Monday, 15 September 2014

Gene hunters and gene saboteurs


There is quiet satisfaction in some quarters that the search for the “intelligence genes” is not making much progress. I, on the contrary, am watching the publication of papers with excitement. We know that intelligence is heritable, because consanguinity is related to a similarity in intellect. We don’t know the names of the genes that bring about this effect, nor how they interact to do it. Given that there may be many hundreds of genes building the brain, each contributing a very small effect, finding them may be hard. For once, researchers are getting together the very large samples required, which may need 1,000,000 genomes. We are reaching towards the biological/social science version of CERN.

But, rather like all other activities, for every enthusiast there is an equal and opposite activist, rather like fox hunters and hunt saboteurs. This may need a little explanation for those readers who have the misfortune not to live in England. Fox hunting is a 16th Century country pursuit, commemorated in paintings of red coated, top hatted gentlemen on fine mounts jumping over hedges in pursuit of the red fox. Its association with the gentry makes it a status marker, implying that fox hunters are Conservatives, Protestants, fond of flagellation and possibly of wearing double breasted pyjamas in bed. Until I started living part of my life in the English countryside, I did not know that for every rider there is at least one countryman hunt follower in a small van. They are farm workers, electricians, plumbers, builders and shopkeepers. I join them by chance on country roads sometimes, and stand with them briefly as with their binoculars they watch the hunt work across the landscape and explain to me, in detail, what is going on. Tradition maintained.

On the other side are a largely urban species, the hunt saboteurs, associated with rabid socialism, multiply-occupied scruffy bed-sitters, vegetarianism and sleeping naked several to a bed. Some have the secret vice of having rich parents. They disrupt hunt meetings, lay false scents, and get into rows, which occasionally lead to minor violence. England being England, it has also led to romantic liaisons with the fox hunting toffs, and at least one marriage. Bless.

(For those still confused, watch Downton Abbey, and remove the small delivery vans).

Back to genetics. The gene saboteurs are defending the citadel of environmentalism, and probably regard each large scale genetic study of intelligence as a waste of resources. The gene hunters are, according to different views, either upholding the highest traditions of the Enlightenment, or dragging us into social Darwinism, nature red in tooth and claw. All these conflicting views contend in scholarly publications. Who really doubts that this is a great time to be alive?

The gene hunting enthusiasts are making progress, in my view. They are raising their standards, roping in even more researchers across the world, and using statistical approaches which are likely to minimise false positives. The gene saboteurs may be laying false scents, arguing that slow progress in phase 2 (gene identifying) means that doubt can be cast on phase 1 results (heritability estimates), but against their will they are driving up the standard of proof, though I very much doubt they want research on the genetics of intelligence funded, and certainly don’t want the databases opened up for racial comparisons, an obvious additional pathway to cracking the code.

Nature (the magazine) described the state of play as: Smart genes prove elusive: Study of more than 100,000 people finds three genetic variants for IQ — but their effects are maddeningly small.

The article says that 106,000 genomes have yielded 69 genes of interest for their links with scholastic achievement, of which 3 may be linked to intelligence. Every article should have a critical comment, in addition to the author of the study saying something, and in the sceptical box Nature says: “With effects this small, the chances that they represent false positives are vastly increased,” says Kevin Mitchell, a neurogeneticist at Trinity College Dublin who says he was decidedly underwhelmed by the study. “While intelligence — and proxy measures such as cognitive test performance or educational attainment — are quite heritable, the idea that this trait is determined by common variants in the population at large is really unproven,” he says.

Interesting that the response is “underwhelmed”. Very few psychological studies muster 106,000 subjects, and if you ensure that your statistics are rigorous then a small effect size need not indicate that the findings are false positives: they may be entirely true but weak positives.

To my mind the history of progress in science tends to oscillate between practice and theory. In my cynical moods I think that theory takes the back seat much of the time, because observation and plain data crunching can take you a long way. At the very least, you can work out the shape of the terrain, and often find flakes of gold. However, there comes a time when just working through the gravel pit looking at individual chips of stone has to give way to an engaging, broad impact but testable hypothesis. Someone is going to have to come up with a better theory as to how the genome builds the proteins that build the organism. The code breakers at Bletchley Park used cribs much of the time: they knew that enemy weather stations would have to report the weather, submarines the positions of shipping, and the very most important orders might include the sequence of letters “Hitler”. (I have encoded this sequence on an Enigma machine myself, wearing white gloves, not out of deference to afore-named assassin, but as a courtesy to the curator of the Bletchley Park museum, to protect the machine I was using). There must be genetic cribs of some sort, so that instead of looking at the limited 4 letter alphabet of the genetic code researchers begin to recognise the odd word, or even short sentence.

If a million genomes are required, then that should be the next step, and those who are “underwhelmed” with the current state of play should urge on the funding of a global program so that we be overwhelmed when we break the code of what makes us a thinking, knowing species.


  1. There is a long running thread on 23&me about the association between neanderthal dna and IQ. People post their IQ/neandertha dna percentage amid a bunch of odd and often snide comments mostly put there to keep the thread bumped up so new people will see it and post.
    Some woman got on there and suggest IQ tests correlate with Neanderthal dna because the people who created the IQ tests had Neanderthal dna. Apparently this is the new and improved genetic version of saying IQ test don't measure anything of relevance at all. I got the impression this was the sort of woman who would have spent a ridiculous amount of money making sure her children were tutored in Mandarin and simply can't take the idea anyone she regards as of the lower (white) classes might have something that would benefit her children a little more. Or, perhaps she's a school teacher looking to disregard those pesky standards. I found her response quite depressing and it actually took me a moment to realize she wasn't suggesting something involving time or space travel.

  2. "Against their will they are driving up the standard of proof."

    I know little about these things but I know that soft science has an inherent interest in lowering standards of proof, among other reasons because it leads to more publications, and there is a wider constituency for soft science. I could easily imagine data dredgers and the like getting in on the genomic act in the hope that their noise will drown out the actual signal. Is there any sign of it happening?

  3. Not as far as I know. I think that these researchers are trying to do a good job, and have learned from the early phase of research that small samples and multiple comparisons lead to false positives, and they very much want to avoid that. Agree that social sciences usually want things soft, but not the behavioural geneticists this time round.

  4. This comment has been removed by the author.

  5. At OpenPsych we're currently running a low-cost GWAS. We've found a set of candidate IQ genes or a "polygene". For a preliminary report, visit this link:

  6. James,

    As you know, despite the small number of intelligence-enhancing alleles uncovered, Davide Piffer (2013, 2014) has made progress in addressing ethnic differences. The early findings, while subject to modification with additional research, are quite striking: the allele frequencies are highest among East Asians, followed by ethnic Europeans, and are rather low among sub-Saharan Africans--Pygmies and Bushmen score lowest. A win for "race realists" is in sight in this respect. On the other hand, if one takes an average British IQ of 100 and an average score of 85 for African-Americans (a one standard deviation difference), the allele frequencies have some interesting implications for genotypic IQs. In Piffer (2014, published this month) the difference in the allele factor scores for Brits and African-Americans amounted to about 2 standard deviations. The implication is that, if the average genotypic IQ of African-Americans is (at least) 85, then a 2-standard-deviation difference in the allele factor score corresponds to (no more than) 7.5 genotypic-IQ points. If one takes the allele factor scores from Piffer (2014, Table 2), the difference between Han Chinese and Brits in terms of average genotypic IQ would be only a few IQ points. The difference between Brits and Bangladeshi Bengalis would be about 5 genotypic IQ points. Vietnamese apparently have higher genotypic IQs than Brits. According to Piffer (2013), Cambodians would be similar to Brits. Mexican Americans in Los Angeles would have a deficit of 5-6 IQ points. In summary, race realists stand to score a win with respect to people of sub-Saharan African descent (Bushmen, Yoruban-West Africans and Luhya Kenyans scoring lower than African Americans), but other ethnic differences in genotypic IQ are arguably rather modest.

  7. The first one gives 90 for Africans out of the 10. Even Piffer had article on it. The second out of 7 alleles the gap is closer because one of them is actually more common in Sub Saharan Africans, but I haven't checked the numbers for genotype using it.

    "despite the small number of intelligence-enhancing alleles uncovered"

    Yeah the problem here is that they didn't uncover IQ enhancing alleles. They correlated some using a tentative significance value.

    1. Also only 6 replicate(sometimes). 7th one is from small samples and didn't show up in any of the big ones, which in some, some of the big ones didn't either.

  8. Also there is recent evidence of the IQ gap in UK closing. In some of the data its completely gone.

  9. just more sulfurous wind from a retarded psychologist and british snob, but that's redundant.

    nessie has also proved elusive. as has bigfoot and ufos.

    maybe dats 'cause dey don't exist govna

    ---BGI volunteer