Forum: Science & Society

Our fragile intellect. Part I

Gerald R. Crabtree
Beckman Center, B211, Stanford University, 279 Campus Drive, Stanford, CA 94305, USA

New developments in genetics, anthropology, and neurobiology predict that a very large number of genes underlie our intellectual and emotional abilities, making these abilities genetically surprisingly fragile. I would wager that if an average citizen from Athens of 1000 BC were to appear suddenly among us, he or she would be among the brightest and most intellectually alive of our colleagues and companions, with a good memory, a broad range of ideas, and a clear-sighted view of important issues. Furthermore, I would guess that he or she would be among the most emotionally stable of our friends and colleagues. I would also make this wager for the ancient inhabitants of Africa, Asia, India, or the Americas, of perhaps 20006000 years ago. The basis for my wager comes from new developments in genetics, anthropology, and neurobiology that make a clear prediction that our intellectual and emotional abilities are genetically surprisingly fragile. To assess this fragility we must rst know how many genes are required for our intellectual abilities. The larger the number of genes required, the more susceptible we are as a species to random genetic events that reduce our intellectual and emotional tness. Recently, the means to answer this question have emerged from genetic studies and insights into the human genome. Several lines of evidence indicate that the number of genes required for normal human intelligence might be quite large. Perhaps the most effective way to estimate the number of genes in humans that are needed for full intellectual function is to rely on studies of X-linked intellectual deciency (XLID). Because males have only one X chromosome, the effects of X-chromosome mutations cannot be rescued or compensated for by the second copy, in contrast to mutations on other chromosomes. Present studies indicate that mutation of about 215 intellectual deciency (ID) genes on the X chromosome give rise to XLID and/or emotional disability [1,2]; this represents about 25% of the genes on the X chromosome. Of these, 86 have been characterized and do not seem to be neomorphs (a gain of inappropriate function). This gives a conservative estimate that about 10% of all human genes are implicated in intellectual function. Because mutation of any one of these genes can give rise to intellectual disability, it can be concluded that they do not operate as a robust network, but rather as links in a chain, failure of any one of which leads to intellectual disability. The X chromosome does not appear to be enriched for genes required for intellectual development [3], and therefore we can extrapolate that
Corresponding author: Crabtree, G.R. (crabtree@stanford.edu).

between 2000 and 5000 genes are needed for intellectual and emotional function. This is supported by the nding that autosomal recessive mental retardation seems to be very heterogeneous, even within a genetically similar background, indicating that it is due to mutations in many genes [3,4]. Many of these genes appear to function indirectly, such as the subunits of nBAF chromatin regulatory complexes, which are global transcriptional regulators [5]. This highlights that a gene need not be functionally brainor even human-specic to be essential for our specic human intellectual abilities. Finally, the number of genes that function as links in a chain to support normal intellect is reected in the frequency with which human genetic diseases in general have an ID component. A recent study of the Online Mendelian Inheritance in Man (OMIM) database, although incomplete, indicates that about half of all human genetic diseases have a neurologic component [6], frequently including some aspect of ID, consistent with the notion that many genes are required for intellectual and emotional function. The reported mutations have been severe alleles, often de novo mutations that reduce fecundity. However, each of these genes will also be subject to dozens if not hundreds of weaker mutations that lead to reduced function, but would not signicantly impair fecundity, and hence could accumulate with time. Based on estimates of the frequency with which deleterious mutations appear in the human genome (Box 1), and the assumption that 20005000 genes are required for intellectual ability, it is very likely that within 3000 years (120 generations) we have all sustained two or more mutations harmful to our intellectual or emotional stability. Recent human genome studies revealed that there are, per generation, about 60 new mutations per genome and about 100 heterozygous mutations per genome that are predicted to produce a loss of function [7], some of which are likely to affect genes involved in human intellect. However, heterozygous mutations (affecting only one copy) are generally not considered problematic until they are reduced to homozygosity. But new discoveries indicate that the human nervous system is uniquely susceptible to loss of heterozygosity (LOH). LINE-1 (L1) repetitive elements were recently reported to transpose in human neurons, leading to neuronal gene inactivation [8]. The somatic origin of these transpositions was demonstrated by direct sequencing of different brain regions [9], which revealed that other repetitive elements could also transpose and insert into or regulate critical neurodevelopmental genes. Indeed, they have a tendency to insert into transcribed genes, modifying transcription [10]. The L1 insertions occur in neural stem cells and lead to clones of neurons with specic insertion sites. Each
1

Forum: Science & Society

Trends in Genetics January 2013, Vol. 29, No. 1

Box 1. Estimation of the rate of accumulation of harmful mutations in ID genes

If the proper function of 20005000 genes is necessary for our intellectual ability, then in the simplest case emotional and intellectual fitness will drift with reduced or absent selection 2000 5000-fold more rapidly that a trait specified by a single gene. Studies in humans using phenotypic methods have estimated that the germline suffers about one new deleterious mutation per average protein-coding gene per 100 000 generations [9]. These are probably mostly point mutations that compromise gene function without totally inactivating it. Recently, an independent estimate of the rate of germline mutations was made from direct genomic sequencing of parents and their children, which found about 25 new mutations (if the father was under 20 years old) and about 65 new mutations (if the father was over 40) over the non-repetitive regions of the genome per generation [14,16]. This analysis predicts about 5000 new mutations in the past 3000 years (120 generations). Of these new germline mutations in non-repetitive regions only a small fraction (variously estimated to be about 110%) will produce a change within a gene or its regulatory regions that will be harmful, and a vanishingly small fraction will increase fitness. This approach gives estimates of the rate of appearance of new mutations consistent with the older phenotypic estimates. However, the most recent estimate is that each child suffers six new harmful heterozygous mutations [16]. Thus the probability of any random gene suffering a harmful mutation in a given generation is at the very least 1/100 000. If indeed 20005000 genes are necessary for our intellectual and emotional stability, then we need to multiply this rate by 2000 or 5000 for the full collection of ID genes. Recall that mutation in any one of these genes produces intellectual or emotional deficiency. Thus the probability that any child will have a new mutation affecting intellectual ability is between 2000 and 5000 over 100 000. This figure predicts that about one newborn child in 2050 should harbor such a mutation. Put another way, every 20 50 generations we should sustain a mutation in one copy of one of our many ID genes. In the past 3000 years then (120 generations), each of us should have accumulated at the very least 2.56 mutations in ID genes. Of course, these will be haploid mutations and will usually be rescued by the other normal allele, highlighting the importance of examining the effect of being haploid for any one of these genes.

neuron is estimated to sustain about 80 L1 insertions, indicating that gene expression is dysregulated in most neurons. These insertions can lead to inactivation of the remaining normal allele of a heterozygous essential gene in a clone of neural stem cells, thereby creating a focal defect in the brain. Many neurons with deleterious insertions might be eliminated by their failure to form effective neural circuits, but this is clearly not always the case because L1 insertions into ID genes have been documented [9]. A practical implication of these studies is that identical twins will have different neuronal subpopulations, and hence the contribution of genetic factors will be underestimated in classic twin studies. It is also worth noting that the number of genes estimated to underlie intellectual function by this means would be much larger than that estimated by the X chromosome analysis because even genes whose mutation causes embryonic lethality could be inactivated by the insertion of mobile elements such as L1 transposons. Another less obvious consequence is that that this route to homozygosity will make intellectual ability less heritable, which necessitates stronger selective pressure to maintain neurologic traits (more on this later). Another route to homozygous inactivation, in individuals already bearing a germline mutation in one allele of a
2

gene required for intellectual tness, is a feature of the nervous system that has recently come to light: apparently between 10% and 50% of human neurons are aneuploid (i.e., they have chromosomal abnormalities that lead to breaks, losses, and duplications of genetic material) [11]. Again, it appears that aneuploidy might originate in neural stem cells and hence be clonal, thereby resulting in a focal loss of function in a specic region of the brain. Furthermore, neurons with aneuploid genomes form genetically mosaic neural circuitries as part of the normal organization of the mammalian brain [12]. Aneuploidy of chromosome 21 is of course the basis of Down syndrome, which is accompanied by a reduction in intellectual function, and illustrates the effect of alterations in gene copy number. Copy-number variation appears to have a role in several neurologic diseases including autism [13]. The above two arguments suggest that focal LOH might be an underlying feature of neurologic diseases, which would be difcult to detect by present-day genome sequencing approaches. To identify focal LOH, neurons from many regions of the brain would need to be sampled and their DNA sequenced. Because aneuploidy and transposon insertion are non-germline routes to homozygous gene inactivation, both would lead to misinterpretation of studies with identical twins and make neurologic traits more difcult to maintain by selection. A third, and possibly even more harmful, effect of heterozygosity occurs when alleles of two or more genes involved in intellectual or emotional function are mutated. The calculations in Box 1, and recent population genome sequencing studies [14], suggest that most of us are heterozygous for mutations affecting two or more of the 2000 5000 genes estimated to be required for intellectual function. Heterozygous inactivation of two or more genes encoding proteins within the same biochemical pathway, genetic circuit, or protein complex can reduce function, similarly to a single homozygous mutation. One recent example is the nding that ID can be produced by mutation of at least six subunits of the nBAF complex. In a given individual, different heterozygous mutations appear to lead to reduced function and ID [5]. In general, it is difcult to know if loss of one allele in, for example, an enzyme removing a neurotoxic intermediate would cause defects in an individual heterozygous for a gene required for dendritic morphogenesis. These considerations make human genetic studies designed to nd the genes at fault in human cognitive disorders difcult, but double or compound heterozygosity would almost certainly contribute to reduced function among the estimated 20005000 genes required for full intellectual and emotional function, and compound heterozygosity will operate exponentially over time as deleterious heterozygous mutations accumulate in our genome at a linear rate. Taken together, the large number of genes required for intellectual and emotional function, and the unique susceptibility of these genes to loss of heterozygosity, lead me to conclude that we, as a species, are surprisingly intellectually fragile and perhaps reached a peak 20006000 years ago. But if we are losing our intellectual abilities, how did we acquire them in the rst place? This will be the topic of the next section [15].

Forum: Science & Society

References
1 Stevenson, R.E. and Schwartz, C.E. (2009) X-linked intellectual disability: unique vulnerability of the male genome. Dev. Disabil. Res. Rev. 15, 361368 2 Gecz, J. et al. (2009) The genetic landscape of intellectual disability arising from chromosome X. Trends Genet. 25, 308316 3 Inlow, J.K. and Restifo, L.L. (2004) Molecular and comparative genetics of mental retardation. Genetics 166, 835881 4 Kuss, A.W. et al. (2011) Autosomal recessive mental retardation: homozygosity mapping identies 27 single linkage intervals, at least 14 novel loci and several mutation hotspots. Hum. Genet. 129, 141148 5 Tsurusaki, Y. et al. (2012) Mutations affecting components of the SWI/SNF complex cause CofnSiris syndrome. Nat. Genet. 44, 376 378 6 Marin, O. and Gleeson, J.G. (2011) Function follows form: understanding brain function from a genetic perspective. Curr. Opin. Genet. Dev. 21, 237239 7 MacArthur, D.G. et al. (2012) A systematic survey of loss-of-function variants in human protein-coding genes. Science 335, 823828 8 Coufal, N.G. et al. (2009) L1 retrotransposition in human neural progenitor cells. Nature 460, 11271131

Trends in Genetics January 2013, Vol. 29, No. 1

9 Baillie, J.K. et al. (2011) Somatic retrotransposition alters the genetic landscape of the human brain. Nature 479, 534537 10 Han, J.S. et al. (2004) Transcriptional disruption by the L1 retrotransposon and implications for mammalian transcriptomes. Nature 429, 268274 11 Westra, J.W. et al. (2010) Neuronal DNA content variation (DCV) with regional and individual differences in the human brain. J. Comp. Neurol. 518, 39814000 12 Kingsbury, M.A. et al. (2005) Aneuploid neurons are functionally active and integrated into brain circuitry. Proc. Natl. Acad. Sci. U.S.A. 102, 61436147 13 Cook, E.H., Jr and Scherer, S.W. (2008) Copy-number variations associated with neuropsychiatric conditions. Nature 455, 919923 14 1000 Genomes Project Consortium (2010) A map of human genome variation from population-scale sequencing. Nature 467, 10611073 15 Crabtree, G.R. (2012) Our fragile intellect. Part II. Trends Genet. http://dx.doi.org/10.1016/j.tig.2012.10.003 16 Kong, A. et al. (2012) Rate of de novo mutations and the importance of fathers age to disease risk. Nature 488, 471475
0168-9525/$ see front matter 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tig.2012.10.002 Trends in Genetics, January 2013, Vol. 29, No. 1

Our fragile intellect. Part II

Gerald R. Crabtree
Beckman Center, B211, Stanford University, 279 Campus Drive, Stanford, CA 94305, USA

Analysis of human mutation rates and the number of genes required for human intellectual and emotional tness indicates that we are almost certainly losing these abilities. If so, how did we get them in the rst place, and when did things begin to change? In the previous piece I argued that we are slowly losing emotional and intellectual traits [1]; but how did we get them in the rst place? This is one of the most important questions of modern anthropology and the subject of much investigation and debate. Although not my area of expertise, I offer the following thoughts on the topic. Expansion of the human frontal cortex and endocranial volume (Figure 1), to which we likely owe our capacity for abstract thought, predominately occurred between 50 000 and 500 000 years ago [2,3] in our prehistoric African ancestors, well before written language and before we had the modern voice-box to produce sophisticated verbal language [2,4], but after the rst tools. Thus, the selective pressures that gave us our mental characteristics operated among nonverbal huntergatherers living in dispersed bands or villages, nothing like our present-day high-density, supportive societies. Furthermore, it seems that our intellectual capacity has not evolved at different rates since our African ancestors began their migrations, based on the fact that geographically disparate societies have near-identical intellectual capacities. For example, written language was independently invented by the group with the longest migration path, the Indians of Middle and South America, and by the people with the one of the shortest migration
Corresponding author: Crabtree, G.R. (crabtree@stanford.edu).

paths, the Sumerians. In addition, whether a migration group lived a high-density urban life or as dispersed huntergatherers did not greatly inuence their intellectual development. Thus, to understand how 20005000 genes [1] were optimized for abstract thought, we almost certainly have to look to this period 50 000500 000 years ago and to ancestors common to all humans today. Somehow the selective pressures on these ancestors led to the evolution of a brain capable of writing symphonies and performing higher mathematics, indicating that life as a huntergatherer was more intellectually demanding than we might think. To understand the extremes of selection that must have occurred as our ancestors went from using speed, strength, and agility, to employing intellect to survive (a process that occurred over about 1 million years), consider the difculty of optimizing 20005000 genes. Because retrotransposon insertion and aneuploidy of neurons reduce the heritability of neuronal traits [1], the selective pressure required is increased. Present studies indicate that the heritability of intelligence, judged largely by IQ scores, is between 0.5 and 0.7, and have called attention to important variables affecting these estimates [57]. This level of heritability indicates that greater natural selection is necessary to maintain this trait compared to eye color, for example, where the heritability is much higher. In addition, one would need to sum the selective pressure for each of the genes operating independently to produce the trait. Thus, extraordinary natural selection was necessary to optimize and maintain such a large set of intelligence genes. This optimization probably occurred in a world where every individual was exposed to natures raw selective
3

Forum: Science & Society

References
1 Stevenson, R.E. and Schwartz, C.E. (2009) X-linked intellectual disability: unique vulnerability of the male genome. Dev. Disabil. Res. Rev. 15, 361368 2 Gecz, J. et al. (2009) The genetic landscape of intellectual disability arising from chromosome X. Trends Genet. 25, 308316 3 Inlow, J.K. and Restifo, L.L. (2004) Molecular and comparative genetics of mental retardation. Genetics 166, 835881 4 Kuss, A.W. et al. (2011) Autosomal recessive mental retardation: homozygosity mapping identies 27 single linkage intervals, at least 14 novel loci and several mutation hotspots. Hum. Genet. 129, 141148 5 Tsurusaki, Y. et al. (2012) Mutations affecting components of the SWI/SNF complex cause CofnSiris syndrome. Nat. Genet. 44, 376 378 6 Marin, O. and Gleeson, J.G. (2011) Function follows form: understanding brain function from a genetic perspective. Curr. Opin. Genet. Dev. 21, 237239 7 MacArthur, D.G. et al. (2012) A systematic survey of loss-of-function variants in human protein-coding genes. Science 335, 823828 8 Coufal, N.G. et al. (2009) L1 retrotransposition in human neural progenitor cells. Nature 460, 11271131

Trends in Genetics January 2013, Vol. 29, No. 1

9 Baillie, J.K. et al. (2011) Somatic retrotransposition alters the genetic landscape of the human brain. Nature 479, 534537 10 Han, J.S. et al. (2004) Transcriptional disruption by the L1 retrotransposon and implications for mammalian transcriptomes. Nature 429, 268274 11 Westra, J.W. et al. (2010) Neuronal DNA content variation (DCV) with regional and individual differences in the human brain. J. Comp. Neurol. 518, 39814000 12 Kingsbury, M.A. et al. (2005) Aneuploid neurons are functionally active and integrated into brain circuitry. Proc. Natl. Acad. Sci. U.S.A. 102, 61436147 13 Cook, E.H., Jr and Scherer, S.W. (2008) Copy-number variations associated with neuropsychiatric conditions. Nature 455, 919923 14 1000 Genomes Project Consortium (2010) A map of human genome variation from population-scale sequencing. Nature 467, 10611073 15 Crabtree, G.R. (2012) Our fragile intellect. Part II. Trends Genet. http://dx.doi.org/10.1016/j.tig.2012.10.003 16 Kong, A. et al. (2012) Rate of de novo mutations and the importance of fathers age to disease risk. Nature 488, 471475
0168-9525/$ see front matter 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tig.2012.10.002 Trends in Genetics, January 2013, Vol. 29, No. 1

Our fragile intellect. Part II

Gerald R. Crabtree
Beckman Center, B211, Stanford University, 279 Campus Drive, Stanford, CA 94305, USA

Analysis of human mutation rates and the number of genes required for human intellectual and emotional tness indicates that we are almost certainly losing these abilities. If so, how did we get them in the rst place, and when did things begin to change? In the previous piece I argued that we are slowly losing emotional and intellectual traits [1]; but how did we get them in the rst place? This is one of the most important questions of modern anthropology and the subject of much investigation and debate. Although not my area of expertise, I offer the following thoughts on the topic. Expansion of the human frontal cortex and endocranial volume (Figure 1), to which we likely owe our capacity for abstract thought, predominately occurred between 50 000 and 500 000 years ago [2,3] in our prehistoric African ancestors, well before written language and before we had the modern voice-box to produce sophisticated verbal language [2,4], but after the rst tools. Thus, the selective pressures that gave us our mental characteristics operated among nonverbal huntergatherers living in dispersed bands or villages, nothing like our present-day high-density, supportive societies. Furthermore, it seems that our intellectual capacity has not evolved at different rates since our African ancestors began their migrations, based on the fact that geographically disparate societies have near-identical intellectual capacities. For example, written language was independently invented by the group with the longest migration path, the Indians of Middle and South America, and by the people with the one of the shortest migration
Corresponding author: Crabtree, G.R. (crabtree@stanford.edu).

paths, the Sumerians. In addition, whether a migration group lived a high-density urban life or as dispersed huntergatherers did not greatly inuence their intellectual development. Thus, to understand how 20005000 genes [1] were optimized for abstract thought, we almost certainly have to look to this period 50 000500 000 years ago and to ancestors common to all humans today. Somehow the selective pressures on these ancestors led to the evolution of a brain capable of writing symphonies and performing higher mathematics, indicating that life as a huntergatherer was more intellectually demanding than we might think. To understand the extremes of selection that must have occurred as our ancestors went from using speed, strength, and agility, to employing intellect to survive (a process that occurred over about 1 million years), consider the difculty of optimizing 20005000 genes. Because retrotransposon insertion and aneuploidy of neurons reduce the heritability of neuronal traits [1], the selective pressure required is increased. Present studies indicate that the heritability of intelligence, judged largely by IQ scores, is between 0.5 and 0.7, and have called attention to important variables affecting these estimates [57]. This level of heritability indicates that greater natural selection is necessary to maintain this trait compared to eye color, for example, where the heritability is much higher. In addition, one would need to sum the selective pressure for each of the genes operating independently to produce the trait. Thus, extraordinary natural selection was necessary to optimize and maintain such a large set of intelligence genes. This optimization probably occurred in a world where every individual was exposed to natures raw selective
3

Forum: Science & Society

Wrien language 2000 Spoken language

Trends in Genetics January 2013, Vol. 29, No. 1

Modern human endocranial volume +/- 2 SD 1500

Earliest spears 1000

500

Modern chimpanzee Endocranial volume +/- 2 SD

0 0 500 000 1 000 000 1 500 000 2 000 000 2 500 000

Years before present

TRENDS in Genetics

Figure 1. Expansion of endocranial volume during the past 2.5 million years among Homo sapiens ancestors. Modified from R.G. Klein [2]. Note that language follows the increase in endocranial volume and therefore the use of language cannot cause the expansion. Also, note that the appearance of first tools such as spears correlates well with increased cranial volume.

mechanisms on a daily basis. In the transition to surviving by thinking, most people (our non-ancestors) probably died simply due to errors of judgment or a lack of an intuitive, non-verbal comprehension of things such as the aerodynamics and gyroscopic stabilization of a spear while hunting a large, dangerous animal. How did we get from accurately throwing a spear to the theory of relativity? The eld of articial intelligence (AI) may shed light on this question. AI promised household robots that would wash dishes, mow the lawn, and bring us freshly cooked croissants and coffee in the morning. Needless to say we do not have these robots now and none of the readers of this piece will probably ever see them, despite the immense nancial impetus to build them. This is because common tasks are actually conceptually complex. However, AI is very good at things we supercially consider intellectual, such as playing chess, winning Jeopardy, ying a jet plane, or driving a car. To understand this paradox, consider the game Foldit, in which players predict protein structures [8]. Humans beat supercomputers at this game in much the same way that we can wash the dishes and put them away better than a robot. We win because Foldit uses spatial reasoning skills that were perfected and selected for in our huntergatherer ancestors. Many kinds of modern rened intellectual activity (by which our children are judged) may not necessarily require more innovation, synthesis, or creativity than more ancient forms: inventing the bow-andarrow, which seems to have occurred only once about 40 000 years ago, was probably as complex an intellectual
4

task as inventing language. Selection could easily have operated on common (but computationally complex) tasks such as building a shelter, and then computationally simple tasks, such as playing chess, became possible as a collateral effect. Loss of any one of 20005000 genes prevents us from effectively doing everyday tasks, and selection for the ability to perform them would optimize the function of the entire group of genes. When might we have begun to lose these abilities? Most likely we started our slide with high-density living, which was enabled by the transformative invention of agriculture. Selection may have begun operating on resistance to the diseases that naturally grow out of high-density living, switching the pressure from intelligence to immunity. It is also likely that the need for intelligence was reduced as we began to live in supportive societies that made up for lapses of judgment or failures of comprehension. Community life would, I believe, tend to reduce the selective pressure placed on every individual, every day of their life. Indeed that is why I prefer to live in such a society. Several considerations could mitigate the validity of the argument that intellectual and emotional tness are slowly decaying. For example, genes required for intellectual and emotional function could be needed for early development or even fertility, and would thus be maintained through selection. Indeed, some genes associated with Xlinked intellectual deciency (XLID) are also involved in the development or function of other tissues or organs. However, these other syndromic features are not lethal, and many do not impair reproduction. Although the multiple usage of genes could slow the rate of accumulation of mutations in intellectual tness genes, if the estimates for the number of genes required and mutation rate are correct, and selection is only slightly relaxed, one would still conclude that nearly all of us are compromised compared to our ancient ancestors of 30006000 years ago. Another common counter-argument is that we are under constant selection for our intellectual traits (Box 1).
Box 1. The Flynn effect
The famous Flynn effect, in which absolute IQ scores increased during the first 50 years after the institution of these tests, seems at first glance to contradict the hypothesis that we are losing our intellectual abilities. However, these changes in IQ scores are probably linked to environmental influences including reduction in lead and other heavy metals used in gasoline and paint and the virtual elimination of hypothyroidism in children due to the widespread use of iodinated salt. These and many other advances in prenatal care and prevention of anoxia during childbirth have clear effects on our average intellectual abilities. In addition, scores on these tests have been shown to correlate well with preschooling and other societal influences instituted during the period over which test scores were compared. Consistent with these hypotheses, the gains recorded are predominantly in the raising of lower scores. However, since about 1985 or 1990 these absolute IQ scores have been dropping in some studies, despite considerable teaching to the test as well as the general awareness among children that it is important to score well on the test. Most likely these short-term effects are not genetic because the genetic effects at issue are only likely to operate over hundreds of years, not decades. In addition, Flynn points out that we are not getting more intelligent, but instead we are getting smarter at taking the tests because our everyday experiences are becoming more like the tests. This is similar to the way that practicing tennis makes one a better badminton player.

Endocranial volume (cc)

Forum: Science & Society

Years before present day Individual 3000 2500 2000 E F G

Trends in Genetics January 2013, Vol. 29, No. 1

5000 A

4500 B

4000 C

3500 D

1500 H

1000 I

500 J

0 K JK t IK t

Hunter-gather period

Time
TRENDS in Genetics

Figure 2. Genetic triangulation to measure rate of change of ID genes over the past 5000 years based on genome sequences of present-day individuals with last common ancestors separated by specific times, Dt (500 years in this case for illustration). The bar at the top indicates the transition from huntergatherer to a more high-density lifestyle when selection based on resistance to infection might have begun to dominate over selection for intellectually or emotional fitness.

Intellectual capacity and emotional stability have mating advantages that would reduce the rate at which mutations affecting these traits become xed in our genome. This is true, but I fear does not take into account the extreme selection required to maintain traits dependent upon thousands of genes with reduced heritability. A huntergatherer who did not correctly conceive a solution to providing food or shelter probably died, along with his/her progeny, whereas a modern Wall Street executive that made a similar conceptual mistake would receive a substantial bonus and be a more attractive mate. Clearly, extreme selection is a thing of the past. The hypothesis that genes critical to intellectual function are decaying could be tested by a form of genetic triangulation (Figure 2). Sequencing many individuals whose last common ancestors ranged from present day to 5000 years ago should produce an estimate of the rapidity of change and the level of selection operating on these genomes at various time-intervals during this 5000 year period (an interval that would span the emergence of cities for several population groups). These genomes would have sequence variations in intellectual deciency (ID) genes but, because each generation produces only about 40 new signature mutations, these would not be enough to guide the temporal ordering. However, using the 400 million retroviral insertion sites as signatures of a specic ancient generation to estimate the age of nearby new mutations, sufcient neness might be produced to permit dating of mutations in ID genes. Because mutations that control the evolution of specic characteristics have often been found in regulatory rather than coding regions, full genome sequences would need to be assayed. I would be very happy to learn from this test that there is no substance to my argument. However, if such a study found accelerating rates of accumulation of deleterious alleles over the past several

thousand years then we would have to think about these issues more seriously. But we would not have to think too fast. One does not need to imagine a day when we could no longer comprehend the problem, or counteract the slow decay in the genes underlying our intellectual tness, or have visions of the world population docilely watching reruns on televisions they can no longer build. It is exceedingly unlikely that a few hundred years will make any difference for the rate of change that might be occurring. Remarkably, it seems that although our genomes are fragile, our society is robust almost entirely by virtue of education, which allows strengths to be rapidly distributed to all members. The sciences have come so far in the past 100 years that we can safely predict that the accelerating rate of knowledge accumulation within our intellectually robust society will lead to the solution of this potentially very difcult problem by socially and morally acceptable means. But in the meantime Im going to have another beer and watch my favorite rerun of Miami CSI (if I can gure out how to work the remote control).
References
1 Crabtree, G.R. (2012) Our fragile intellect. Part I. Trends Genet. http:// dx.doi.org/10.1016/j.tig.2012.10.002 2 Klein, R.G. (1999) The Human Career, University of Chicago Press 3 Aiello, L.C. and Dunbar, R.J.M. (1993) Neocortical size, group size and the evolution of language. Curr. Anthropol. 34, 184192 4 Holden, C. (1998) No last word on language origins. Science 282, 1455 5 Turkheimer, E. et al. (2003) Socioeconomic status modies heritability of IQ in young children. Psychol. Sci. 14, 623628 6 Devlin, B. et al. (1997) The heritability of IQ. Nature 388, 468471 7 Bouchard, T.J., Jr et al. (1990) Sources of human psychological differences: the Minnesota Study of Twins Reared Apart. Science 250, 223228 8 Cooper, S. et al. (2010) Predicting protein structures with a multiplayer online game. Nature 466, 756760
0168-9525/$ see front matter 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tig.2012.10.003 Trends in Genetics, January 2013, Vol. 29, No. 1

La

st c

om

on

an

ce

st e

rs