Source: here.
Most people know that there is something tangible called intelligence. Ancient Hindus saw it as a tangible entity called buddhi, while modern psychometrics uses the concept of general intelligence called g. In the modern theory it is proposed that irrespective of the domain of specialization, and despite the presence of general abilities such as verbal ability, or spatial ability or numerical ability that contribute to intelligence, the main common determinant of intelligence is this generic factor termed g. IQ tests are designed to be as good a proxy for this g as possible. Generally this theory of g also makes intuitive sense and fits in with our anecdotal observations: Most scientists of the type who need to use their mental faculties for their profession have IQ higher than 135. There is normal distribution of IQ, which fits well with the anecdotal observation that the majority of people in the population fall around a certain average intelligence, a yard-stick used for educational testing. There are few highly intelligent people in any population. g is a quantitative trait with broad sense heritability H^2~0.5-0.8 showing that it is substantially genetically determined. It is correlated with brain anatomy like frontal lobe gray matter volume and brain physiology like glucose utilization. Thus, the available evidence strongly supports intelligence as a biologically determined trait with a strong genetic component.
So what are the genes that affect human intelligence? As I was studying brain function genes in context of the pathogen-response of mammalian hosts to a well-known vertebrate parasite, I was distracted to survey intelligence related genes. I gathered a bit of biology in this regard mainly for my own edification, but in course of a conversation with R decided that it might be worthwhile putting down some of it here in the near future.
A brief summary:
CHRM2: The cholinergic muscarinic 2 receptor gene encodes a 7 TM receptor of acetylcholine expressed in the brain (and heart). This gene shows a highly significant association with intelligence with the strongest association seen with the SNP: rs324650. The T allele (as against the ancestral A allele) of this gene was found to correlate with an increase in what the authors call performance IQ by about 4.6 points. This SNP is seen in the 3’UTR intron suggesting that its biological action is regulatory, probably at the transcriptional level. In studied human populations the T allele shows about 91% frequency in Chinese and Japanese, 47% in White Americans, and 27% in West Africans.
DTNBP1: Several SNPs in this gene were associated with IQ. This gene encodes a cytoskeletal coiled-coil protein that affects dystrobrevin localization in axons and might thereby affect axon architecture. rs760761, rs2619522, rs2619538 are the 3 main SNPs that affect what the authors term full-scale IQ. The first 2 derived SNPs reduce IQ while the 3rd one increases IQ, each by about 6-7 points. Again these seem to cause regulatory effects as they are found in non-coding intronic regions. Each allele shows clear population differences: For example in the last SNP the T allele is similar in proportion in both West Africans and White Americans, but is extremely rare in the East Asians.
SNAP25: This gene encodes a protein involved in vesicular fusion, with two coiled coil SNARE modules and cysteines which are palmitoylated. 4 SNPs in the intron-1 of this gene: rs363043, rs353016, rs363039 and rs363050 have an effect of IQ. These SNPs appear to alter transcription factor binding sites and alter the expression pattern of the SNAP-25 gene. Some of these alleles again show dramatic population differences.
NRG1: Neuregulin-1 is a cell-surface protein which is the ligand for the ERBB3 and ERBB4 receptor tyrosine kinases. It induces the expression of acetylcholine receptor and induces Schwann cell proliferation. One variant of it in the promoter region showed reduced IQ and frontal/temporal lobe activity and pre-disposition for pyschosis.
LIMK1: The LIM domain kinase 1 with two LIM domains fused to a kinase domain, has been implicated in William’s syndrome. Effects on visuo-spatial cognition and the tendency of individuals with LIMK1 deletion to anthropomorphize non-human entities implicate it in both spatial ability and general reconstruction of imagery. Mice with LIMK1 deletion also show altered spatial abilities and fear responses. Its potential interaction with the cytoplasmic tail of NRG1 implicate it in a common pathway with that gene product. This might be explored in the future for a role in intelligence.
RIMS1: Encodes a protein with N-terminal Zn-chelating Rabphilin-effector domain fused to a C-terminal PDZ, and 2 C2 domains. Appears to be important in regulating synaptic membrane exocytosis via the Rab3 GTPase during neurotransmitter release. Importantly, it physically interacts with SNAP-25 which has also been implicated in intelligence and vesicular fusion. Here a mutation in the coding region, resulting in a R844H (gi:2224621) substitution, in turn results in increased IQ, especially verbal IQ [ekanetra, I wonder if you have this mutation!]. This mutation is in the C2 domain and is close to the residues interacting directly with Calcium. We believe that the enhanced RIMS1 IQ phenotype might be linked to altered calcium affinity of the molecule. The down side of this IQ gain is an associated vision defect phenotype. Mice lacking this gene show severely impaired learning and memory
COMT: catechol O-methyltransferase which is involved in catechol amine degradation has a common polymorphism V158M that has been implicated in cognitive differences with individuals with the M allele performing better. The M allele results in a less active enzyme and concomitantly greater dopamine concentrations, suggesting that its effects are a consequence of dopamine concentration. However, its effect might be relatively subtle. The M/M homozygotes are relative rare globally compared to V/V homozygotes or M/V heterozygotes. However, there is some population differentiation of the variants with the maximum M/M presence in White population and relatively low presence in Chinese/Japanese and intermediate presence in Sub-Saharan Africans.
Further, quantitative trait associations studies have found linkages between IQ and two chromosomal regions namely 2q24.1-31.1 and 6p25.3-22.3. These respectively overlap with chromosomal regions linked with autism and reading disability/dyslexia. Several genes lie in these chromosomal locations that have emerged as potential candidates in other cognitive disorders. E.g. : SLC25A12: autism; NR4A2, DTNBP1, KIF13A, NQO2: schizophrenia; RANBP9: fragile X syndrome; BBS5: Bardet-Biedl syndrome. There are also other uncharacterized candidates in these regions like: KCNH7, Neuritin 1, HTR5A, HTR3A and HTR3B. Thus, there several indications that other genes with a role in individual differences in intelligence might be uncovered in the near future.
Location and function of the gene products of major genes implicated in intelligence (Click to enlarge).
An examination of the major genes implicated in intelligence shows an interesting pattern. 4 out of the 7 allelic variants with an effect on intelligence/cognition affect the regulatory regions, most likely transcription control sites in non-coding regions of the gene. The only known LIMK1 allele with a cognitive effect seems to be a null allele resulting in complete gene deletion. This suggests that a notable fraction of the variation in the genes resulting in intelligence differences is subtle and does not alter protein function. Instead, the chief effects of the variants are predicted to be in changing concentrations or amounts of proteins that are available in within the neuron. Only in two cases (RIMS1 and COMT) do we actually observe a change in protein functional properties, through alteration of the coding region. In at least the latter case the net result seems to be a change in dopamine concentration. These observations suggest that most effects of genetic variation on intelligence can in part be relatively simply modeled a changes in protein or neurotransmitter concentration which either directly affect neuronal architecture (e.g. DTNBP1) or amount of available neurotransmitter or its receptor. This also raises the possibility of relatively easily phenocopying such alterations through non-genetic, biochemical interventions. Thus, intelligence altering drugs, which have been a fascination from the earliest days of Hindu medicine might not be out of place.
Beyond these genes, a recent study (Cochran et al) on the possible genetic determinants for high IQ in the Ashkenazi Jewish population uncovered a series of potential candidate genes. So far there is no evidence for these having a general role in intelligence in association studies. However, as the authors suggest some of this may have a been uniquely selected in a particular Jewish population: elsewhere their negative fitness effects might have eliminated them from the population. The best of these candidates include genes like torsin encoding an AAA+ ATPase, mutated in Torsion dystonia. The mutation appears to affect a glutamate in the helical C-terminal module of the AAA+ domain and might hence affect its target protein interaction, and there by affect protein translocation. The other candidate is CYP11B1 gene which encodes a steroid 11-beta hydroxylase, which is mutated in non-classical congenital Adrenal Hyperplasia but its exact role if any in elevating IQ is unclear. The DNA repair group including BRCA1, the Bloom’s syndrome helicase and some others have also been implicated along with brain-size related genes with related functions such as ASPM and MCPH1. However, at least in the case of the latter two no direct association with IQ was found. However, ASPM and MCPH1 have been linked to the emergence of tonal languages and alphabetic writing, suggesting that in combination with some of the established IQ-affecting genes these and other DNA repair genes could impact cognitive capacity in facets other than that measure by proxies for g.
The case of FADS2: It has been known from earlier studies that breast-feed infants develop significantly higher IQ than non-breastfed infants. However, this increase in IQ due to breast-feeding is dependent on the presence of a particular allelic variant (SNP: rs174575) in the FADS2 gene. This polymorphism is close to a predicted sterol response element and might affect transcription factor binding. Individuals with the more prevalent FADS2 allele in the tested population (mainly White Anglo-Saxon) responded dramatically breast-feeding in terms of IQ, whereas those with the less prevalent allele showed neither IQ elevation nor depression in response breast-feeding. FADS2 encodes a membrane associated delta-6 fatty acid desaturase with a N-terminal cytochrome b5 domain and a C-terminal multi-TM desaturase domain that is involved in the synthesis of highly unsaturated fatty acids. It is possible that the activity of this enzyme on precursors derived from breast-milk are critical for generation of specific highly unsaturated fats accumulated in the brain. An earlier association study had linked the intronic SNP rs498793 in the same gene (different from the above one) with attention-deficit/hyperactivity disorder. This SNP shows dramatic difference between Chinese/Japanese one hand and Africans/Whites on the other. It would be interesting to see if this allele might account for behavioral differences between these ethnic groups, and whether it interacts with other alleles in the same locus. Here again the polymorphisms linked to phenotypic differences are merely involved in regulatory effects rather than changing protein biochemistry.
The case of FADS2 again raises the possibility that dietary supplements of particular fatty acids might be beneficial for development of intelligence. More generally, it shows how environmental factors themselves might uncover a further set of genes that affect development of intelligence.
This finally leads to the politically contentious issue of intelligence difference between different populations, ethnic groups or races. Psychometry has consistently suggested differences in average g as measured by IQ between different populations. The above survey of genetic variation implicated in individual intelligence differences also show marked population differences in their frequencies. This is consistent with a genetic basis of the inter-population differences in intelligence. However, we must keep in mind that we have not exhausted all the genes involved in intelligence differences, nor have exhausted the effects of all possible allelic variants found in different populations. So we cannot yet reconstruct purely ground up a measure of the amount of inter-population intelligence difference using the current molecular data.
मानसतरङ्गिणीकृत्
Terminal sialidase Klotho allele VS causes a 6pt IQ increase in heterozygotes: http://cell.com/cell-reports/abstract/S2211-1247(14)00287-3 Note its frequency in Ashkenazi Jews