness, they will be captured in the additive genetic component, while epistasis (in twin studies [4]) does the same. It is also unclear that dominance and/or epistasis contribute to any of the variance in IQ scores. While it is clear that if measurement error exists, then it would attenuate the heritability of IQ scores, but again it is not clear that a classical test theory that claims there are “true scores” that observed scores deviate from is an accurate reflection of reality (Muir 1977). Finally, there have been specific unshared environmental factors identified that causally influence IQ scores (Daw et. al 2015), as well as modifiers of those factors.

So far, after intensive searching, no one has found any, and the current consensus is that variation in IQ reflects variation in the underlying biology rather than in the social environment

Cochran et. al (2006)

This is both erroneous (as described above) and absurd. The “consensus” is that IQ test scores arises from a cascading escalation of development (Briley et. al 2019; Johnson et. al 2011) involving both processes of genes and environments. There are clearly identified causal environmental factors for intelligence, while the genes have been difficult to identify reliably at this point in time.

Quantitative traits like height or IQ are influenced by many genes. The response of quantitative traits to selection is described by the fundamental relationship

R = h^2\cdot S

where R is the response to selection, S is the selection differential, the difference between the mean value in the population and the mean value in parents, and h^2 is the narrow sense heritability of the trait.

Cochran et. al (2006)

In contrast to Cochran et. al’s claims that “this simple robust formulation is applicable to animal breeding, laboratory experiments, and evolution in natural populations”, the assumptions of the Breeder’s equation are almost ubiquitously violated in natural populations (Pujol et. al 2018) making its application to humans especially problematic (Morrissey et. al 2010).

Detailed demographic data about early medieval Ashkenazim are lacking, but we can infer plausible parameters from the scarce information that we do have. First, their jobs were cognitively demanding since they were essentially restricted to entrepreneurial and managerial roles as financiers, estate managers, tax farmers, and merchants. These are jobs that people with an IQ below 100 essentially cannot do. Even low-level clerical jobs require something like an IQ of 90 (Gottfredson, 2003).

Cochran et. al (2006)

The most notable claim in Cochran et. al (2006)‘s reasoning here is that the jobs that the Ashkenazim inhabited during that period were jobs that only people with exceptionally high IQs could perform, giving us reason to believe that the only jobs available to them forced their IQs to increase. However, it is not clear that there is any meaningful correlation between occupational attainment and intelligence (Hauser 2010; Huang 2001). Moreover, it is manifestly false that those jobs “require” certain IQs as Gottfredson claims, especially given the fact that we are talking about hundreds of years ago, where the actual tasks that comprised particular jobs differed dramatically from our increasingly technologized world.

Moreover, there’s good reason to believe that Cochran’s understanding of Jewish economic history is similarly flawed. For instance, it has been documented that Jews were actually slightly poorer than gentiles in Anglo-Saxon England (Mell 2017, p. 155-216). Given that Cochran claims that the boost in IQ occurred after the classical period (specifically he claims that there has never been an observation of the Jewish population being smart before 1400), the fact that the Jews of Perpignan were prominent moneylenders in the 1200s (Emery 1959) should at least be a disconfirmation of his thesis [5]. In fact, the timing of Jewish restriction to particularly occupations and Jewish success are not historically aligned in a way that makes his thesis coherent. For instance, the transition of Jews into the trades, finance and craft professions occurred in the 8th century (Botticini & Eckstein 2003) [6], not past the 1400s, meaning that the purported selection pressures would have had to have started much earlier than Cochran supposes.

Despite their later discussion of the actual timing of the occupational transition, it is clear that the actual magnitude of the transition of Ashkenazi Jews into the urban professional occupations he discusses happens far before the period of Jewish “smartness” he allegedly observes throughout the historical literature (see Botticini & Eckstein 2004, table 1). Indeed, “by 900 the overwhelming majority of the Jews in Mesopotamia and Persia were engaged in a wide variety of crafts, trade, moneylending, and medicine” (Botticini & Eckstein 2013), while Jewish philosophy became prominent centuries before the first period Jews are professed to be intelligent (Jospe 2009). In contrast, Cochran et. al (2006) claims that:

The Ashkenazi occupational pattern was different from that of the Jews living in the Islamic world. The Jews of Islam, although reproductively isolated, did not have the concentration of occupations with high IQ elasticity. Some had such jobs in some of the Arab world, in some periods, but it seems it was never the case that most did.

Cochran et. al (2006)

Moreover, when it comes to the authors’ citation of Lewis in describing the occupations of “the Jews of Islam”:

In fact, to a large extent, and especially during the last six or seven
hundred years of relative Moslem decline, the Jews of Islam tended to have “dirty” jobs (Lewis, 1984). These included such tasks as cleaning cesspools and drying the contents for use as fuel—a common Jewish occupation in Morocco, Yemen, Iraq, Iran, and Central Asia. Jews were also found as tanners, butchers, hangmen, and other disagreeable or despised occupations. Such jobs must have had low IQ elasticity; brilliant tanners and hangmen almost certainly did not become rich

Cochran et. al (2006)

there is a misrepresentation. While the discussion of “dirty jobs” does appear in Lewis (1984), it is contextualized by a discussion of what is actually constituted by these “dirty jobs”. For example, the Ottomans considered dealing with foreigners the “dirtiest trade of all” (Van den Boogert 2013). As such, Jews were heavily concentrated in “diplomacy, commerce, banking, and brokerage”. However, both Lewis’ claims of the presence of Jews in the “dirty jobs” and in “the dirtiest trade of all” (Van den Boogert 2013) come without sources, leading us to question the actual occupational composition of the Jews of Islam in this time period as stated by Lewis, and therefore by Cochran.

However, a quick perusal of different sources tells us a different story than the one that Cochran presents. The Jewish community in Istanbul was primarily Sephardic and Romaniot, although there were often small communities of Ashkenazim (Rozen 2002, p. 51). Moreover, they were prominent in the upper echelons of the bureaucracy, taking roles as statesmen (ibid, p. 209), physicians (ibid, p. 209) and businessmen (ibid, p. 230, p. 239). A review of the Jewish population in the Maghreb will tell similar stories of their wealth and accomplishment (Fenton & Littman 2010, p. 152), as does an analysis of Jews throughout the Middle East (Kavon 2017). Just like that has characterized all of Jewish history, there as intense social stratification with some Jews being “very rich–merchants and tax farmers, court doctors, and people with close ties to the Ottoman court; they lived in stone palaces, owned slaves and maidservants, dressed in silk and velvet, and were surrounded by luxurious furniture and objects from abroad” (Rozen 2002, p. 241), whereas others were “workers who lived from hand to mouth; they and their families dwelled in a single room of a wooden house shared with many other families” (ibid, p. 241). In summary, the economic conditions of Ottoman Jews did not significantly differ from the actual conditions of Ashkenazim in Europe (Mell 2017, 2018). If Cochran et. al (2006)‘s thesis as to how intelligence was selected for in Ashkenazi Jews is to be considered legitimate:

The third is that they experienced unusual selective pressures that were likely to have favored increased intelligence. For the most part they had jobs in which increased IQ strongly favored economic success, in contrast with other populations, who were mostly peasant farmers. They lived in circumstances in which economic success led to increased reproductive success.

Cochran et. al (2006)

then it should also be represented in the descendants of Middle Eastern Jews

An unsanitized history of the Ashkenazim in Europe will not tell the stories of Jews wholly in the professional ‘cognitively demanding’ occupations Cochran’s theory require. They were heavily involved in the arts and crafts (Abrahams 1932, p. 217), the production of silk (ibid, p. 219-220), the dying of cloth (ibid, p. 219), tailoring cloth and wool (ibid, p. 222-224) [7].

He describes the Jews of Roussilon:

In some cases, we have fairly detailed records of this activity. For example (Arkin, 1975, p.58), concerning the Jews of Roussilon circa 1270: “The evidence is overwhelming that this rather substantial group of Jews supported itself by money lending, to the virtual exclusion of all other economic activities. Of the 228 adult male Jews mentioned in the registers, almost 80 percent appear as lenders to their Christian neighbors. Nor were loans by Jewish women (mostly widows) uncommon, and the capital of minors was often invested in a similar manner. Moreover, the Jews most active as moneylenders appear to have been the most respected members of the community.”

Cochran et. al (2006)

In contrast, the Jews of Marseille participated in the ‘cognitive demanding’ occupations of trade at similar rates to which they were represented in the population (Mell 2018, page 120). Indeed, the Jewish labour movement has an extremely long history (Abrarasky 1971; Gorny 1983; Lockman 1996; Mendes 2013; Weinstein 2018), one hard to explain by positing Jews were universally and grossly disproportionately in the upper echelons of society, in bourgeois and petit bourgeois occupations that had high cognitive prerequisites. As such, the alleged uniformity of Jewish occupations during this time period should be dismissed (Mell 2018, p. 192), instead replacing our understanding of Jewish labour in terms of its temporal, geographic and historical specificity. It has long been noted that “neither in the Middle Ages nor during the early modern period was every Jew a merchant”, but that “the majority of the Jewish population was occupied in small-scale urban trades and crafts” (Trivellato n.d.).

Assume, for example, that the correlation between income and IQ is 0.4 (about the correlation in the United States today)

Cochran et. al (2006)

As a matter of fact, the correlation between income and IQ is much lower at .15 to .2 (Strenze 2007), but we have doubts as to the statistical accuracy of those reports (Taleb 2019). We should assume that whatever the correlation today, it was even smaller in the past.

and that individuals in the top 10% of income have twice the average fitness.

Cochran et. al (2006)

The reasoning for this figure is not given – whether it comes from observed data or pure guesswork is not reported, and it does not seem very plausible.

The mean wealth of parents would be .16 standard deviations above the population average and the mean IQ of parents would be 0.4×0.16 or 0.064 IQ standard deviations, that is 1 IQ point above the population mean. This is the selective differential, and with a heritability of 0.8 IQ would increase by 0.8 points per generation. In 500 years—20 generations—average IQ would increase by 16 points.

Cochran et. al (2006)

Under a number of restrictive assumptions, particularly that the fitness advantage above actually causes selection (Preston & Campbell 1993) and that the variability in fitness effects is small enough (Graves & Weinreich 2017).

Jewish History

More Genetics

Assume that a population experienced selective pressures similar to those posited for the Ashkenazim, such that the parents of the next generation averaged one half an IQ point higher than the current average while also experiencing significant gene flow from the general population. For 10% gene flow and a narrow-sense heritability of 0.7, the maximum IQ increase over many generations would be 3.2 points. For 20% gene flow, the maximum increase would be 1.4 points. Clearly, any significant amount of gene flow greatly inhibits local adaptation. We know, however, that the Ashkenazim experienced very limited inward gene flow, on the order of 0.5% per generation (Hammer et al., 2000).

Cochran et. al (2006)

Again, if we look closely at the claims the authors make, specifically “We know, however, that the Ashkenazim experienced very limited inward gene flow, on the order of 0.5% per generation”, citing Hammer et. al (2000). However, Hammer et. al’s analysis was very limited, specifically due to its use of haplotype analysis to infer the total contribution of European ancestry. More recent estimates, like those of Xue et. al (2017) or Carmi et. al (2014) indicate that Hammer et. al’s estimates ranging from 13% to 23% are quite low. Consequently, rates of gene flow should be adjusted upwards [8]. However, there should be more careful consideration given to the timing of the influxes of genes from other populations, which needs to be explicitly modeled with population genetic techniques. Cochran et. al do not do that.

Social History

The ancient Jewish population suffered remarkable vicissitudes—the Babylonian exile, the Hellenistic conquest and Hasmonean state, the revolts against the Roman Empire—but most of that history is probably irrelevant to our thesis, except to the extent that it helped create necessary cultural preconditions

Cochran et. al (2006)

The issue with Cochran et. al’s statement here is their uncritical citation of the standard historical narrative. It is, in fact, false that there exists such a thing as the “Babylonian exile” (Grabbe 2014; Moore & Kelle 2011; Sand 2009).

The key cultural precondition among the Jews was a pattern of social organization that required literacy, strongly discouraged intermarriage, and that could propagate itself over long periods of time with little change

Cochran et. al (2006)

If Jews had a pattern of social organization that discouraged intermarriage, it was not very successful. It has been observed that the inbreeding coefficient for Ashkenazi Jews is actually lower than of Europeans, indicating that intermarriage was more common among Ashkenazi Jews (Bray et. al 2010).

Jewish Diseases

Lysosomal Storage Disorders

Gaucher Disease

Cognitive and Neural Effects

Gaucher Disease Type I is known to be highly comorbid with deleterious neurocognitive diseases such as Parkinson’s disease (Cherin et. al 2010; Lal & Sidransky 2017).

Despite Cochran’s claims that:

A prediction is that Gaucher, Tay-Sachs, and Niemann-Pick heterozygotes will have higher tested IQ than control groups, probably on the order of 5 points.

Cochran et. al (2006)

That particular claim has already been falsified. McNeill et. al (2012) has documented that both “Gaucher disease patients … and carriers” had lower cognitive assessment scores than controls, despite their sample being comprised primarily of Ashkenazi Jewish patients. Gaucher disease type III patients are also known to have IQs in the range of 80-85, with cognitive deficits being common (Abdelwahab et. al 2017; Goker-Alpan et. al 2008). As noted by Ferguson (2008), unpublished data from the Shaare Zedeck Gaucher clinic indicates that carriers and their normal siblings have identical IQs, indicating that there is no heterozygote advantage.

Slatkin (2004) demonstrated that drift can explain the high frequency of Gaucher disease in Ashkenazi Jewish populations if the effects of the disease are purely recessive. Research by Peleg et. al (1998) indicates that a double heterozygote advantage for carriers of Gauchers and Tay-Sachs is not observed in the data, meaning that any purported heterozygote advantage for the two diseases are likely to be distinct, making the common IQ cause implausible. Finally, Diaz et. al (2000) gave good reason to suspect drift in a bottleneck is the reason for its high frequency.

Mucolipidosis Type IV

Mucolipidosis type IV is another sphingolipid storage disorder known for high frequency in Ashkenazi Jewish populations (Bargal et. al 2001).

Neurocognitive Effects

Both neuropsychiatric and neuroimaging data show that Mucolipidosis type IV patients have delayed brain growth (Fisher et. al 2017; Geer et. al 2010; Schiffman et. al 2014). In model organisms, heterozygotes have the same phenotypes as controls (homozygous for not having the allele) (Boudewyn & Walkley 2018; Chacon et. al 2019, fig 1.d; Walker & Montell 2016).

Founder, Selection or Neither?

Data from Ashkenazi populations in Israel indicate that the prevalence of the disorder in the population can be explained almost completely by two alleles (Bargal et. al 2001), indicating founders effects, as they correspond with only two haplotypes (Slaugenhaupt et. al 1999).

Tay-Sachs Disease

The evolution of Tay-Sachs, a highly lethal recessive disorder, in the Ashkenazi Jewish population has long been an outstanding question in evolutionary biology and population genetics (Cavalli-Sforza 1973; Chase & McKusick 1972; Ewens 1978; Myrianthopoulous, Naylor & Aronson 1966; Valles 2009).

Drift, Selection or Neither?

Various models have been developed as to the likelihood of observing the prevalence of the disease in the Ashkenazi Jewish population under different conditions (e.g. genetic drift) that have come to varying conclusions. Rao & Morton (1973) argue that genetic drift is particularly plausible for Tay-Sachs given particular statistical assumptions, as have others (Wagener et. al 1978), while others have reached more equivocal results (Chakravarti & Chakaborty 1978; Myrianthopoulous, Naylor & Aronson 1972; Yokoyama 1979). If we consider the results from Yokoyama in tandem with recent results for the effective population size of Ashkenazi Jewish populations, then we can see that drift is quite a plausible hypothesis. The particular geographic composition of Tay-Sachs carriers (Petersen et. al 1983) has led some authors to conclude that drift is a more likely explanation than selection (Risch et. al 2003a).

All evidence for heterozygote advantage thus far has been quite weak (Myrianthopoulous & Aronson 1966; Shaw & Smith 1969; Spyropoulos et. al 1981), and there are several genomic indicators that drift is a better explanans for the frequency of the allele (Frisch et. al 2004; Slatkin 2004).

There have been models other than drift and selection proposed to explain the frequency of the disorder. For instance, some have proposed reproductive compensation for children with rare lethal diseases (Koeslag & Scach 1984; Koeslag & Scach 1985; Koeslag, Scach & Melzer 1984), as well as other social factors (McCormick et. al 1986; McKusick et. al 1990) can explain the frequency of the alleles. As another example, Fraikor (1977) proposed that the unique population expansions, contractions and cultural practices combine to make drift a parsimonious explanation.

The raised frequency of the alleles in Moroccan and Iraqi Jews (Karpati et. al 2003), French Canadians (De Braekeleer et. al 1992), the Pennsylvania Amish (Kelly et. al 1975; Myerowitz & Hogikyan 1986), as well as particular subsets of the Ashkenazi Jewish community (Broide et. al 1993) similarly suggests that there are common founder effects at play here, particularly indicating the absence of an IQ heterozygote advantage, given that the purported Ashenkazi Jewish IQ boost is absent in non-Ashenkazim (Cochran et. al 2006).

Neurocognitive Impacts

Despite the claims of Dunkel et. al (2019) that Kohn et. al (1988) provides support for the heterozygote advantage theory of Ashkenazi Jewish IQ, it actually provides more robust support against the theory. Specifically, the authors found that another spingolipidosis, metachromic leukodystrophy (discussed below), had lower spatial cognitive abilities, while Tay-Sachs patients were normal. The alleged common mechanism that Cochran et. al (2006) made such a big deal of was not found to have any impact on scores. The theory would actually predict that the two disorders would have similar impacts on IQ. Moreover, the authors found that the Tay-Sachs patients had scores that were below those of brain-damaged patients in several instances. Moreover, Tay-Sachs carriers have been shown to have higher rates of cognitive impairment, opposite the prediction by the theory (Zelnik et. al 2000).

Niemann-Pick type A

Neurocognitive Impacts

Model organisms with forms of the Niemann-Pick Type A disease have been shown to have motor, cognitive and emotional deficits (Dodge et. al 2005; Horinouchi et. al 1995; Iyer 2018; Ng & Griffin 2006; Trovo et. al 2015). In humans, SPMD1 mutations contribute to Parkinson’s disease risk, including carriers, indicating that the predicted positive benefits of Niemann-Pick Type A do not materialize (Dagan et. al 2015a; Dagan et. al 2015b; Dagan et. al 2016; Foo et. al 2013; Gan-Or et. al 2017). Moreover, there have long been documented deleterious impacts of NPD Type A on neurons and other brain pathologies (Ledesma et. al 2011). Indeed, more recent work in model organisms (monkeys) has demonstrated that acid sphingomyelinases help prevent neuronal cell death by “inhibiting lysosomal destabilization or lysosomal rupture” (Zhu et. al 2014). Moreover, it is well-known that NPD Type A patients suffer from rapid neurodegeneration (McGovern et. al 2017) and experience high burdens of neurological disorders (Cox et. al 2018). Despite this, SMPD1 is actually less involved in the brain than other sphingomyelin genes (Stoffel et. al 2018).

Non-Lysosomal Storage Disorders

Glycogen storage disease type Ia

Ashkenazi Jews have a frequency of Glycogen storage disease type Ia that is approximately 5 times that of the general European population (Ekstein et. al 2004). Despite having slightly higher carrier frequencies than East Asians, East Asians have a much higher residual risk for the disorder. Moreover, there is only one prevalent allele for the disorder, which has a quantitatively small overall frequency (Ekstein et. al 2004; Pavari et. al 1997), indicating a founder effect.

Torsion Dystonia

The only example of a “Jewish disease” that Cochran et. al (2006) provided any modicum of evidence for IQ-increasing effects is torsion dystonia. In a series of studies published by Roswell Eldridge in the 1960s and 70s, the group found that individuals with torsion dystonia had enhanced IQ (Eldridge et. al 1970).

Neurocognitive Impacts

There are several concerns to be had here. The first is that the result has not been replicated in any independent samples. For instance, Sofaer & Emery (1981) could not replicate the effect in a sample of extremely high IQ individuals.

Case reports from the literature have evinced multiple examples of torsion dystonia patients with mental disabilities and cognitive deficits (Nakashima et. al 1992), making one question the physiological impacts that purportedly lead to increased IQ. Other dystonias are known to be associated with negligible negative or substantial negative cognitive deficits (Balas et. al 2006; Coenen et. al 2017; Jahanashahi et. al 2003). In particular, torsion dystonia is typically said within the literature to have no extreme cognitive phenotypes (Kamm 2006).

With particular respect to Jewish populations, it has been shown that only the autosomal recessive form of the allele is associated with an increased IQ, while the dominant form actually is associated with reduced IQ (Eldridge et. al 1971). Recent research has shown that despite previous claims that Ashkenazi Jews had the recessive allele, they in fact have the dominant one (Risch et. al 1995; Zilber et. al 1984). The confusion of the Eldridge publications in combination with later results has been examined, showing that there is no consistent advantage, if any (Ferguson 2006). Indeed, in accordance with Ferguson’s claims, torsion dystonia genes have been shown to inhibit neuron formation (Carbon et. al 2008; Hewett et. al 2006).

Selection, Founders Effects, Drift or What?

There is some considerable evidence that the high frequency of idiopathic torsion dystonia in Ashkenazi Jewish populations is not due to heterozygote advantage, but due to drift (Risch et. al 1995).

Familial Hypercholesterolemia

Familial hypercholesterolemia, a disease known for being common in Ashkenazi Jewish populations (Seftel et. al 1989) as well as Sephardic Jews (Leitersdorf et. al 1993), has been shown to result from a mutation originating in the Lithuanian Jewish population (Meiner et. al 1991), as a result of the founders effect (Durst et. al 2001; Risch et. al 2003).

Breast / Ovarian Cancer

The best explanation for why Ashkenazi Jews have high frequencies of the BRCA1 and BRCA2 genes is not that they were positively selected, but due to drift because of bottlenecks (Im et. al 2011).

Congentinal Adrenal Hyperplasia

Despite initial claims that CAH exists at a higher frequency in Ashkenazi Jews than other populations (New et. al 2013), it seems that it may exist at quite similar frequencies to Caucasians (Hannah-Shmouni et. al 2017), and actually may exist at higher frequencies in Moroccan Jews (Rosler et. al 1992).

Neurocognitive Effects

Congentinal Adrenal Hyperplasia has long been posited to be the a source of masculinization in females, providing evidence for biologically determinist theories of the emergence of sex differences in psychological traits (Jordan-Young 2012). Despite this, its impacts on IQ in particular does have some meaningfully robust evidence on the question. In particular, it has been shown that children with CAH do not have IQs higher than their siblings, indicating no beneficial aspect of the condition (McGuire & Omenn 1975).

Bloom Syndrome

Drift, Selection or What?

Data on the frequency of heterozygotes with Bloom Syndrome gives strong indication that it can be explained by drift (Oddoux et. al 1999).

Physiology

Structural abnormalities in the spermatozoa of heterozygous Bloom syndrome patients have been observed (Martin et. al 1994). Some cases have reported “mild mental retardation [sic]” in patients with Bloom syndrome (Ahmad et. al 1977; Shah et. al 2012). Bloom Syndrome is essentially known as a cancer-prone disorder due to somatic instability (Yankiwski et. al 1999).

Fanconi anemia

Symptoms

Patients with fanconi anemia often have developmental delays (Byrd et. al 2011), and it has been posited that damaged mitochondrial function may be characteristic of the disease (Pagano et. al 2014), which is purportedly associated with intellectual functioning (Deary 2018).

Selection, Founders or What?

Evidence suggests, yet again, that the best explanation for the high frequency of fanconi anemia in Ashkenazi Jewish populations is genetic drift and founders effects (Tamary et. al 2008; Verlander et. al 1995). Populations like black South Africans (Morgan et. al 2005), Koreans (Park et. al 2015), the Dutch and Canadian Manitoban Mennonites (de Vries et. al 2012), and the Afrikaans population of South Africa (Rosendorff et. al 1987) suggest that there is either ubiquitous convergent adaption that does not impact population IQ, or that the disease prevalence is due to drift and/or founder effects.

BRCA1 and BRCA2

BRCA1 and BRCA2, genes that Ashkenazi Jewish populations have at quite high frequencies (Metcalfe et. al 2015; Oddoux et. al 1996; Struewing et. al 1995) were posited by Cochran et. al (2006) to be at this high frequency due to its IQ-increasing effects. Despite the claim, BRCA1 has been shown to have no impact on IQ (Bates et. al 2008). Moreover, a more parsimonious explanation for the occurence of the risk gene in Ashkenazi Jews is genetic drift (Fackenthal & Olopade 2007), given that it has also come to high frequencies in Spanish-derived populations (Mullineaux et. al 2003) and other Jewish populations (Bar-Sade et. al 1998). Research has also demonstrated that the 185delAGallele in BRCA1 in Ashkenazi Jewish populations comes from a common haplotype (Struewing et. al 1995), consistent with drift. Moreover, founder mutations have been identified for many families with BRCA1 risk alleles (Levy-Lahad et. al 1997).

Usher Syndrome

The origin of the R245X allele causing Usher syndrome in Ashkenazi Jewish populations are quite recent, indicating founders effects (Ben-Yosef et. al 2003), with similar results for the N48K allele (Adato et. al 2002).

Abetalipoproteinemia

Another “Jewish disease”, abetalipoproteinemia, is also at a high frequency due to founders effects (Benayoun et. al 2007).

Clustering and Other Idiosyncracies

One of the issues mentioned by a recent interlocutor and by Cochran et. al (2006) include the clustering of diseases around a common physiological mechanism: particularly the sphingolipid storage disorders (a particular subset of lysosomal storage disorders – see Platt 2014), including but not limited to Tay-Sachs, Gaucher, Niemann-Pick and mucolipidosis type IV. The aforementioned list is that diseases that occur with unusual frequency in Ashkenazi Jews. Beyond the listed ones, there are also GM2 and GM3 synthase deficiency, Fabry disease, Sandhoff disease, GM1 gangliosidosis, Krabbe disease, prosaposin deficiency, Salla disease, Wolman disease and metachromatic leukodystrophy (Gieselmann et. al 1995). Why is it that Ashkenazi Jews have a higher freuqency of the aforementioned four sphingolipid disorders, but not metachromatic leukodystrophy or Krabbe disease, despite their common biochemical mechanisms [8]? Is the Ashkenazi Jewish frequency of Niemann-Pick Type A really unusual given that East Asians have a higher frequency? Why do the French Canadians Gaspesie not experience an even more massive boost in IQ due to 1/13 of the population being carriers of Tay Sachs? 1/50 of the Romani population are carriers of GM1 gangliosidosis, meaning they should also experience genetically boosted IQ, but are commonly considered one of the populations with the lowest IQs. Where are the Italian and English superiorities in IQ due to Tay-Sachs disease (Branda et. al 2004)? Moreover, shouldn’t Ashkenazi Jews also have Salla disease? Why is it that the allegedly genetically stupider Sephardic Jews have high rates of Wolman disease, while Ashkenazi Jews do not?

Even beyond the questions of whether it’s unusual for Ashkenazi Jews to have high frequencies of a number of sphingolipidoses, we should point out that this is not unique to Ashkenazi Jews, but is also observed in Scandanavians, who have high frequencies of mannosidosis, aspartylglucosaminuria, Salla’s disease, and Gaucher’s disease type 3. High comorbidities for various other lysosomal storage disorders has also been observed in Arab populations (Zlotogora et. al 1988).

Finally, we should examine the statistical assumptions beyond the conclusion that clustering of biochemically similar disorders must be the result of selection, which seems quite implausible in the view of Ashkenazi Jewish population history (Risch 2001).

Niemann-Pick Type A

A single mutation (p.Arg498Leu) [9] for Niemann-Pick Type A has been found to account for about a third of Ashkenazi Jewish cases of Niemann-Pick Type A (Levran et. al 1991), with the specific mutation also causing Niemann-Pick Type B in the heterozygote form. Despite there being over 185 mutations associated with Niemann-Pick disease (Zampieri et. al 2016), just two mutations account for 60% (Jones et. al 2008), and accounting for only two other missense mutations (p.Leu304Pro & p.Phe333SerfsX52) accounts for over 90% of Ashkenazi Jewish cases of Niemann-Pick disease (Ferreira & Gahl 2017; Schuchman & Miranda 1997), which is why Ashkenazi Jewish mutation reference panels only include those 3 Niemann-Pick mutations.

Tay-Sachs

In Ashkenazi Jewish populations, just a small set of 3 alleles account for over 95% of the cases of Tay-Sachs (Bach et. al 2001; Grebner & Tomczak 1991; Landels et. al 1991; Mahuran et. al 1990; Myerowitz 1988; Myerowitz 1997; Myerowitz & Costigan 1988; Paw et. al 1990; Triggs-Raine et. al 1990; Triggs-Raine & Gravel 1990). However, in Moroccan Jews, who also have a substantial incidence of Tay-Sachs disease, there are 7 mutations in the HEXA gene that lead to Tay-Sachs (Kaufman et. al 1997; Navon & Proia 1991). Even French-Canadian populations, known to have high frequencies of the disease due to founders mutations (see Tay-Sachs section above) have over 7 mutations (Hechtman et. al 1990; Triggs-Raine et. al 1995), while Cajun populations in Louisiana have at least two (McDowell et. al 1992).

Indeed, when we consider the Acadians population, we also note that Acadians have the only known cases of Niemann-Pick Disease Type D (Greer et. al 1998; Winsor & Welch 1976), along with their high frequency of Tay-Sachs (Thurmon 1993). Similarly, the Amish have congenital adrenal hyperplasia, BRCA2, cystic fibrosis, nephropathic cystinosis, factor V deficiency, familial hypercholesterolemia, GM1-gangliosidosis type I, GM3 synthase deficiency, Famial Hypercholanemia, Mucolipidosis II alpha/beta, Tay-Sachs disease, while Old Order Mennonites have cystic fibrosis, factor V and XI deficiency, glycogen storage disorders II, IV and VI, mucolipidosis II alpha/beta, and  Salla disease.

Factor XI Deficiency

Three point mutations can account for factor XI deficiency in Ashkenazi Jews (Asakai et. al 1989).

Drift on Lethal Mutations?

One concern about the drift model of recessive diseases is that it may not be able to operate on lethal diseases, which have quite high negative fitness effects.

It has long been established that many humans carry lethal mutations in their genome, even if only they are only heterozygotes for the allele (Gao et. al 2015; Morton 1960; though see Amorim et. al 2017). As such, geneticists have developed models as to how they can persist in the population despite the alleged strong selection pressure against the allele (Slatkin et. al 1979). Zlotogora (1994) reports the possibilities for different classes of mutations and diseases

Many Diseases, Many Mutations

One of the objections raised in Cochran et. al (2006) is that the fact that multiple mutations have been found to contribute to Jewish incidence of the lysosomal storage disorders mentioned in the article (e.g. Tay-Sachs, Niemann-Pick, Gaucher, etc). This was actually an objection raised by Zlotogora & Bach (2003) in their response to Risch et. al (2003) (as noted in Ferguson 2008). The claim is essentially along these lines: if founder effects were to contribute to the high frequency of these disorders in Ashkenazi Jewish populations, then we would expect only one mutation to be at high frequency, not two. The issues with this are manyfold, but we should first visit the response of Risch and colleagues. Risch & Tang (2003) respond by saying:

Zlotogora and Bach (2003 [in this issue]) use the observation of multiple mutations to infer selective advantage for etiologically unrelated diseases, such as metachromatic leukodystrophy (MIM 250100), Hurler syndrome (MIM 252800), hyperoxaluria (MIM 259900), and ataxia telangiectasia (MIM 208900) in Arabs and Bardet-Biedl syndrome (MIM 209900) in Bedouins. An even stronger argument might be applied to BRCA1 in the Dutch population, which, according to a recent study, has at least 12 different recurrent mutations (Peelen et al. 1997). It seems implausible that all these disease mutations have undergone selective advantage unique to one population. We do not agree that the number of mutations found in a population is a good indication of past selective forces. Rather, it is aberrantly high mutation frequencies, despite severe negative selection against homozygotes, that provide the strongest argument for carrier advantage. We believe that care needs to be applied in concluding past selection when disease mutations are specific to single founder populations

Risch & Tang (2003)

(See also Risch 2001).

It would also be prudent to examine other examples of physiologically similar rare diseases at reasonably high frequencies in distinct populations that are likely to be explained by founders effects, mutation biases or other non-selective factors.

For instance, Heiniscih et. al (1995) reported 5 separate mutations causing a lysosomal storage disorder, metachromic leukodystrophy, in a small region:

The disease was diagnosed in Arab patients from two small geographic areas: in the Jerusalem region and in lower Galilee. Ten families with children suffering from metachromatic leukodystrophy were identified. Late-infantile metachromatic leukodystrophy was diagnosed in three unrelated Muslim families originating from a small village now included within the Jerusalem municipality. The seven other families with metachromatic leukodystrophy patients (four Muslim and three Christian) originated from seven different villages, which include a total of some 50,000 inhabitants within a small geographic area of -225 km2 in lower Galilee

Heiniscih et. al (1995)

They found that the explanations for each geographic area were likely to be distinct:

They indicate that the mutations in the Christian Arab patients have most likely occurred on the same haplotype, whereas those of the Muslim patients are found on different arylsulfatase A allele backgrounds.

Heiniscih et. al (1995)

Despite claims by many authors that if drift and/or founders effects are adequate explanations for the frequency of a recessive deleterious disease, then only a single mutation should be responsible for the presence of the disease in the population, it is likely that multiple founders events, admixture, and other population history idiosyncrasies are related to the presence of multiple alleles for a single disease in a population (Feingold 1998). This may resolve the “Newfoundland” and “La Reunion” paradoxes [10] (Katsanis et. al 2001; Zlotogora et. al 1996) over Bardet-Biedl syndrome [11].

And indeed, there is substantial evidence that multiple alleles for a single disease can arise in a population at a relatively high frequency due to founders effects. For instance, Papillon-Lefèvre syndrome has 4 associated mutations (out of the 25 known) in Saudi Arabian populations, which subsequent population genetic analysis has shown to be due to founders effects (Zhang et. al 2001). Similar results have been observed for Familial Mediterranean fever in Lebanese populations, where 5 separate mutations causing the disease emerge at high frequencies in a Lebanese family can be explained by founders effects (Medlej-Hashim et. al 2010). Genetic heterogeneity for rare diseases have also been observed for retinitis pigmentosa (Benayoun et. al 2009), recessive ataxia (Bouhlal et. al 2009), Werner syndrome caused by founders effects (Saha et. al 2013), congenital ichthyosis caused by founders effects (Fachal et. al 2012), and a plethora of mutations for nonsyndromic retininis pigmentosa (Sharon & Banin 2015), but interpretations of purported genetic heterogeneity can be complicated (Frishberg et. al 2007). [12]

There are similar concerns for the fact that Ashkenazi Jews show higher frequencies for over 20 identified disease genes, despite these being quite deleterious. However, analyses for a similar population, the Hutterites, has shown that founders effects with particular compositions of the original founding population provides a more parsimonious explanation than heterozygote selection for why the Hutterites have high frequencies of over 30 rare genetic disorders (Chong et. al 2012 – see also Gerull et. al 2013)

Mucolipidosis type IV

The vast majority of Ashkenazi Jewish cases of mucolipidosis can be accounted for by two or three mutations (Bach et. al 2005). [13]

Genes

Microcephalin

Cochran et. al (2006) cite the microcephalin gene and the purported selection that occurred on it as evidence for their position that intelligence was under divergent selection between Ashkenazi Jews and gentiles:

Recent work (Evans et al., 2004) supports this hypothesis. Evans et al. show that microcephalin, a gene controlling brain size has evolved rapidly throughout the primate lineage leading to humans and that this evolutionary process exhibits strong signs of positive selection (see also Wang and Bing, 2004). Microcephalin and BRCA1 both show signs of positive selection during primate evolution, share BRCT domains, and have critical functions in regulating the development of neural stem cells

Cochran et. al (2006)

[1] Despite Bray et. al (2010)‘s claim that their evidence (low heterozygosity) was inconsistent with a bottleneck, Carmi et. al (2014) demonstrated that the heterozygosity finding has been found alongside findings of strong bottlenecks, meaning there is a probably an idiosyncratic aspect of population history relevant here.

[2] For instance, Vaughn (1994) lists Jews involved in tailoring and woodworking, closely mirroring the crafts and trades Jews were involved in in the medieval period (Abrahams 1932, p. 222-224; Botticini & Eckstein 2013; Wischnitzer 1965). Again, not all Jews were involved in trade/commerce (Toch 2000; Toch 2012, p. 50-52).

[3] Cochran et. al 2006 claim that the Hughes study is important because it “contradicts a widely cited misrepresentation by Kamin (Kamin, 1974) of a paper by Henry Goddard (Goddard, 1917).” An unsanitized version of that history can be found in Gelb et. al (1986).

[4] Because epistasis often involves higher-order interactions, using less-related individuals will decrease the coefficients on epistasis components in all variance decomposition equations, making it difficult to distinguish between epistasis and measurement error.

[5] Cochran recognizes that “the Jews of Roussilon circa 1270” were moneylenders, but doesn’t consider that if Jews had become moneylenders by 1100 AD (Cochran et. al 2006, p. 12), then the cognitive selection benefits of the occupation should have been noticeable quite before 1400 (Cochran claims that the selection could occur rapidly, with 0.8 IQ points per generation). Moreover given the fact that Jewish occupational specialization actually happened far before that time period (circa 700-900 AD), then it should have been evident in the historical record centuries before Cochran claims the first records of Jewish intelligence emerge.

[6] Although I think there are many things to praise about Botticini & Eckstein’s work, I agree with most of the criticisms. See Chazan (2014) and Appelbaum & Appelbaum (2014), as well as Holo (2013).

[7] These occupations, in fact, are quite similar to the occupations that Jews partook in the Levant and under Muslim rule (Wasserstrom 1995, p. 21).

[8] For a more detailed exposition of these criticisms, see this blog post here.

[9] And before it be claimed that they may not exist in the populations to be positively selected, metachromic leukodystrophy is known to have an unusually high frequency among a very small Sephardic Jewish population (Zlotogora et. al 1980), but not the entire Jewish population as a whole, where its prevalence is regular.

[10] Note that even with the fact that the mutation accounts for a third of Ashkenazi Jewish Niemann-Pick type A cases, the mutation only has a frequency somewhere from 0.003 to 0.007.

[11] Note that the digenic model also provides a fairly robust explanation for many instances as well (Allamand et. al 1995; Beckmann 1996).

[12] See, e.g. M’hamdi et. al (2011).

[13] See also Vijzelaar et. al (2019).

[14] See also Schrijver et. al (2007).

Trans”ethnic” Polygenic Scores

Trans”ethnic” polygenic score research is research in human genetics that estimates the applicability of polygenic scores developed in one population to another population; whether or not, and how well, polygenic scores perform (Li & Keating 2014).

Recent discussion over the applicability of polygenic scores (hereafter PGS) has sparked some controversy on Twitter. Some have alleged that recent work on trans”ethnic” genetic correlations provide evidence that they have good external validity (e.g. Lam et. al 2019), while others caution the need to specifically tailor polygenic scores to different populations (Grinde et. al 2018; Gurdasani et. al 2019).

However, there are several issues to the question of the applicability of polygenic scores outside of the group they were designed in (Coop 2019). The first is whether effect sizes are biased; if the SNPs discovered in “European” cohorts (as the vast majority of GWAS research is done on “Europeans” [Bien et. al 2019; Bustamante et. al 2011Duncan et. al 2019; Fullerton et. al 2010Martin et. al 2017Martin et. al 2018Martin et. al 2019; Mills & Rahal 2019; Mogil et. al 2018; Need & Goldstein 2009; Park et. al 2018Peterson et. al 2019; Petrovski & Goldstein 2016]) are not true causal SNPs, but only tag SNPs captured because the tag SNPs are in linkage disequilibrium with the true SNPs, then the fact that these patterns of linkage disequilibrium are destroyed in other populations via recombination will lead to reduced predictive accuracy. Additionally, allele frequency differences can induce different percentages of variance explained per population (Lam et. al 2019; Zanetti & Weale 2018), meaning that scores that are predictive in one population are not nearly as predictive in another (Duncan et. al 2019).

Reduction in Predictive Power

Theoretical research has argued that the predictive power of a polygenic score will reduce about linearly with the $F_{st}$ value of the GWAS population and the application population (Scutari et. al 2016).

  • Bigdeli et. al (2017) – major depressive disorder – CONVERGE PGS explains 0.09% of variance in PGC, PGC PGS explains 0.2% of variance in CONVERGE.
  • Chang et. al (2011) – HDL – \Delta R^2 of 2.3% for “Europeans”, 3.4% for “African-Americans”, and 5.0% for “Mexican-Americans”
  • Chang et. al (2011) – LDL – \Delta R^2 of 9.9% for “Europeans”, 10.5% for “African-Americans”, and 5.3% for “Mexican-Americans”
  • Chang et. al (2011) – TC – \Delta R^2 of 5.0% for “Europeans”, 5.0% for “African-Americans”, and 3.7% for “Mexican-Americans”
  • Chang et. al (2011) – TG – \Delta R^2 of 3.6% for “Europeans”, 1.1% for “African-Americans”, and 7.2% for “Mexican-Americans”
  • Ikeda et. al (2017) – bipolar ‘disorder’ – 86.5% attenuation (from 2% in Japan-Japan to 0.27% in European-Japan)
  • Lee et. al (2018) – EA / IQ – 65.2% attenuation (\Delta R^2 from 4.6% to 1.6%)
  • Lam et. al (2019) – schizophrenia – 33% attenuation (3% to 2%)
  • Monda et. al (2013) – BMI – 22% attenuation (1.67% to 1.3%)
  • Nievergelt et. al (2019) – PTSD –  50$ attenuation (h^2_{SNP} from 4% to 2%)
  • Nylolt et. al (2012) – endometriosis – BBJ SNPs explained 0.54% of variance in QIMRHCS+OX cohort, QIMRHCS+OX cohort explained 1.06% of variance in BBJ cohort.
  • Rabinowitz et. al (2019) – EA – 71% attenuation (R^2 from 4.6% to 1.3%)
  • Vassos et. al (2017) – psychosis – 88.3% attenuation (from 9.4% to 1.1%)
  • Ware et. al (2017) – BMI – 72.7% attenuation (from 5.5% to 1.5%)
  • Ware et. al (2017) – height – 89.2% attenuation (from 7% to .75%)
  • Ware et. al (2017) – EA 2013 – ~91% attenuation (from 3% to .25%)
  • Ware et. al (2017) – EA 2016 – 81% attenuation (from 5.5% to 1%)

Common Causal Variants?

Theory suggests that for traits under local adaptation between regions, that the traits will not have equal effect sizes (Shi et. al 2019), though there are other reasons SNPs could have differential effect sizes by region.

There has been some suggestion that most GWAS variants are in fact common to all populations (e.g. ‘universal’), with approximately identical effect sizes (Akiyama et. al 2017; de Candia et. al 2013Guo et. al 2019; Gurdasani et. al 2019He et. al 2015; Jorgensen et. al 2017Lau et. al 2017Marigorta & Navarro 2013; Monda et. al 2013Nylolt et. al 2012Waters et. al 2010; Xing et. al 2014), though not all work shows similar results (Bigdeli et. al 2017Carlson et. al 2013; Chang et. al 2011; Diagram Consortium et. al 2014Fesinmeyer et. al 2013; Locke et. al 2015; Ware et. al 2017; Wray et. al 2017).

r_{g} and effect sizes?

  • Akiyama et. al (2017) found an r_{g} of .94 for BMI loci.
  • Bigdeli et. al (2017) found r_{g}s ranging from 0.33 to 0.41 for major depressive disorder (MDD) between Chinese and “European” populations.
  • Brown et. al (2016) used traits like gene expression, rheumatoid arthritis and type 2 diabetes and found r_{g}s of 0.32, 0.46, and 0.62 respectively for Yorubans and “Europeans”, and “Europeans” and “East Asians” respectively.
  • Carlson et. al (2013) found that up to 25% of SNPs tagged in a “European” cohort had significantly different effect sizes in a “non-European” cohort.
  • de Candia et. al (2013) found an r_{g} of 0.66 and 0.61 for schizophrenia in two datasets, but the trans”ethnic” genetic correlations depended on minor allele frequencies.
  • Fesinmeyer et. al (2013) found evidence for effect size heterogeneity in 5/13 SNPs.
  • Guo et. al (2019) found that the r_{g} of height for all SNPs between “Europeans” & “Africans” was .75, for BMI was .68, while genome-wide significant SNPs had slightly higher r_{g}s at .82 and .87 for height and BMI respectively. They found, however, that these could not be attributable to allele frequency differences or linkage disequilibrium, but could not rule out issues of power.
  • Ikeda et. al (2017) reported a \rho_{g} of 0.724 for bipolar disorder
  • Jorgensen et. al (2017)‘s meta-analysis of alcohol consumption SNPs found r_{g}s ranging from 0.4 to 0.6.
  • Marigorta & Navarro (2013) found that \rho_{g} for \log(OR) was 0.82 between “European” and “East Asian” populations, but effect sizes were slightly larger (\beta>1). The \rho_{g} differed by whether the variant replicated between populations. They also found that differences in linkage disequilibrium could likely explain some of the failed replications.
  • Ntzani et. al (2011) found averages r_{g}s of about .20 (“Asian”-“African”), .27 (“European”-“African”) and 0.33 (“European”-“Asian”) for asthma, atrial fibrillation, BMI, breast cancer, colorectal cancer, eosinophil count, gout, height, Parkinson’s disease, prostate cancer, schizophrenia, SLE, stroke, systemic sclerosis, type 2 diabetes, and uric acid.
  • Wray et. al (2017) found an r_{g} of 0.33 for major depressive disorder, of 0.34 for schizophrenia and 0.45 for bipolar disorder between “Chinese” and “European” populations.
  • Yang et. al (2013) found an r_{g} of 0.39 for ADHD between “European” and “Chinese” populations.
  • Zhou et. al (2018) finds r_{g} of about .4 to .6 for height and BMI

Directional Consistency and Replication Rates

  • Bigdeli et. al (2017) found that ~50.5-51.1% of SNPs are directionally consistent/replicate between the CONVERGE and PGC cohorts for major depressive disorder.
  • Carlson et. al (2013) found directional consistency rates ranging from 68-88% for BMI, type 2 diabetes, and lipid levels.
  • Chang et. al (2011) found a replication rate between ancestral groups for lipid variants that ranged from 44-67%
  • Diagram Consortium et. al (2014) found that the effect sizes between “ethnic” groups were concordant in ~50-57% of SNPs, with concordance rates increasing as the p-value decreased.
  • Fesinmeyer et. al (2013) found replication rates of 69% for within-“European” analyses, 61% for “European”-“East Asian” analyses, 46% for “European”-“African” analyses, 46% for “European”-“Hispanic” analyses, 62.5% for “European”-“Pacific Islander” analyses, and 55.5% for “European”-“American Indian” analyses.
  • Locke et. al (2015) found 79% directional consistency for “Africans” and 91% directional consistency for “East Asians” for BMI.
  • Marigorta & Navarro (2013) found replication rates of 45.8% between “East Asians” and “Europeans”, which increased to 76.5% following accounting for statistical power. For “Africans” and “Europeans”, the respective figures were 9.6% and 59.2%.
  • Monda et. al (2013) found that 88.8% of BMI loci were directionally concordant between “African” and “European” populations.
  • Waters et. al (2010) found that all 19 loci in their study were directionally consistent between a number of populations.

Comparing PGS

Some individuals have advocated that polygenic scores finally provide an opportunity to test whether group differences in particularly phenotypic traits are the result of genes, environments, or some combination thereof. They argue that we can compute the polygenic score for each population, get some sort of representative sample and compare the means, thus concluding that some proportion of the phenotypic gap can be explained by the genotypic gap. However, there is little reason to believe that this is a valid method of inference, and indeed displays the naivete of these activists. As an example, Kerminen et. al (2019) document that subtle population stratification in Finland induces spurious polygenic differences in propensities for various complex traits like BMI, height and coronary artery disease. Martin et. al (2017) showed that the use of polygenic scores for height in “African” populations predicts that “Africans” would be less than 5 feet tall [1], indicating that the actual levels/intercepts of phenotypes can be severely misestimated (Kim et. al 2018). The reasons for this vary by trait, study and cohort, but include Eurocentric biases in GWAS, meaning that variants at high frequency in Africa, but low-frequency in Europe are not captured (Durvasula & Lohmueller 2019), differences in population frequencies of derived and ancestral alleles (Kim et. al 2018), gene-gene and gene-environment interactions (Coop 2019).

Conclusion

The applicability of polygenic scores developed in one population is likely to be limited in other populations, but the magnitude of this limitation varies by the discovery and target populations, the trait and the methodology used to assess heterogeneity. What is clear, however, is that inferences based on mean PGS values for different populations are severely limited at this point in time.

Footnotes

[1] See also here and Berg et. al (2019).

Reviving Sampling Theories of Intelligence

In my last post, I examined developmental and network models of intelligence as alternatives to ‘g‘ theory as explanations for the positive manifold. However, there are other alternatives as well: the sampling model proposed by Thomson (Thomson 1916) has long been pointed out by critics of ‘g‘ theory as something to fill the void (Shalizi 2007). Essentially, Thomson’s model proposes that there are many different uncorrelated abilities, and the fact that a positive manifold is observed is not reflective of an underlying “general” ability or anything of the sort, but because tests share/tap into these particular neural/cognitive resources/abilities.

The usual arguments brought up against sampling theory come from Jensen (1998):

But there are other facts the overlapping elements theory cannot adequately explain. One such question is why a small number of certain kinds of nonverbal tests with minimal informational content, such as the Raven matrices, tend to have the highest g loadings, and why they correlate so highly with content-loaded tests such as vocabulary, which surely would seem to tap a largely different pool of neural elements. Another puzzle in terms of sampling theory is that tests such as forward and backward digit span memory, which must tap many common elements, are not as highly correlated as are, for instance, vocabulary and block designs, which would seem to have few elements in common. Of course, one could argue trivially in a circular fashion that a higher correlation means more elements in common, even though the theory can’t tell us why seemingly very different tests have many elements in common and seemingly similar tests have relatively few.

Let’s break these parts down one by one:

One such question is why a small number of certain kinds of nonverbal tests with minimal informational content, such as the Raven matrices, tend to have the highest g loadings

Despite the claim that “Ravens matrices … tend to have the highest g loadings”, recent research has shown to be false (Gignac 2015). It actually ends up to be one of Jensen’s many falsehoods.

and why they correlate so highly with content-loaded tests such as vocabulary, which surely would seem to tap a largely different pool of neural elements

This claim is also false. The correlation between Raven’s matrices and tests like vocabulary are only around 0.48, indicating about 23% of shared variance per Johnson & Bouchard (2011). The tests that correlated highest with Raven’s matrices are the tests with the most similar content, such as block design, paper folding and hidden patterns tests. However, the correlations are broadly similar for Raven’s, making it difficult to make specific claims. Regardless, the fact is that we can’t presume the neural elements drawn upon by particular tests are reflected in the actual content or form of the test.

Another puzzle in terms of sampling theory is that tests such as forward and backward digit span memory, which must tap many common elements, are not as highly correlated as are, for instance, vocabulary and block designs, which would seem to have few elements in common

Johnson and Bouchard report correlations of vocabulary and block design ranging from .39 to .43, indicating that they are quite small. As for the forward and backward digit span memory, this is actually quite an interesting area of research. Factor analytic methods suggest that the abilities are broadly similar (Colom et. al 2005; Engle et. al 1999), but experimental research suggests that the difference is that backwards recall employs the visuospatial resources, while forwards recall does not (Clair-Thompson & Allen 2013). Moreover, the small correlation observed (which I cannot confirm at this time) may be the result of the extremely restricted range of the digit span tests (Miller 1956).

Of course, one could argue trivially in a circular fashion that a higher correlation means more elements in common, even though the theory can’t tell us why seemingly very different tests have many elements in common and seemingly similar tests have relatively few.

The issue with Jensen’s critique here is that the explanation that sampling theorists would offer to respond to this is not the argument he proposes they would. Sampling theorists might argue that there isn’t a need for an explanation of any set of particular correlations between tests, or that the proper mode of investigation isn’t to employ subjective and ad hoc interpretations of the elemental similarity based on content similarity, but rather to employ factor analytic methods and/or neurocognitive research.

And how would sampling theory explain the finding that choice reaction time is more highly correlated with scores on a nonspeeded vocabulary test than with scores on a test of clerical checking speed

It is unclear where Jensen is getting his sources from, but the choice reaction time research is quite supportive of a sampling model (Neubauer & Knorr 1997). However, we should be cautious in interpreting the correlations of choice RT tasks to other IQ subtests, as choice RT tasks can vary significantly (Liewald 2013). Given other issues like age differences (Der et. al 2017), and task complexity (Proctor & Schneider 2018), we should be cautious in interpreting the alleged finding.

When people actually do end up testing the actual predictions of the sampling theory, they end up being empirically confirmed (Rabaglia 2012). Modern versions have reworked the mathematical and empirical foundations of the theory to demonstrate that the alleged evidence mounted against it in the past are not so much disconfirmations as non-sequiturs (Bartholomew et. al 2009). Other research has extended the sampling theory into the neurocognitive domain (Kovacs & Conway 2016), while Detterman’s older theories (Detterman 1987, 2000) has been suggested to be similar to sampling and the new process overlap theory (Detterman et. al 2016).

Conclusion

Whether or not you’re an advocate of ‘g‘ theory, dynamic mutualism, or reject ‘g‘ altogether, the sampling model is a model to contend with. It does seem to have some empirical support, particularly with the widespread notion of different cognitive processes in cognitive psychology. While it may have some things to work out, it certainly has not been “refuted”, much less to the extent that Spearman’s g theory has.