Jonathan Haidt still thinks like a liberal

The New Yorker has a piece which follows up on how extremely biased the field of social psychology is, Is Social Psychology Biased Against Republicans?:

…A 2012 survey of social psychologists throughout the country found a fourteen-to-one ratio of Democrats to Republicans. But where were the hard numbers that pointed to bias, be it in the selection of professionals or the publication process, skeptics asked? Anecdotal evidence, the Harvard psychologist Daniel Gilbert pointed out, proved nothing. Maybe it was the case that liberals simply wanted to become professors more often than conservatives. “Liberals may be more interested in new ideas, more willing to work for peanuts, or just more intelligent,” he wrote. The N.Y.U. political psychologist John Jost made the point even more strongly, calling Haidt’s remarks “armchair demography.” Jost wrote, “Haidt fails to grapple meaningfully with the question of why nearly all of the best minds in science find liberal ideas to be closer to the mark with respect to evolution, human nature, mental health, close relationships, intergroup relations, ethics, social justice, conflict resolution, environmental sustainability, and so on.”

…“There’s often a lot of irony in this area,” he said. “The same people who are exquisitely sensitive to discrimination in other areas are often violently antagonistic when it comes to political ideology, bringing up clichéd arguments that they wouldn’t accept in other domains: ‘They aren’t smart enough.’ ‘They don’t want to be in the field.’ ”

…As the degree of conservatism rose, so, too, did the hostility that people experienced. Conservatives really were significantly more afraid to speak out. Meanwhile, the liberals thought that tolerance was high for everyone. The more liberal they were, the less they thought discrimination of any sort would take place.

Call it “liberal privilege” in academia. The assumption is often that the normative framework you are working with is on the cultural Left (these discussions are much more relevant for social issues since the high socioeconomic status backgrounds of many academics means they are less interested in Left populism in economic domains, even if they give lip service to it). Two examples will suffice personally. I was at a dinner hosting a speaker once when someone started talking about how “we” could reach out to conservatives. The assumption was that of the two dozen people around the long table none would be conservative. Second, I had an exchange with a patronizing reader who advised that I tone down any mention of politics on my blog since that would alientate readers. I responded that it seemed to work for PZ Myers, to which he responded “Oh, I hadn’t thought of that.” The issue here is that the reader didn’t think of liberal politics as politics.

Second, the idea that liberalism just aligns with reality is a nice and neat conceit, but it is, as a nice liberal would say, “problematic.” Chris Mooney has a short piece on the priors of sociologists and human nature. “Blank slate” dead-enders aren’t just found on the political Right, though some of that exists there too (e.g., homosexuality, the important of shared family environment). Additionally, the natural sciences tend to be less politically liberal overall than fields like sociology and social psychology. I’m skeptical that this suggests that the best minds are in sociology and social psychology.

Finally, I do have to note that Haidt still offers a fundamentally liberal solution to the conservative deficit: “Haidt believes that the way forward is through a system of affirmative action: engaging in extra recruitment and consciousness-raising efforts for political conservatives in much the same way as for women or ethnic minorities.” All the same issues that afflict racial and ethnic affirmative action might be relevant in the case of conservatives. It may actually be the case that conservatives don’t have the inclination and aptitude for particular fields. This doesn’t negate the reality of discrimination, but it questions axiom that discrimination in a crude sense is shaping the different demographics we see in different fields, and a simple fiat fix can solve the problem. It is true that social psychology probably suffers from the lack of ideological diversity, but it won’t prevent the field from publishing lots of fluffy results which are picked up by the media oblivious to p-value fishing. Yes, people won’t take social psychology seriously, but most scientists probably never have. Bias is real. And discrimination is real too. But sometimes the solutions are more pernicious than the problems. That’s a conservative insight.

Men of Bronze, Men of Iron

Ancient DNA has made it big time, with a write up in The New York Times by Carl Zimmer, From Ancient DNA, a Clearer Picture of Europeans Today. The primary sources of the piece are two papers published recently in Europe, Ancient human genomes suggest three ancestral populations for present-day Europeans (ungated version) and Genome flux and stasis in a five millennium transect of European prehistory. Like the earlier mtDNA work on ancient DNA what these results are suggesting is that European demographic history has been punctuated by periodic discontinuities. Here’s an important bit from the Zimmer article:

Archaeologists have found that early farming culture didn’t change drastically for the next 3,700 years. But about 4,000 years ago, the Bronze Age arrived. People started using bronze tools, trading over longer networks and moving into fortified towns.

Dr. Pinhasi and his colleagues found that the era also brought a sudden shift in human DNA. A new population arrived on the Great Hungarian Plain, and Dr. Reich believes he knows who they were: the northern Eurasians.

It seems really unlikely that Europeans are special in this tendency, with broad world-wide trends as outlined in Towards a new history and geography of human genes informed by ancient DNA. For decades there have been books written about the coming for the Indo-Europeans. David Anthony’s is probably the best recent summation of the “they came out of the steppes” viewpoint. If there is a major update on this it now looks like the demographic impact of the Indo-Europeans was much greater than we had previous imagined. But are Indo-Europeans special? Probably not…earlier work suggested major discontinuities in Europe with the arrival of agriculture. Later in Africa you had the Bantu expansion, which replaced most of the local people. And as John Hawks points out the ancient Siberian who lived ~45,000 years ago probably comes from a group with no modern descendants. With the disappearance of the Ma’lta boy’s people ~20,000 years later from eastern Siberia it suggests that the heart of Eurasia has been roiled multiple times since the arrival of anatomically modern humans.

Addendum: I would take minor issue with the title of The New York Times piece. The picture isn’t really clearer, but cloudier. It’s just that the old clear picture was wrong, and the new cloudy picture is less wrong. Ultimately the clouds may clear, but we need more samples for that.

Bay Area Population Genomics Meeting XI, December 6th


In case you are one of the few people interested in population genomics who is not religiously following Twitter, the 11th Bay Area Population Genomics Meeting will be held at UC Davis on December 6th. Save the date. It is free. But you have to register. Details the Coop lab website (they are hosting). I’ll be tweeting from it.

R1a1a, what is best in life?

300px-R1a1a_distributionOver at Greg Cochran’s blog he’s been posting on Indo-Europeans. He’s had many of these ideas for a long time, but after I recounted to him some more information from ASHG 2014 it crystallized a lot in terms of specifical detail. For example, the Kalash of Pakistan share a lot of drift with “Ancestral North Eurasians” (ANE). By “a lot”, I mean in the same range as North Caucasus and Eastern European groups. Other HGDP samples from Pakistan are somewhat lower in their signals, but it still noticeable.* In Iosif Lazaridis’ presentation at ASHG 2014 he outlined the likelihood that the widespread distribution of ANE ancestry in Europe probably had something to do with the migrations of the Yamna culture, from which derived the Battle Axe Culture. The genetic variation you see in eastern and central Europe today is representative of the Yamna people. They know because they have ancient samples from those regions. The Yamna themselves are a mix of an Armenian-like Middle Eastern population, and “Eastern Hunter-Gatherers” (EHG) which resemble those to the west but have a higher fraction of ANE (so the are WHG + ANE, while the Armenian-like population is similar to, but not exactly the same as, the “European First Farmers” (EFF).

But that’s not the point of this post. There were two Y chromosome posters which were of interest. One showed a Bayesian skyline plot which illustrated that many of the Y chromosomal lineages you know and love went through very rapid population expansion on the order of 5 to 10 thousand years ago. A second poster had a phylogeny of Y chromosomes derived from high coverage whole genome sequencing. They had four individuals from the R1 lineages, two of them from R1a1a. One individual was Indian and the other was Russian. The coalescence was ~5,000 years ago. The individual who did this analysis was not aware of the Bayesian skyline plot poster, so she immediately ran off to look at it when I told her. The coalescence with R1b for the R1a individuals was ~10,000 years ago.

I know that there are lots of debates about clocks and calibration when it comes to Y chromosomes. But the archaeology, ancient DNA, autosomal work, and uniparental lineages are all coming together with a coherent picture. The Y chromosomal data strongly suggests that we’re talking about “star phylogenies” in the recent human male past.

* And for what it’s worth the Kalash are not descended from the soldiers of Alexander. Rather, they seem an early example of the admixture which led to modern South Asians. Their drift from other populations is due to them being isolated and endogamous.

Gattaca will never be

Yesterday CBS had a segment on hyperbolically titled Breeding out Disease. First, we will never “breed out” disease. Part of the reason is that a large fraction of disease is due to non-genetic factors. Perhaps in the future with nanotech we might get at all the biological misfires due to developmental problems which emerge out of the “environmental” (a word for stuff we basically can’t understand in any causal sense) effects. But genes aren’t everything.

Second, the CBS piece had two segments, which differ a lot in terms of their implications. The first involves preimplantation genetic diagnosis (PGD). This is already happening, so what you will see in the future is a matter of scale or magnitude, not a paradigm shift. I do think it is possible that in the next generation we will see the diminishing of recessive diseases due to highly penetrant deleterious alleles. Every birth of a child who is diagnosed with such a disease will allow us to predict future births, because presumably their parents will have rare variants which can then be put in the database. I don’t think this is controversial or scary in any way. It’s classic “science makes the world better.” Your child having a recessive disease or a karyotype abnormality is not part of some grand plan.

But the next element of the segment dealt with the firm GenePeeks. I saw the founder speak at the Consumer Genetics Conference in 2013, and it seemed to be a reasonable idea. Basically right now the play is to simulate the outcome of genotypes for combinations of sperm (donors) and eggs (the founder herself has a child with a recessive disease due to herself and her sperm donor being carriers for a rare disease). Enter Lee Silver, a famous geneticist before genetics was even quite so big. He makes many claims, some entirely reasonable, and some which I view to be a stretch. It seems that in concert with PGD simulating genotypes and looking to avoid highly penetrant alleles is very smart. In fact this is just carrier screening on steroids. But then Silver begins to imply that genetic methods are going allow to predict complex traits. On the face of it this seems likely to be true. The work on height is just a trial run for all sorts of complex traits, in particular diseases. In the next 10 years it is entirely likely that genomic techniques will allow us to capture most of the heritable variation which we now classify as “missing heritability”. Making a prediction which is actionable is a different thing altogether.

If you have a trait whose genetics is distributed across thousands of loci then simulating the gentoypes is going to be a brute force affair. I trust computation to catch up to this problem, but then it is making predictions on the individual level. It is one thing to capture the heritable variation on the population scale, but predicting in an individual case is going to be harder. Then, once you have the prediction you have to screen an enormous number of genetic combinations. If you want more than one complex trait, and they are independent, then the problem becomes exponentially more difficult.

There are two things which I think can get around this. One, which I’ve already mentioned, is to skew the embryos which are enriched for a grandparent whose quantitative trait you value (intelligence, height, or agreeability). Second, as I have said, the 2010s are the decade of reading the genome. The 2020s are going to be the decade of writing the genome. That seems a more viable and probable solution than screening for variants which are “in house.”

Finally, there is the standard question about selecting for non-disease traits like eye color. Silver doesn’t blink, and admits that this might happen. Norah O’Donnell is unsurprisingly concerned. I would reassure that we already select for non-disease traits in our children by selecting our spouses. It’s not that big of a deal. I’m rather sure that O’Donnell’s husband didn’t marry just because she’s a great journalist.

In the end Gattaca is a great movie with contemporary relevance. And thanks to the statistical shenanigans which went on in fMRI research it seems that genetics is unchallenged today as the queen of the biological sciences in our age.* But a movie is not reality, and geneticists have not bit into the apple of knowledge and are not as the gods. Relax, though expect a better future.

* Neuroscience made a play, but I think that’s done.

Heritability, autism, & fear of breeding

Cite: Sandin, Sven, et al. “The familial risk of autism.” JAMA 311.17 (2014): 1770-1777.

My wife and I are hoping to have more children. I am pretty convinced of paternal age effect,* so aside from issues like Down Syndrome, which is due to maternal age, the risk for various behavioral issues such as autism are on the radar for us. Also, there are people in the extended pedigree who are probably on the Asperger spectrum (I probably would characterize myself as on the spectrum, but many people who have met me contend that this just isn’t a good description and so might confuse or mislead readers who only know me from the blog. It might be more accurate to say that I’m low on “agreeability”). But should I be worried?

This weekend I heard about a documentary, Mimi and Donna, which is about a 90 year old woman who has to confront letting her 60 year old daughter be institutionalized. Because of her age the woman has never been properly diagnosed, but she definitely has some sort of “intellectual disability.” Today she might have been diagnosed with autism. The film-maker is the grand-daughter of Mimi, and the niece of Donna. What she said was very interesting to me. Her brother has a son with autism, and she herself has a son with autism. As she admits, it runs in the family. Though as I stated above some of my family members exhibit behavior which might seem on the Aspergers end  of the spectrum (and I do have a lot of physical scientists and engineers in my pedigree), none of them have been diagnosed autism, nor are they mentally retarded in any way.

As of now my daughter is nearly 3 years old and displays no sign of autism, and my son is young yet but makes eye contact and is socially typical (yes, I am aware of later onset). So I began to realize perhaps I should at least update my odds a bit. I may contribute de novo mutations for risk, but it seems that I don’t carry them from previous generations. I found a large Swedish study, The Familial Risk of Autism, which outlines clearly the odds conditional on affected or unaffected siblings. The figure above is from that paper, and here are the results:

Results In the sample, 14 516 children were diagnosed with ASD [Autism spectrum disorder], of whom 5689 had autistic disorder. The RRR [relative recurrence risk] and rate per 100 000 person-years for ASD among monozygotic twins was estimated to be 153.0 (95% CI, 56.7-412.8; rate, 6274 for exposed vs 27 for unexposed ); for dizygotic twins, 8.2 (95% CI, 3.7-18.1; rate, 805 for exposed vs 55 for unexposed); for full siblings, 10.3 (95% CI, 9.4-11.3; rate, 829 for exposed vs 49 for unexposed); for maternal half siblings, 3.3 (95% CI, 2.6-4.2; rate, 492 for exposed vs 94 for unexposed); for paternal half siblings, 2.9 (95% CI, 2.2-3.7; rate, 371 for exposed vs 85 for unexposed); and for cousins, 2.0 (95% CI, 1.8-2.2; rate, 155 for exposed vs 49 for unexposed). The RRR pattern was similar for autistic disorder but of slightly higher magnitude.We found support for a disease etiology including only additive genetic and nonshared environmental effects. The ASD heritability was estimated to be 0.50 (95% CI, 0.45-0.56) and the autistic disorder heritability was estimated to 0.54 (95% CI, 0.44-0.64).

I am aware that the basal risk is low. But the cumulative sum of a lot of independent low risks can be quite stressful.

* A presentation at ASHG 2014 which had huge sample sizes (~1000 trios) with whole genome sequencing convinced me that paternal age effect shouldn’t be doubted. It’s real.

The Austronesian explosion was crazier than we thought

The Austronesian languages, credit
The Austronesian languages, credit

The Austronesians were crazy and extraordinary. Starting about ~5,000 years ago they set off from the environs of Taiwan, and began to push outward. For ~30,000 years the people of Melanesia had defined the eastern edge of human habitation, but the Polynesian branch of the Austronesians blasted past that, going alway the way to Hawaii and Easter Island. At the other extreme the ancestors of the Malagasy settled Madagascar, and island which the peoples of Africa had not reached as of yet despite ~200,000 years of human habitation. We don’t know what was happening here, but it is hard to pinpoint particular cultural, environmental, or genetic forces which might result in these sorts of radical change in mores. Humans are conservative and cautious by nature. But our particular lineage of modern humans far less so than our forebears or cousins. After all we did make it Oceania and the Americas, while the others did not.

But a great unresolved question is contact with the Americas. There’s a lot of suggestive evidence, but no clincher. But two recent papers increase the probability considerably. Both are in Current Biology, Two ancient human genomes reveal Polynesian ancestry among the indigenous Botocudos of Brazil and Genome-wide Ancestry Patterns in Rapanui Suggest Pre-European Admixture with Native Americans. Alexander Kim has already reviewed the nuts & bolts of the first paper. Here’s the major finding: heretofore the reasonable assumption about these Polynesian remains in interior Brazil were the product of escaped slaves, but there is an 80-90% probability that they died before any such enslavement of Polynesians could have occurred. In fact both remains may be pre-Columbian!

Cite: Genome-wide Ancestry Patterns in Rapanui Suggest Pre-European Admixture with Native Americans
Cite: Genome-wide Ancestry Patterns in Rapanui Suggest Pre-European Admixture with Native Americans

The second paper has a somewhat more subtle result. The inhabitants of modern day Easter Island are descended in the main from the Polynesians who arrived from the west. This has long been known from classical genetics and non-genetic fields. There has also been suggestion of European and Amerindian admixture. Entirely reasonable in light of Easter Island being a possession of Chile, and 19th century migratory events. What these authors did is that by looking at the distribution of ancestry outcomes in the genomes of Easter Islanders, they inferred that the admixture with Amerindians far predated that with Europeans. The rationale here is simple: recent ancestry from divergent groups tends to exhibit patterns of long alternating blocks, due to a relatively small number of recombination events. In contrast older ancestry tends to be broken up by many recombination events over the generations, until deconvolution can’t separate the two elements and they fuse as one. As an example of the latter case modern day Europeans and South Asians are compound populations whose admixture dates of ~4,000 years or more makes it difficult to trivially deconvolute their ancestral components on a genome-wide scale (though ancient DNA from Mal’ta likely can help in the case of Europeans).

Figure 4 above shows the match of two demographic models with the empirical results. M2 is one where Mestizos from Chile bring European and Amerindian ancestry into the genomes of Easter Islanders. M1 is where there is an ancient Amerindian admixture, followed by a later European one. The solid lines show the predictions, while the points show the empirical results from the samples. It is clear visually that M1 fits the data. There are many short Amerindian blocks, evident of an old admixture, as opposed to more varied and longer European blocks. The rough dates for Amerindian ancestry admixture are in the range of 1300 to 1400 A.D., which match reasonably well with when Easter Island was settled.

These results are strong. Not definitive and probably not the last word, though more Easter Islander samples can end the debate of admixture at least. But they make us wonder how incredible human migrations have been over the past ~50,000 years! Ancient people were far more daring than we had imagined, and I think we need to reconsider what “crazy” exactly is in many ways.

Open Thread, 10/26/2014

What’s up?

So I recently listened to an interview with the author of . One datum which is of interest is that apparently we as human beings are conditioned to ambient sound, and environments which are truly silent (chambers designed for scientific experiments) are very uncomfortable for us, in large part because our own sounds begin to overwhelm us (e.g., the beating of the heart). Anyway, the issues at the heart of the book turn out to be very relevant to me. Normally I have an on my person. I have had one since early 2008. But at ASHG 2014 I forwent it in the interests of being able to hear someone if they wanted to get my attention. This caused an unanticipated problem. It turns out I really hate the sounds that you encounter in a public restroom, and I’ve been habituated toward just turning the shuffle on whenever I feel like it. This gets to the point that we all create our own aural environment, and in the age of portable digital devices this is been taken to a new level.

The forgotten first ages of mankind

iceage01By now you have read the paper in Nature, Genome sequence of a 45,000-year-old modern human from western Siberia. In The New York Times Carl Zimmer has an excellent write up, Man’s Genome From 45,000 Years Ago Is Reconstructed. The two major findings that are getting a lot of attention are that the Neandertal ancestry tracts in his genome are considerably longer than in modern humans and that he is basal to modern non-African populations. In regards to the first the distribution of Neandertal ancestry in the genome allowed them to infer backward to the point at which a pulse admixture might have occurred. Seeing as this individual has been dated to ~45,000 years before the present, Neandertal admixture occured 50-60,000 years ago. This happens to right around the time of the “Out of Africa” expansion.

But for me the Neandertal aspect is not the most interesting, as that simply refined our prior understanding. Rather, it is the relationship to modern human beings. The same first author gave us DNA analysis of an early modern human from Tianyuan Cave, China a few years ago, and in it she reported that a 39,000 year old individual found in China already exhibited clear affinities to eastern Eurasians and Oceanians. Now, one of the inferences which fit the results in this paper is that this 45,000 year old Siberian derives from a period when west and east Eurasians were not fully diverged (or, that divergence had been a recent event). Is ~6,000 years sufficient to account for drift away of the Tainyuan sample? My intuition is that it isn’t (they had ~85,000 positions at chromosome 21 to make the inference about the ancient Chinese sample, more than sufficient). I suspect that the dating is off somewhere here, though I don’t know which (if only one) sample.

Second, the issue of Basal Eurasians comes up again in the paper, and more extensively in in the supplements. I have a very hard time not believing that there is a paper on Basal Eurasians in the works, because they are very sketchy on specifics (also, Lazaridis did not really talk about the Basal Eurasians in the ASHG presentation). From page 60 of the supplements:

We caution that the TreeMix model is sensitive to which present-day populations are used for the analysis and the tree changes when we use different present-day human populations (Tabl S10.1). In particular, the analysis is sensitive to which African individuals are included. When we including only one African individual and all available Eurasian individuals, the Ust’-Ishim individual separates before the Eurasian split (position a) with bootstrap support between 68% and 100% depending on which population the individual is from (Table S10.1). These results suggest that recent gene flow between African and from European populations may influence the placement of Ust’-Ishim in the maximum likelihood tree. Nevertheless, the TreeMix models typically find that the Ust’-Ishim individual separates either before Eurasian split (position a in Figure S10.7) or just after the split and already on the eastern non-African lineage (position c in Figure S10.7).

The Sardinians have a bit of Sub-Saharan ancestry from the Roman era, but is this what they’re talking about? The Basal Eurasians are part of the greater Eurasian clade. That is, they are a branch of “Out of Africa.” But who knows where they are localized? No one knows. Thinking about it it probably is the case that they’re talking about recent African gene flow into Europe, even if it’s a few percent in Sardinia. But who knows? It’s all quite mysterious.

We’re at a very confused and exciting time right now. About a decade ago there was a stylized model of a rapid “Out of Africa replacement” all across the world. Mitochondrial Eve had even convinced people that Africans were subject to this. That is not the case. It seems clear that the Khoisan people diverged from the rest of the human lineages on the order of 200,000 years before the present. This predates any “Out of Africa” event. Much of Africa’s genetics has been reshaped by massive demographic expansion by the Bantus. There have long been hints out of the Reich lab about “Out of Africa” gene flow back into Africa that isn’t obvious (i.e., not Ethiopia or North Africa) leading to the genesis of contemporary Sub-Saharan populations, but excluding the hunter-gatherers. Such hints don’t emerge from a vacuum. The results are perplexing. It strikes me that now we know a fair amount about the demographic events which reshaped the Holocene, but the period before the Last Glacial Maximum is now far more clouded. Attempting to reconstruct the deep past with the algebraic variables of the recent past might be part of our problem here….

Update: From the comments:

It’s very likely Tianyuan was not really closer to East Asians than Ust-Ishim. U-I branches with East Eurasians with 100% bootstrap support, but branches off before Oceanians and Asians (supplements, page 57-58). This is similar to how Tianyuan branched, which was the main basis for that paper claiming it’s proto-East Eurasian. I think they did not realize during the Tianyuan paper’s writing that this could be simply because Tianyuan lacked the Middle Eastern/Basal/African admix of Europeans used in the tree (French and Sardinians).

Formal testing (f-statistics from Rasmussen et al 2014) which were not done in Tianyuan’s own paper show no difference between Tianyuan’s relation to East Asians and Europeans.
(F & G)

This sounds reasonable to me. On the other hand this possibility is not acknowledged in the supplements to the paper.

California Olive Ranch, the real deal

About a month ago I was listening to the host of America’s Test Kitchen, Christopher Kimball, talk about olive oil on the radio. I’m aware that most of the “extra virgin olive oil” we buy in this country is basically fradulent. A study came out of UC Davis on this several years ago, but it’s long been an open secret. It doesn’t really matter much if you use olive oil to cook, as a lot of the flavor disappears in that process anyhow. Rather, it matters if you use olive oil as dressing on salads and such. I do. Still, as I have a rather unrefined palette I’ve usually plunked down for the biggest container I could find. Yes, even . But Kimball was asked by the host about his preferred brand, and he offered , which also came out high on the UC Davis study, though to be fair the firm apparently had a hand in the funding. This extra virgin olive oil is often described as having a strong aromatic taste, almost pungent. I don’t mind strong tastes, so I ordered some on .

Probably because my taste sensitivity has been modulated by my high consumption of spices I felt that that the tangy aspect some mention was rather subtle. The oil is very “clean,” and leaves less of a cloying taste in my mouth than what I’m used to. Though I’m conscious of the fact that this might be a subjective perception of the fact that I know that someone with refined tastes prefers this olive oil, I’ve decided to make the switch and made a large order. I’d be curious what olive oil readers prefer.