Paper on proposed mechanism of the Fragile X genetic variant:
- mRNA of the extended repeat region binds to the promotor region of the gene
- Example of ‘epigenetic silencing’ driven by promotor-bound trinucleotide repeat.
- FMR1 silencing is mediated by the FMR1 mRNA- which contains the transcribed CGG-repeat tract as part of the 5’UTR. This tract hybridizes to the complementary CGG repeat portion of the FMR1 gene to form a RNA-DNA duplex (in a direct interaction)
- Disrupting the interaction of the mRNA with the CGG repeat portion of the FMR1 gene prevents promotor silencing.
Colak, 2014, Science 343, 1002
Demonstration of women women heterozygous for FXS mutation who are not actually mentally retarded (i.e. known gene is systematiclly affecting IQ within the main body og the IQ distribution.
Series of tests for cognitive functioning, 2 main comparisons:
- Women heterozygous for FXS full mutation compared to women with a premutation and women without FXS who grew up in FXS families.
- Women with FXS compared to women without FXS matched on age and IQ.
- NB Studied women heterozygous for FXS and who were not actually mentally retarded. All had completed 12 years of education. Average full scale IQ was 82.7.
Results:
- Women with FXS had a significantly lower IQ than the other groups. Analyses of subtest profiles showed they had a relative weakness on Arithmetic and strength on Picture Completion. In general, FXS women performed worse on executive function, spatial ability and visual memory.
- Overall, results indicated that executive rather than visuospatial deficits were primary in the cognitive profile of FXS.
- Measured, in individual women, the bias in activation of one or other X chromosome- interestingly this measure correlated in the expected direction with performance IQ and EF.
- Carriers of pre-mutations differ little from controls.
Bennetto et al 2001
Recently discovered form of non-syndromic intellectual disability is associated with single-nucleotide variations (SNVs) of gene IQSEC2
- As part of a recent X chromosome exome resequencing project, identified for different nonsynonymous SNPs in IQSEC2 in 4 separate families with XLID (X chromosome linked intellectual disability)
- All 4 nonsynonymous sequence variants were predicted to lead to nonconserved aa substitutions.
- In the 4 families originally studied, the carrier females typically exhibited mild learning disabilities- although authors note this could not be unequivocally attributed to their IQSEC2 carrier status.
Shoubridge et al 2010 Nature Genetics
Another case where heterozygotes fall between the mean scores of the control and homozygous groups.
Method/background
- Individuals with cystinosis exhibit specific cognitive deficits in visual spatial function.
- Investigated whether obligate heterozygotes of the CTNS mutation have the same pattern of cognitive functioning seen in homozygotes (aberrant visual-spatial functioning against a background of relatively intact visual-perceptual functioning and overall cognitive ability.)
- 254 adults (100 heterozygotes, 154 controls) aged 17-74.
- Tested intelligence, visual perceptual functioning and visual spatial functioning
Results:
- Cystinosis heterozygotes demonstrated intelligence within the normal range, performed similarly to controls on tests of visual-perceptual ability, but showed significantly poorer performance on tests of visuo-spatial ability.
- Thus obligate heterozygotes with the CTNS mutation (i.e. carriers) display a similar pattern of visual processing decrements as individuals with cystinosis.
- Decrements were not as marked as those found in cystinosis, suggesting a gene dosage effect.
Niemiec et al 2012 Amer. J. Med. Genet. 158A 1902
Summary: measured the IQs of early-treated 8 year old PKU patients. Mean Wechsler FSIQ was only 91.1. Performance IQ was lower (89.4) than verbal (94.7). Therefore PAH gene is still contributing to the genetic variance within the main body of the IQ distribution.
- Mildly depressed IQ is common in treated PKU. Explored whether a particular intellectual ability typifies early and continuously treated PKU, and whether component skills comprising the IQ relate to socioeconomic & treatment factors.
- IQ scores collected retrospectively from variants of the WAIS for children performed at age 8, for 57 children with early treated, classic PKU. Analysed subscale and subtest scores and dietary factors.
Results:
- Childrens mean full scale IQ = 91.11, significantly below the healthy population norm.
- Significant discrepancy between mean verbal IQ and mean performance IQ suggesting a spatial deficit, but date did not support a biochemical or sociological explanation.
- PKU control at age 2 was predictive of overall IQ
Concluded: early & continuous treatmen of PKU does not necessarily lead to normalisaiton of overall IQ. Discrepancy in skill deficit is not the result of social status or treatment variables.
Griffiths et al 2000
GWAS.
- No individual SNP was significantly correlated with intelligence after statistical corrections were made
- However, authors grouped candidate genes into functional groups (e.g. NT metabolism, ion balance/ transport)- found that one of these groups, G proteins, was strongly related with intelligence with the probabilities for the individual SNPs were cumulated.
Ruano et al 2010 American Journal of Human Genetics
- Curiously does not cite the Ruano study
- Used data from 5 different GWAS studies.
- Similarly find no SNP that is significantly associated with either crystallised or fluid intelligence after correction for multiple comparisons
- Then used recently developed method to test the cumulative effects of all of the genotyped SNPs. Calculate how much variance is accounted for by all the 500,000 SNPs they tested. (Essentially calculate the genetic similarity between each possible pair of participants, and then ask how similar they are in phenotype)
- Found heritability of 40% for variation in crystalized intelligence and 51% for variation in fluid intelligence.
Davies et al 2011 Molecular Psychiatry
- Most important figure: estimated that cumulatively the ~630,000 SNPs in the data account for 47% of the variance in g.
- Sought to replicate published associations between 12 specific genetic variants and g, sing three independent, longitudinal data sets of 5571, 1759 and 2441 well characterized individuals.
- Although analyzed the data sets separately, combined sample size is almost 100000- gives considerable statistical power.
- In the three data sets, out of 32 comparisons, onl one was sattistically significant: involved SNP rs27660118. Association between this SNP aland intelligence also approached significance in analyses of another data set, but in the opposite direction.
- If null hypothesis was correct, would expect to see 1.6 significant associations (5% of 32 tests)- so the actually observed one nominally significant association was slightly less than would be expected by change. This result is not likely to be due to lack of statistical power. Power analysis shows that is all associations tested had been true positive with effect size of (Rsquared) = 0.1%- then would still expect 10.6 total associations, i.e. 10x more than were found.
Finally, used recently developed method for testing cumulative effects of all of the genotyped SNPs, applied this to the STR sample from study 3
Estimated that the ~630,000 SNPs in the data account for 47% of the variance in g.
‘These and other results, together with the failure of whole-genome association studies of g to date, are consistent with general intelligence being a highly polygenic trait on which common genetic variants individually have only small effects.
Chabris et al 2012
Method: Used genome-wide SNP data from 1940 unrelated individuals whose intelligence was measured in childhood (11 years) and again in old age (65, 70 or 79),
Esimated genetic (co)variance.
Results:
- Estimate that causal genetic variants in linkage disequilibrium with common SNPs account for 0.24 of the variation in cognitive ability change from childhood to old age.
- Estimate genetic correlation between childhood and old age of 0.62.
- Conclude: this warrants the search for genetic causes of cognitive stability and change.
Deary 2012
2 hypotheses:
- The same genes affect g throughout development.
- Even though, throughout development, heritability increases
- These hypotheses are based on genetic findings from twin research- this study sought to test them based on DNA alone. From 1.7 million DNA markers and g scores at ages 7 and 12 on 2875 children.
Results:
- DNA genetic correlation from 7 to 12 was 0.73, highly similar to genetic correlation of 0.75 estimated from 6702 pairs of twins from the same sample.
- DNA-estimated heritability increased from 0.26 at age 7 to 0.45 at age 12; twin-estimated heritabilities also increased from 0.35 to 0.45
- These DNA results confirm the results of twin studies: indicating strong genetic stability, but increasing heritability for g, despite mean changes in brain structure and function from childhood to adolescence.
- NB genetic correlation = the correlation between genetic effects on g at the two ages independent of heritability. The high genetic correlation implies that if a gene is found to be associated with g in childhood, that gene is also highly likely to b e associeated with g in adolescence*
Trzakowski 2014
Meta analysis of reported associations between the COMT Val158/108Met polymorphism and measures of memory and executive function.
- Found no association between COMT genotype and the majority of phenotypes.
- There was evidence of association with IQ score (d=.06), which did not differ significantly by ancestry, sex, average sample age or patient status.
- BUT there was also evidence of publication bias and decreasing effect sizes with later publication.
- Conclusion: Despite initially promising results, the COMT Val158/108Met polymorphism appears to have little if any association with cognitive function. Publication bias may hamper attempts to understand the genetic basis of psychological functions and psychiatric disorders.
Barnett et al 2008
